From comsat@vtcs1 Tue Jan 7 09:33:06 1986 Date: Tue, 7 Jan 86 09:33:02 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a007763; 6 Jan 86 14:28 EST Date: Mon 6 Jan 1986 11:10-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #1 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 7 Jan 86 07:19 EST AIList Digest Monday, 6 Jan 1986 Volume 4 : Issue 1 Today's Topics: Policy - Welcome & Technology Export Policy, Games - Wargamers List & Othello Tournament & Computer Chess Tutor ---------------------------------------------------------------------- Date: Sun 5 Jan 86 23:21:57-PST From: Ken Laws Subject: Introduction to Volume 4 Welcome to AIList'86. We went through 193 issues last year, and a high proportion of that was interesting and perhaps even useful. For those who haven't seen the official welcome message in the last 2 1/2 years, I've been telling the new arrivals that the list topics are: Expert Systems AI Techniques Knowledge Representation Knowledge Acquisition Problem Solving Hierarchical Inference Machine Learning Pattern Recognition Analogical Reasoning Data Analysis Cognitive Psychology Human Perception AI Languages and Systems Machine Translation Theorem Proving Decision Theory Logic Programming Computer Science Automatic Programming Information Science I like to think of AIList as the forum for AI and pattern recognition, although we've had precious little of the latter. There are a number of related lists, some sparked by the success of AIList. Prolog-Digest@SU-SCORE was here first, of course, and I thank Chuck Restivo for the help he gave me in getting started. Human-Nets@RUTGERS also served as a template for AIList. Recently created lists are Soft-Eng@MIT-XX for programming languages and man-machine interfaces; Vision-List@AIDS-UNIX for vision algorithms; AI-Ed@SUMEX-AIM for AI in education (CAI, tutoring systems, user modeling, cognitive learning, etc.); PARSYM@SUMEX for parallel symbolic computing; IRList%VPI.CSNet@CSNet-Relay for information retrieval; MetaPhilosophers%MIT-OZ@MIT-MC list for philosophy discussions; and the Usenet net.math.symbolic and computers-and society discussions. Discussions of workstations and individual languages are carried on WorkS@Rutgers, SLUG@UTexas (Symbolics), and Scheme@MIT-MC. (If someone wants to spin off other topics, such as linguistics, seminar announcements, etc., I'll be glad to help.) The digest goes out to a great many readers via bboards, redistribution nodes, and Usenet forwarding. I know that places like MIT and Xerox have hundreds of readers, but I don't have even a rough estimate of the total readership. My direct distribution (after 738 revisions) is to Arpanet Hosts: ACC(BB+1), AEROSPACE(8), AFSC-SD, AIDS-UNIX, ALLEGRA@BTL, AMES-NAS, AMES-VMSB(4), AMSAA(3), ANL-MCS, APG-3(2), ARDC(3), ARE-PN@UCL-CS, ARI-HQ1(10), BBN(1), BBNA(BB+1), BBNCC4, BBNCCH(2), BBNCCS, BBNCCT(3), BBNCCX, BBNCCY(2), BBN-CLXX, BBNF, BBNG(14), BBN-LABS-B, BBN-MENTOR, BBN-META, BBN-SPCA, BBN-UNIX(9), BBN-VAX(6), B.CC@BERKELEY, D@BERKELEY.EDU, UCBVAX.BERKELEY.EDU, UCBCAD@BERKELEY(2), UCBCORY@BERKELEY, UCBDALI@BERKELEY(2), UCBERNIE@BERKELEY(4), UCBESVAX@BERKELEY, UCBIC@BERKELEY, UCBLAPIS@BERKELEY, BNL44, BRL(BB+1), BRL-VOC, C.MFENET@LLL-MFE, CECOM-1, CECOM-2, PCO@CISL-SERVICE-MULTICS, CIT-20, CIT-HAMLET, CIT-VAX, CMU-CS-A(BB+5), CMU-CS-G(2), CMU-RI-ISL1, COLUMBIA-20, CORNELL(BB+1), CRYS.WISC.EDU, CRDC-VAX2, CSNet-SH, DCA-EMS(2), DCT%DDXA@UCL-CS, DDN, HUDSON.DEC.COM(2), DEC-MARLBORO(2), Other.DEC@DECWRL(23), DMC-CRC, DOCKMASTER(2), DREA-XX, EDN-UNIX, EDN-VAX(2), EDWARDS-2060, EGLIN-VAX, ETL-AI, FORD-COS1, FORD-SCF1(2), FSU.MFENET@LLL-MFE(3), GE-CRD(2), SCH-GODZILLA@SCRC-STONY-BROOK(2) GSWD-VMS(BB+1), GUNTER-ADAM, GWUVM@MIT-MULTICS(3), HARV-10, HAWAII-EMH, HI-MULTICS(BB+1), CSCKNP@HI-MULTICS, HOPKINS, IBM-SJ, ISIA, ISI-VAXA(10), JPL-VAX, JPL-VLSI(7), KESTREL, LANL, LBL-CSAM, LLLASD.DECNET@LLL-CRG, LLL-CRG, LLL-MFE(6), CMA@LLL-MFE, DMA@LLL-MFE, ORN@LLL-MFE, PPL@LLL-MFE(2), SAI@LLL-MFE, LL-VLSI, LL-XN, LOGICON, MARYLAND, MCC-DB@MCC(2), AI@MCC(2), CAD@MCC, PP@MCC, MIT-MC, MIT-MULTICS, ADL@MIT-MULTICS, MIT-OZ@MIT-MC, MITRE(14), MITRE-BEDFORD, MITRE-GATEWAY(3), MOUTON, MWCAMIS@MITRE, MWVM@MITRE, NADC(9), NBS-VMS, NCSC(3), NLM-MCS, NOSC(BB+4), CCVAX@NOSC, NOSC-F4(BB+5), COD@NOSC(5), TETRA@NOSC, NPRDC(BB+3), NRL-AIC, NRL-CSS, NSWC-WO(2), NTA-VAX(BB+2), NTSC-74, NUSC, NYU, NYU-CSD2, OAK.SAINET.MFENET@LLL-MFE, MDC@OFFICE-1, OMNILAX, ORNL-MSR(BB+1), OSLO-VAX, PAXRV-NES, PURDUE, RADC-MULTICS, RADC-TOPS20, RAND-UNIX(BB+1), RDG.AM.UTS@UCL-CS, RIACS, RICE, ROCHESTER(3), RUTGERS(BB+1), SAIL(BB+3), SAN.SAINET.MFENET@LLL-MFE, SANDIA-CAD, SCRC-STONY-BROOK(5), SECKENHEIM-EMH, SIMTEL20, SRI-AI(BB+6), SRI-CSL, SRI-KL(19), SRI-NIC(BB+1), SRI-SPAM, SRI-TSC(3), SRN-VAX, STL-HOST1, SU-AMADEUS@SU-SCORE, SU-CSLI(BB+1), SU-GSB-HOW(2), SUMEX(BB+3), SU-PSYCH(3), SU-SCORE(BB+8), SU-SIERRA(BB+2), SU-SUSHI(4), SYMBOLICS(2), TKOV02.DEC@DECWRL, UCBKIM, UCL-CS(BB+1), CAMJENNY@UCL-CS, UK.AC.EDINBURGH@UCL-CS(2), RLGM@UCL-CS(3), UCLA-LOCUS(BB+2), UCSD, UDEL, A.CS.UIUC.EDU, MIMSY.UMD.EDU, VAX.NR.UNINETT@NTA-VAX, VAX.RUNIT.UNIT.UNINETT@NTA-VAX(3), USC-ECL, USC-ISI(8), USC-ISIB(BB+5), USC-ISIF(6), UTAH-20(BB+2), UTEXAS-20, WASHINGTON(4), WHARTON-10(2), WHITNEY, WISC-AI, WISC-CRYS(5), WISC-GUMBY, WISC-PIPE, WISC-RSCH(2), WISCVM, WPAFB-INFO1, WPAFB-AFITA, WSMR04, WSMR06, XEROX, YALE CSNet: BGSU, BOSTONU(3), BRANDEIS, BROWN, BUFFALO, CLEMSON(3), COLGATE, COLOSTATE, DEPAUL, GATECH, GERMANY, GMR(12), GTE-LABS(2), HP-BRONZE, HP-LABS, SJRLVM1%IBM-SJ, WLVM1%IBM-SJ, IRO.UDEM.CDN%UBC, CSKAIST%KAIST(2), LOSANGEL%IBM-SJ, LSU, NMSU(2), NORTHEASTERN(11), OKSTATE, PITT, RPICS, SCAROLINA(3), SMU(BB+1), SPERRY-RESTON, TAMU, SPY%TEKTRONIX, TEKCHIPS%TEKTRONIX, TEKIG5%TEKTRONIX, TEKGVS%TEKTRONIX, TEKLDS%TEKTRONIX, TENNESSEE, CSL60%TI-CSL(BB+1), TI-EG, UTAI%TORONTO, TUFTS, UBC, UCF-CS, UCI, UCSC, UIOWA(BB+1), ARCHEBIO%UIUC, UIUCDCSB%UIUC, ULOWELL(2), UMASS-CS, UMN-CS, UNC, UPENN, VIRGINIA, VPI, WWU(2), YKTVMV@IBM-SJ(7) BITNET@WISCVM: BOSTONU(2), BROWNVM, BUCASA, BUCKNELL, CARLETON, CBEBDA3T, CGEUGE51, CUNYVM(2), CZHRZU1A, DB0TUI11, DBNGMD21, DBNRHRZ1(2), DBSTU1, DDATHD21, DDATHD21, DHDURZ2(2), HNYKUN52, HNYKUN53, HWALHW5, ICNUCEVM, IDUI1, IPACUC, ISRAEARN, NJECNVM, NSNCCVM, RYERSON(2), SBBIOVM, SLACVM.WISC.EDU, SUCASE, UCF1VM(2), UCONNVM, UHUPVM1, UKCC(2), ULKYVX, UMCVMB, VTVM1, WISDOM, WSUVM1 BITNET@BERKELEY: CORNELLA(2), HLERUL5, UTCVM(2), VPIVM2, WESLYN Mailnet@MIT-MULTICS: Grinnell, NJIT-EIES, RPI-MTS, UMich-MTS, VANDERBILT Usenet Paths: bellcore@BERKELEY, franz@BERKELEY, ucscc@BERKELEY, sdcsvax!sdamos!crash@NOSC, mcvax!inria!imag!csinn@SEISMO.CSS.GOV, dec-rhea!dec-gvaic3@DECWRL, mcvax!enea!erix@SEISMO mcvax!cernvax!ethz@SEISMO.CSS.GOV, packard!ihesa@SEISMO.CSS.GOV, mcvax!ircam@SEISMO.CSS.GOV, mcvax!inria!lasso@SEISMO.CSS.GOV, mcvax!cernvax!unizh@SEISMO, tflop@SU-SHASTA, vitesse!vec_j@S1-C This includes something like 35 government and military sites, 15 national laboratories and research institutes, 55 companies and nonprofit corporations, and 100 universities around the world. That's just the direct mailings; my thanks to the many people who have established and maintained local bboards and remailers. About a year ago I began to worry that the international nature of the list might violate President Reagan's directives concerning unclassified technical (export restricted) and unclassified national security-related (UNS-R) information. (The list goes to Canada, Britain, Australia, West Germany, Norway, the Netherlands, Switzerland, Japan, South Korea, etc. Readers in these countries have also contributed to the list, of course.). I sent out some queries and received a great deal of informed discussion, but there were no firm precedents for determining whether we were headed for trouble. The whole file is available for those who want it; just write to AIList-Request@SRI-AI.ARPA. I have attempted, at least four times, to summarize the material, but have been unable to do so without losing the critical context of each opinion. The policy I have settled on (subject to revision) is the following: AIList is a public information service provided to the Arpanet community and others though my own efforts, indirect support from my company, and the help of numerous individuals and organizations at other sites. Readers are advised not to submit any material that is export controlled or classified. As moderator, I must assume that individuals have obtained all required clearances for their submissions to the list and for the university bboard messages that AIList occasionally reprints. The export control laws are both broad and vague, but material that could be published in news magazines or publicly available scientific journals is probably safe. Scientific information "without engineering or military significance" is always permissible, but technical details of specific military or government- controlled systems should not be discussed in this forum. I would also like to point out that, in my own opinion, technology transfer via informed discussion and incremental question/answer exchanges can be far more effective than by flooding a channel with printed technical material. Indeed, that is the very reason for AIList's existence -- to put people in touch with those who can help them the most. Readers at government-supported sites should keep in mind that any exchanges of reports or technical data resulting from "friendly contacts" on the AIList are their own responsibility, and that care should be taken when communicating over unsecure channels or with unknown individuals. For those participants who regard the above as paranoid, I apologize for any offense. The critical decisions concerning U.S. policy and network policy are not mine to make; I merely interpret them as best I can. I am comfortable with the level of exchange that AIList has promoted, and grateful for the broad participation that has made the list such a success. -- Dr. Kenneth I. Laws Computer Scientist SRI International ------------------------------ Date: Fri 3 Jan 86 23:39:12-EST From: "Daniel F. Lane" Subject: Wargamers! [Forwarded from the MIT bboard by Laws@SRI-AI.] To anyone interested in wargames, or strategy games in general, myself and SWF at OZ are starting a wargamers mailing list. Discuss the latest games out on the market, etc. Also, as soon as we get it all organized (if?) we will be running a game over the net called "Battle for North America" or, "The Second Battle-Between-the-States" (whichever you prefer). But, more about that after we get the group rolling. Send any addition requests to: WAR-GAMES-REQUEST%MIT-OZ@MIT-MC. Thanks, Daniel Lane (GZT.LANE@OZ) ------------------------------ Date: Thu, 2 Jan 86 17:10 EST From: Kurt Godden Subject: Computer Othello Tournament I received an announcement in the mail and am passing it along to ailist out of the goodness of my heart. I am not connected in any way with the tournament (other than as an entrant): 1986 North American Computer Othello Championship Tournament Host: CS Association at California State University, Northridge When: February 15-16, 1986. Where: Cal State campus in Northridge (LA area) Sanctioned by:U.S. Othello Association >From flier: "...an eight-round, Swiss-style event with awards for the winners, and is open to computers of all makes, models, and sizes. Participation from programmers anywhere in the world is welcome; entrants need not be present, as they may play via phone or submit software and/or hardware to be run by volunteer representatives." For detailed info you are requested to contact 1) North American Computer Othello Championship CSUN Computer Science Association School of Engineering, Box 31 18111 Nordhoff Street Northridge, CA 91330 2) Brian Swift or Marc Furon (apparently pronounced ['fju ren] -KG) at 213-852-5096 Please don't contact me. -Kurt Godden p.s. Presumably it's necessary or at least polite to note that 'Othello' is a registered trademark of CBS Toys. ------------------------------ Date: Mon 6 Jan 86 07:17:05-EST From: "Fred Hapgood" Subject: computer chess tutor Would anybody know who might be thinking about tutor/annotator functions in chess computers? The simplest imaginable computer chess tutor might work like this: After one had played a game against it one would indicate the moves one wished to see annotated. The machine would retrieve the positions in that range on which you had the move. For each it would run its evaluation routine to see what move it would have made had it been playing. It would then score both (a) the position resulting from your move and (b) that resulting from the move generated by its own routine. This done, it would move on to the next move in the series and repeat the procedure. One could of course enter an entire score, perhaps from a newspaper, and have the computer perform this function for the moves of both sides. When the list was exhausted the machine would find all the cases in which the evaluator routine scored a difference between (a) and (b) of more than a defined amount. It would then display these cases either by replaying the game and stopping at the points found, or in order of greatest disparity, i.e., biggest blunder first. In either case display would consist of: (i) the original position, (ii) the move actually made, and (iii) the improvement claimed by the machine, together with a short list of the best subsequent moves for both sides. This is only the simplest instance of how a machine might comment on a position or 'explain' itself. From a marketing point of view, one virtue of these devices is that tutors can never get too strong. A person buying a chess computer as a opponent is likely to drop out of the market for new versions once the machines have gotten strong enough to be a challenge. What is being sold in an annotator is authority, and one can never get enough of that. In fact, it is possible that as chess computers improve, the forces driving that market will shift from using chess computers as calculators to using them as annotators. ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Thu Jan 9 18:57:45 1986 Date: Thu, 9 Jan 86 18:57:41 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a018861; 8 Jan 86 12:17 EST Date: Wed 8 Jan 1986 08:57-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #2 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 9 Jan 86 09:11 EST AIList Digest Wednesday, 8 Jan 1986 Volume 4 : Issue 2 Today's Topics: Corrections - Feigenbaum's Comments & Xerox Reader Count, Query - AI Paradigm, Review - Stanford SDI Debate (12/19) ---------------------------------------------------------------------- Date: Sun 29 Dec 85 22:18:33-PST From: Edward Feigenbaum Subject: re the news report on my speech in the Netherlands Saw my name in the Fri 27 Dec. 1985 AIList Digest V3 #192. Since it's best not to let silly things propagate, let me say here what I said (I actually said many many things; I don't understand why those few things were picked out). I said that among the most commercially important applications of expert systems in the next ten years would be factory management applications and financial service applications. (I didn't even mention factory automation). I said that speech understanding applications would become economically very important. (I never mentioned speech generation.) Best wishes for a journalistically accurate New Year (fat chance), Ed Feigenbaum ------------------------------ Date: 6 Jan 86 16:22 PST From: Newman.pasa@Xerox.ARPA Subject: Reader Count I don't know if you want to post this to the net or not, but in the interest of accuracy, Xerox has approximately 248 readers of the AIList. >>Dave ------------------------------ Date: Tuesday, 7 January 1986 02:19:31 EST From: Duvvuru.Sriram@cive.ri.cmu.edu Subject: AI Paradigm I have seen the word "AI Paradigm" in several papers/reports. My dictionary [Random House] says that a Paradigm is either an example or a model. Is there any other meaning to it or is it just a better word for "example"? Sriram ------------------------------ Date: Tue, 31 Dec 85 11:55:37 pst From: jrisberg@aids-unix (Jeff Risberg) Subject: Stanford SDI debate (12/19) summary The following is a somewhat long summary of the technical debate on SDI entitled "SDI: How Feasible, How Useful, How Robust?" that was held at Stanford on December 19th. Since this debate was announced on AILIST, we felt that readers would be interested in this summary. [I was reluctant to permit the initial announcement and I am reluctant to permit the summary. I have decided to forward them because SDI may well involve major funding in the area of AI. Please restrict any discussion in AIList to the areas of AI, pattern recognition, or the feasibility of distributed decision making. Political discussions would be more appropriate on Arms-D@MIT-MC, Risks@SRI-NIC, or perhaps Space@MIT-MC. -- KIL] The panelists at the debate were: Advocates: Professor Richard Lipton, Professor of Computer Science at Princeton University, current member of SDIO's Panel on Computing and Support of Battle Management. Major Simon Peter Worden, the Special Assistant to the Director of the SDIO and Technical Advisor to the Nuclear and Space Arms Talk with the USSR in Geneva. Opponents: Dr. Richard L. Garwin, IBM Fellow and Adjunct Professor of Physics at Columbia University, Physicist and Defense Consultant. Professor David Parnas, Lansdown Professor of Computer Science at the University of Victoria, former member of the SDI Organization's Panel on Computing and Support of Battle Management. Dr. Goldberger, President of CalTech, served as the moderator of the discussion. He presented a bit of history relating to the subject of defensive warfare and then allowed the panelists to speak. There are certainly historical precedents for defensive systems, in fact, each US leader since the 1950's has sought a defense. SDI is simply the largest scale and most visible concept to date. Because of the complexity of the issue, a question like "can it work?" can only be answered by determining 'what does "can" mean?', 'what does "it" mean?', and 'what does "work" mean?'. There have been various justifications proposed for SDI, and the technical and political community has raised numerous questions. The format of the debate consisted of three sections: during the first section, each speaker was allowed 20 minutes with which to present his case; following that, each speaker had a 5-minute rebuttal period; finally, the audience was allowed to ask questions via screened 3 x 5 index cards. Major Worden spoke first and discussed some positive aspects of SDI. In his view, the principal justification is for Arms Control. In that view, a major goal is that the SDI system have a lower marginal cost than that of building additional offensive systems. Survivability is another goal. He said that the object of SDIO is to establish the feasibility of the system, but not to build it. Similarly, it may not involve space weapons, although most of the current concepts include a space segment. We would also like to get the Soviets to admit their own work in such systems. They regularly deny such work, but when we show them ariel photos of their high-power lasers, they say, "Oh, we do have laser research for medical purposes". The numerical aspects of a Defensive Reliant Deterance are that each layer of defense drives the number of offensive warheads needed up further. There are a series of layers, with each layer consisting of sensors, weapons, and battle management systems. He showed some of the standard slides of this design. The concept is "Proliferated and Distributed". They are planning for the late 1990's. The key issue as he sees it is countermeasures, such as the fast burn booster. There are three types of threats and five types of countermeasures. He expected that countermeasures will develop much as aircraft gained shielding, speed, and proliferation countermeasures after WW I. He gave a cost context of the project in comparison to the cost of insurance in the private sector. The cost of insurance is over $300 billion/year, while the SDI work is currently costing $1.5 billion. Major Worden admitted that President Reagan caught everyone off guard with his speech about SDI two years ago. Dr. Garwin spoke next. He recommended that we think carefully about just what are the goals, costs, and likely Soviet Response to SDI. He said that the Scowcroft Commission reported that U.S. security could be maintained without SDI. His opinion is that while SDI has been proposed to replace deterrance, it is really simply another form of deterrance. He is concerned with the layered approach of the system in that there is catastrophic failure if one layer does not do its job. For example, the design of each layer assumes that the prior one does its job in reducing the number of incoming objects. Examples of potential areas of failure were given: space mines could easily knock out any space segment units and midcourse intercept could be overwhelmed by large numbers of decoys. Dr. Garwin feels that the systems needed for SDI can not be built under the ABM treaty. There has been a progression of goals, and in effect "replace deterance" has become "strengthen deterance". He closed by describing his view of a viable strategic balance, which would be to limit each side to 1000 warheads, deployed on small missiles, small subs, and cruise missiles, with no counterforce threats against strategic targets. Preserve the ADM treaty. BMD research may be continued, in order to confirm that there is no threat to the system. Dr. Lipton was the next speaker. He joined the technical panel of SDIO last summer. (An interesting point is that Dr. Lipton is a former student of Dr. Parnas.) His major focus was in the importance of the de-centralization of software. The Fletcher panel design was centralized, with software in charge of everything. "Is it possible to build a system without these problems?" Discussion of feasibility must encompass all design possibilities, and Lipton stressed the merits of a decentralized design. He led into this by an analogy with the banking industry. Banking works because it is a large collection of loosely organized components. In the SDI example, he refered to large numbers of satellite groups handling independent battle management functions. Fault tolerance would provide reliability, like the concept of the strategic weapons triad. He argued that these seperate groups would be testable, by putting a few into orbit and shooting missiles at them. The false alarm problem could be controlled by activating different numbers of systems. Coordination problems were raised by the Fletcher panel toward a goal of conserving "bullets". Dr. Lipton's studies indicated that the shot overhead of low coordination is not that high. Dr. Parnas spoke last. He had good points in his speech, but had a problem with becoming quite caustic in his remarks about the SDIO members. He said that (loosely quoted) "I used to feel that arms control people are guilty of wishful thinking, but I have now seen a whole new standard." His major complaint against SDI is that SDI forces us to trust the system; if SDI need not be perfect, it must at least be trustworthy, and he feels that this is not possible. Conditions for validation include: mathematical analysis, exhaustive case analysis, or prolonged, realistic testing. Even after one or more of these conditions have been met, the system must still be operated under controlled conditions. The validation of software is inherently different from the standard engineering problems such as bridge design. Something different about software. It is made up of discrete, rather than continuous functions. Thus design principles such as building for twice the weight do not truly apply, but instead, the number of discrete cases must be examined, along with thorough testing. Even after a number of years of use, bugs may still be found. True testing requires thousands of years. For most software, we can allow unreliable software, as long as we trust it. For SDI, we cannot. He doesn't believe that de-centralization provides added trustworthiness to the system. He stated that he never took the Fletcher design seriously in the first place, feeling it that was no more than a rough sizing of the problem. There are a series of myths around de-centralization. Dr. Parnas' final point is that SDI is not a limit of computers, but of human beings. The rebuttals were then held. Major Worden questioned the meaning of deterence and then mentioned some possible alternatives to SDI: automatic launch under attack, preventive attacks, and bombs under U.S. cities. He indicated that Dr. Garwin had shown only that he could design a system that SDIO wouldn't buy. His final comment re-iterated the linkage of SDI to the arms control process. Dr. Garwin (and the others) kept mentioning the Scowcroft report which produced possible defensive measures other than SDI. Dr. Garwin pointed out that Dr. Lipton had only found that the system proposed by Fletcher might not work, but that Lipton believed others might. In any case, so long as Soviets can deliver by other means (cruise missiles), we will continue to need deterance. Dr. Lipton restated his belief in the need for independent systems. He recognized that nothing is perfect, that even computers are not reliable, but they are used on a daily basis. The use of independent battle stations would stress the sensors, but he argued that teraflops would alleviate the need for independent views. Dr. Parnas again made a couple of inappropriate shots..."I've been to a lot of Mickey Mouse meetings, but the ones sponsored by SDI had the biggest ears and biggest nose I've ever seen." He thinks that the idea of separate systems does not remove the size or complexity of what is needed; dividing 10 million lines of code into small modules of 1000 lines does still not ensure error-free code. People do not write independent code. Questions raised from the floor asked about different types of lasers, the time to phase-in to SDI, and about the non-ICBM threat. Worden replied that cruise missiles are not strategic weapons because of their flight time, and that smuggling bombs into the US would not be a realistic approach for a Soviet leader to taken. The speakers then made closing comments: Dr. Garwin said that we currently have a real opportunity for arms reduction. This would be much more survivable than continued escalation and research into defensive weaponry. He feels that both sides should abandon defense efforts. Control of nuclear proliferation is essential. Major Worden agreed with these points of Dr. Garwin, but said that it is necessary and vital to carry forward a defensive program within the ABM treaty to provide a different kind of security. Dr. Parnas said that in software, the engineering term of "tolerance" depends on continuity. "Almost right" does not make sense in the context of SDI. He fears espionage that would result in someone getting a copy of the software. Reasons for not going ahead with SDI anyway include the lost opportunity for other projects, low quality of results, and weakening of the strategic position. Dr. Lipton said that if the independent segments of the SDI system do not interact, the code is not vulnerable. He pointed out that there are simple systems, such as elevators, that we do trust. Dr. Goldberger then made some closing comments. He said that strategic defense and arms control must be approached seriously. The laws of physics are immune to political views and we are currently at a critical political point. A decision to push forward defensively without a reduction in offense would be a mistake. SDI has been proposed as part of current moves toward lowering threat of destruction, yet it is difficult, with verification problems, and major risks. He hopes that the human spirit will prevail in the decisions that must be made. In summary, the debate was quite interesting, although inconclusive if judged in a strict manner. We were most surprised that software technical details were hardly mentioned, and that political and non computer technology issues were the focus of the discussion. Dr. Parnas and Dr. Lipton made several comments against each other, which detracted from the technical discussion. It didn't appear that Dr. Lipton was overly familiar with the SDI problems; he continually talked in generalities, with few facts with which to back up his statements. Dr. Garwin and Major Worden were much more prepared in their talks and didn't take any cheap shots with which to score points with the audience. The comments above are strictly our personal opinions and not representative of any organization. Jeff Risberg (jrisberg@aids-unix) Susan Rosenbaum (susan@aids-unix) ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Sun Jan 12 17:05:15 1986 Date: Sun, 12 Jan 86 17:05:12 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a002339; 12 Jan 86 1:28 EST Date: Sat 11 Jan 1986 22:13-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #3 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sun, 12 Jan 86 04:41 EST AIList Digest Sunday, 12 Jan 1986 Volume 4 : Issue 3 Today's Topics: Bindings - AI-Related Lists, Definition - Paradigm, Logic - New CSLI Reports, Reviews - Spang Robinson Report 2/1 & Rational Agency Seminars (CSLI) ---------------------------------------------------------------------- Date: Fri 10 Jan 86 12:18:09-PST From: Christopher Schmidt Subject: AI-Related Lists [...] To add to your list of AIList-related lists, Info-1100@SUMEX and Bug-1100@SUMEX are DL's concerning the Xerox 1100 series lisp machines and Interlisp, and Info-TI-Explorer@SUMEX and Bug-TI-Explorer@SUMEX are DL's concerning the TI Explorers and associated software. --Christopher ------------------------------ Date: Wed, 8 Jan 86 16:38:34 EST From: Bruce Nevin Subject: paradigm The term paradigm was specialized in philosophy of science by Thomas Kuhn in his 1965(?) book _The_Structure_of_Scientific_Revolutions and subsequent works. I would question whether AI is a mature enough field to have a paradigm in the sense that Kuhn intends for a mature science. Instead, there appears to be a fair selection of more or less divergent examples/models/agendas for each area of investigation. Many of these are associated with the more prominent investigators in AI. Bruce Nevin bn@bbncch.arpa BBN Communications 33 Moulton Street Cambridge, MA 02238 (617) 497-3992 [Disclaimer: my opinions may reflect those of many, but no one else need take responsibility for them, including my employer.] ------------------------------ Date: Wed, 8 Jan 1986 19:37 EST From: MINSKY%OZ.AI.MIT.EDU@MC.LCS.MIT.EDU Subject: AIList Digest V4 #2 about "paradigm" -- the dictionary is out of date because this word now almost universally refers to the notion in Thomas Kuhn's "Structure of Scientific Revolutions." It seems to mean powerful and influential idea, or something. ------------------------------ Date: Thu, 9 Jan 86 11:04:40 GMT From: Mmaccall%cs.ucl.ac.uk@cs.ucl.ac.uk Subject: Re: AI Paradigm An approximate meaning for the word `paradigm' is `template'. Gordon Joly gcj%qmc-ori@ucl-cs.arpa ------------------------------ Date: Fri, 10 Jan 86 16:53:46 GMT From: Mmaccall%cs.ucl.ac.uk@cs.ucl.ac.uk Subject: Re: AI Paradigm As an afterthought. The place where I first saw the term "paradigm" was in "Games People Play" by Eric Berne. Here, he has a model of the (transactional) relationship between two people, with three states of parent-adult-child. They are then put side by side with the parent above adult and the adult above child, each being represented by a circle. Lines are drawn to indicate which relationships are active in a given "game". The Chambers 20th Century Dictionary, as well as the Random House, gives the notion of "side by side". I hope this has a meaning for the "AI Paradigm"! Gordon Joly, gcj%qmc-ori@ucl-cs.arpa ------------------------------ Date: Thu 9 Jan 86 12:09:33-PST From: Wilkins Subject: Paradigm Your dictionary is correct about "paradigm". This word has been used extensively in the Ai literature in an incorrect way. People incorrectly use it to mean "methodology" or "school of thought" or some such. David ------------------------------ Date: Thu 9 Jan 86 15:29:34-PST From: Michael Walker Subject: ai paradigm? If you have a paradigm, there's always a chance that you'll get a paradigm shift, in which case people will fund your research for the next 20 years. On the other hand, if you say your example shifts, they'll think you're fudging your data. Mike ------------------------------ Date: Wed 8 Jan 86 16:53:32-PST From: Emma Pease Subject: New CSLI Reports on Logic NEW CSLI REPORTS Report No. CSLI-85-41, ``Possible-world Semantics for Autoepistemic Logic'' by Robert C. Moore and Report No. CSLI-85-42, ``Deduction with Many-Sorted Rewrite'' by Jose Meseguer and Joseph A. Goguen, have just been published. These reports may be obtained by writing to Trudy Vizmanos, CSLI, Ventura Hall, Stanford, CA 94305 or Trudy@SU-CSLI. ------------------------------ Date: Fri, 10 Jan 86 17:28:42 cst From: Laurence Leff Subject: Spang Robinson Report, Volume 2 No 1 Summary of Spang Robinson Report, Volume 2 Number 1, January 1986 featuring AI Hardware Vendors state that the biggest problem in marketing AI hardware is educating both internal people and the market place. An interview with a gentleman who evaluated AI type machines for use in developing software for silicon compilation research at Philips Labs. Discussion of various ways to enhance IBM PC's for AI (or other development needs) and the use of the Macintosh and Commodore's Amiga for AI research. C. J. Petrie of MCC described a system to parse text from a "how to" book into rules. Interview with Dag Tellefsen of Glenwood Management, a venture capitalist. They have funded Natural Language Products and AION. Kurzweil Applied Intelligence, that develops voice recognition hardware, has signed a joint marketing agreement with FutureNet which supplies electronic engineering work stations. Reasoning Systems has signed an agreement with Lockheed Missiles and Space Corporation to develop knowledge based systems for communications. (Reasoning Systems is involved with the commercialization of some of the techniques from the University of Southern California work in automating software development. See the IEEE Transactions on Software Engineering November 1985 Special Issue on AI and Software Engineering for more info.) "Logicware Inc. and Releations Ltd., both in Canada, have signed a long-term agreement to design an Artificial Intelligence language leading to a computer system which will emulate the thinking process of the human brain. It will be the first AI language designed for vector-processing by a super computer." Composition Systems has released two Artificial Intelligence kit that links VAX Lisp with such DEC product as FMS, RDB, GKS and DECNET." Review of the IEEE Computer Society Second Conference on Artificial Intelligence. ------------------------------ Date: Wed 8 Jan 86 16:53:32-PST From: Emma Pease Subject: Review - Rational Agency Seminars (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] RATIONAL AGENCY GROUP Summary of Fall 1985 Work The fall-quarter meetings of the Rational Agency Group (alias RatAg) have focused on the question: what must the architecture of a rational agent with serious resource limitations look like? Our attempts to get at answers to this question have been of two kinds. One approach has been to consider problems in providing a coherent account of human rationality. Specifically, we have discussed a number of philosophically motivated puzzles, such as the case of the Double Pinball Machine, and the problem of the Strategic Bomber, presented in a series of papers by Michael Bratman. The second approach we have taken has been to do so-called robot psychology. Here, we have examined existing AI planning systems, such as the PRS system of Mike Georgeff and Amy Lansky, in an attempt to determine whether, and, if so, how these systems embody principles of rationality. Both approaches have led to the consideration of similar issues: 1) What primitive components must there be in an account of rationality? From a philosophical perspective, this is equivalent to asking what the set of primitive mental states must be to describe human rationality; from an AI perspective, this is equivalent to asking what the set of primitive mental operators must be to build an artificial agent who behaves rationally. We have agreed that the philospher's traditional 2-parameter model, containing just ``beliefs'' and ``desires'', is insufficient; we have further agreed that adding just a third parameter, say ``intentions'', is still not enough. We are still considering whether a 4-parameter model, which includes a parameter we have sometimes called ``operant desires'', is sufficient. These so-called operant desires are medial between intentions and desires in that, like the former (but not the latter) they control behavior in a rational agent, but like the latter (and not the former) they need not be mutually consistent to satisfy the demands of rationality. The term ``goal'', we discovered in passing, has been used at times to mean intentions, at times desires, at times operant desires, and at times other things; we have consequently banished it from our collective lexicon. 2) What are ``plans'', and how do they fit into a theory of rationality? Can they be reduced to some configuration of other, primitive mental states, or must they also be introduced as a primitive? 3) What are the combinatorial properties of these primitive components within a theory of rationality, i.e., how are they interrelated and how do they affect or control action? We have considered, e.g., whether a rational agent can intend something without believing it will happen, or not intend something she believes will inevitably happen. One set of answers to these questions that we have considered has come from the theory of plans and action being developed by Michael Bratman. Another set has come come from work that Phil Cohen has been doing with Hector Levesque, which involves explaining speech acts as a consequence of rationality. These two theories diverge on many points: Cohen and Levesque, for instance, are committed to the view that if a rational agent believes something to be inevitable, he also intends it; Bratman takes the opposite view. In recent meetings, interesting questions have arisen about whether there can be beliefs about the future that are `not' beliefs that something will inevitably happen, and, if so, whether concomitant intentions are guaranteed in a rational agent. The RatAg group intends to begin the new quarter by considering how Cohen and Levesque's theory can handle the philosphical problems discussed in Bratman's work. We will also be discussing the work of Hector-Neri Castaneda in part to explore the utility of Castaneda's distinction between propositions and practitions for our work on intention, belief and practical rationality. Professor Castaneda will be giving a CSLI colloquium in the spring. RatAg participants this quarter have been Michael Bratman (project leader), Phil Cohen, Todd Davies, Mike Georgeff, David Israel, Kurt Konolige, Amy Lansky, and Martha Pollack. --Martha Pollack ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Sun Jan 12 17:02:33 1986 Date: Sun, 12 Jan 86 17:02:28 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a002392; 12 Jan 86 1:43 EST Date: Sat 11 Jan 1986 22:28-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #4 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sun, 12 Jan 86 04:43 EST AIList Digest Sunday, 12 Jan 1986 Volume 4 : Issue 4 Today's Topics: Seminars - Organization of Semantic Knowledge Systems (MIT) & LISP architectures (NASA Ames) & Computational Networks in Silicon and Biology (PARC), Course - Values, Technology, and Society (SU) & Highly Parallel Architectures for AI (UPenn), Conference - 3rd Symposium on Logic Programming ---------------------------------------------------------------------- Date: Sun, 5 Jan 86 03:44:31 EST From: "Steven A. Swernofsky" Subject: Seminar - Categorical Organization of Semantic Knowledge Systems (MIT) Monday 2, December 4: 00-6:00pm Room: E25-117 HARVARD UNIVERSITY-MIT DIVISION OF HEALTH SCIENCES AND TECHNOLOGY "The Categorical Organization of Semantic Knowledge Systems" Elizabeth K. Warrington Professor of Neurology The National Hospitals for Nervous Diseases Queen Square, London Patients with cerebral lesions provide an important source of evidence about the organization of semantic systems. Striking instances of the selective preservation and selective impairment in the comprehension of particular categories of verbal and visual stimuli have long been reported in the neurological literature and more recently such dissociations have been investigated and assessed using experimental methods. The issue of modality specificity will be discussed and it will be argued that there are at least partially independent systems that subserve verbal and visual semantics. Evidence for both broad category specific impairments, such as knowledge of concrete and abstract concepts, and more fine grain category impairments such as knowledge of animate and inanimate objects will be reviewed. It will be argued that there are modality specific semantic systems and that these are categorised in their organization. Host: Lucia Vaina ------------------------------ Date: Fri, 10 Jan 86 07:58:25 pst From: eugene@AMES-NAS.ARPA (Eugene Miya) Subject: Seminar - LISP architectures (NASA Ames) National Aeronautics and Space Administration Ames Research Center SEMINAR ANNOUNCEMENT Computational Research Branch SPEAKER: Raymond S. Lim Computational Research Branch TOPIC: LISP Machine Architectures of MIT CADR, Symbolics 3600, & TI Explorer ABSTRACT: Common LISP is becoming a standard, and MULTI-LISP is contemplating for parallel LISP Processing. A modern LISP machine is a conventional virtual memory, Von Neuman machine with addded hardware to support runtime data type checking and incremental garbage collection. This presentation will discuss the architecture issue of LISP machine, starting from the MIT CADR. DATE: 23 Jan 1986 TIME: 9:30-11:00 BLDG: 233 ROOM: 172 POINT OF CONTACT: Becky Getz PHONE NUMBER: (415)-694-5197 VISITORS ARE WELCOME: Register and obtain vehicle pass at Ames Visitor Reception Building (N-253) or the Security Station near Gate 18. See map below. Do not use the Navy Main Gate. Non-citizens (except Permanent Residents) must have prior approval from the Director's Office one week in advance. Submit requests to the point of contact indicated above. Non-citizens must register at the Visitor Reception Building. Permanent Residents are required to show Alien Registration Card at the time of registration. ------------------------------ Date: 10 Jan 86 14:27:59 PST (Friday) From: Kluger.osbunorth@Xerox.ARPA Reply-to: Kluger.osbunorth@Xerox.ARPA Subject: Seminar - Computational Networks in Silicon and Biology (PARC) Xerox Palo Alto Research Center Forum Thursday, January 16, 1986 4:00 pm, PARC Auditorium J.J. Hopfield Divisions of Chemistry and Biology Caltech and AT&T Bell Laboratories will speak on Computational Networks in Silicon and Biology The brain as a piece of computer hardware violates most of the sensible design criteria for good computers, yet manages to be extremely effective. We investigate the kinds of behavior which circuits built in a neuronal fashion--emphasizing large connectivity, large size, analog response, and self-timed--naturally have. The collective properties of such systems lead naturally to the behaviors needed for associative memory, or pattern recognition, error decoding, visual information processing and many complex optimization problems. At the same time, the circuits are relatively robust (fail soft), like their biological relatives. Such circuits may be of use as high density associative memories and as signal processors. The effectiveness of biological computation may in part result from the use of the collective decision capabilities of neural networks. This Forum is OPEN. All are invited. Host: Larry Kluger (Information Systems Division, 496-6575) Refreshments will be served at 3:45 pm Visitors: Welcome! The PARC Auditorium is located at 3333 Coyote Hill Road. The street is between Page Mill Road (west of Foothill) and Hillview Avenue, in the Stanford Research Park, Palo Alto. Enter the building through the *auditorium's* entrance, at the upper level of the building. ------------------------------ Date: 03 Jan 86 1404 PST From: John McCarthy Subject: Course - Values, Technology, and Society (SU) [Forwarded from the Stanford bboard by Laws@SRI-AI.] The following course will be given by John McCarthy in Winter 1986 in the Values, Technology and Society program. As will be noticed from the description, it will emphasize opportunities rather than problems. It will meet 14:15-15:30 Tuesdays and Thursdays in room 202 History corner (bldg 200). Technological Possibilities for enhancing man This course surveys the technological possibilities for increasing human capability and real wealth. It is oriented toward what people will want rather than around what we might think is good for them. Some of the improvements discussed are in the direction of (1) making housework trivial (2) making government responsive (3) increasing the ability of one person to build an object like a car, airplane or house to suit him without organizing others (4) allowing groups to live as they prefer less hindered by general social laws and customs. We will emphasize computer and information technology and ask what will be genuinely useful about computers in the home and not just faddish or flashy. To what extent are futurists and science fiction writers given to systematic error? Can we envisage advances as important as electricity, telephones, running water, inside toilets? The second topic concerns the social factors that determine the rate of scientific and technological progress. Why was scientific advance a rare event until Galileo? Why didn't non-Western cultures break through into the era of organized scientific and technological progress and why did it take Western culture so long? Why isn't the rate of progress faster today? As examples, we shall inquire into the obstacles that made cellular telephone systems and electronic funds transfer take so long. ------------------------------ Date: Wed, 8 Jan 86 16:33 EST From: Tim Finin Subject: Course - Highly Parallel Architectures for AI (UPenn) From: Lokendra Shastri on Wed 8 Jan 1986 at 15:44, 45 lines COURSE ANNOUNCEMENT CIS704 Highly parallel architectures for Artificial Intelligence PREREQUISITES: This is an advanced course in artificial intelligence. It will be assumed that the participants are familiar with basic issues in AI. DESCRIPTION: There is a growing interest in highly interconnected networks of very simple processing elements. These networks are referred to as Connectionist Networks and are playing an increasingly important role in artificial intelligence and cognitive science. This course is intended to discuss the motivation behind pursuing "connectionism" and to survey the state of current research in this area. We will review connectionist models of language understanding, parsing, knowledge representation, limited inference, and learning, and compare the connectionist approach to traditional AI approaches. TEXTS: None. A reading list will be provided. ASSIGNMENTS: Students will be expected to prepare a presentation of (or lead a discussion on) a paper on the reading list. There will be two or three assignments and a term paper. PLACE: TB 309. M, W 4:30-6:00 ------------------------------ Date: Mon, 6 Jan 86 20:33:58 MST From: keller@utah-cs.arpa (Bob Keller) Subject: Conference - 3rd Symposium on Logic Programming [Forwarded from the Prolog Digest by Laws@SRI-AI.] '86 SLP Call for Papers Third Symposium on Logic Programming Sponsored by the IEEE Computer Society September 21-25, 1986 Westin Hotel Utah Salt Lake City, UT The conference solicits papers on all areas of logic programming, including, but not confined to: Applications of logic programming Computer architectures for logic programming Databases and logic programming Logic programming and other language forms New language features Logic programming systems and implementation Parallel logic programming models Performance Theory Please submit full papers, indicating accomplishments of substance and novelty, and including appropriate citations of related work. The suggested page limit is 25 double-spaced pages. Send eight copies of your manuscript no later than 15 March 1986 to: Robert M. Keller SLP '86 Program Chairperson Department of Computer Science University of Utah Salt Lake City, UT 84112 Acceptances will be mailed by 30 April 1986. Camera-ready copy will be due by 30 June 1986. Conference Chairperson Exhibits Chairperson Gary Lindstrom, University of Utah Ross Overbeek, Argonne National Lab. Tutorials Chairperson Local Arrangements Chairperson George Luger, University of New Mexico Thomas C. Henderson, University of Utah Program Committee Francois Bancilhon, MCC William Kornfeld, Quintus Systems John Conery, University of Oregon Gary Lindstrom, University of Utah Al Despain, U.C. Berkeley George Luger, University of New Mexico Herve Gallaire, ECRC, Munich Rikio Onai, ICOT/NTT, Tokyo Seif Haridi, SICS, Sweden Ross Overbeek, Argonne National Lab. Lynette Hirschman, SDC, Paoli Mark Stickel, SRI International Peter Kogge, IBM, Owego Sten Ake Tarnlund, Uppsala University ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Sun Jan 12 17:02:24 1986 Date: Sun, 12 Jan 86 17:02:15 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a002464; 12 Jan 86 1:55 EST Date: Sat 11 Jan 1986 22:36-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #5 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sun, 12 Jan 86 04:45 EST AIList Digest Sunday, 12 Jan 1986 Volume 4 : Issue 5 Today's Topics: Conferences - Intelligent Systems Symposium & Workshop on AI for Generic Avionics & Uncertainty and AI Workshop & User-System Interfaces Workshop & IFIP Conference on Knowledge and Data & 2nd Expert Systems in Government, Re-Revised Version ---------------------------------------------------------------------- Date: 6 January 1986 1412-EST From: Peter Andrews@A.CS.CMU.EDU Subject: Conference - Intelligent Systems Symposium An International Symposium on Methodologies for Intelligent Systems (ISMIS'86) will be held October 23-25, 1986 in Knoxville, Tennessee. Papers are solicited in the following areas: (1) Expert Systems (2) Knowledge Representation (3) Learning and Adaptive Systems (4) Intelligent Databases (5) Approximate Reasoning (6) Logics for Artificial Intelligence Papers will be due on March 1, 1986, and papers which are accepted will be published in the proceedings of the symposium. A copy of the Call for Papers is posted on my office door (WEH 7216), and you can get a personal copy by sending a message to Zbigniew Ras (ras%tennessee.csnet@CSNET-RELAY.ARPA). ------------------------------ Date: Tue, 7 Jan 86 17:52:17 est From: Scott C McKay Subject: Conference - Workshop on AI for Generic Avionics AVIONICS LABORATORY WORKSHOP ON ARTIFICIAL INTELLIGENCE FOR GENERIC AVIONICS Georgia Tech Research Institute Atlanta, Georgia March 26-28, 1986 The Avionics Laboratory, located within the Air Force Wright Aeronautical Laboratories at Wright Patterson Air Force Base, is the primary organization responsible for planning and executing the Air Force basic research, exploratory and advanced development programs for aerospace avionics. A current major focus of that program is to explore the applicability of artificial intelligence to many functional avionics domains. The results have been very encouraging and we are convinced that AI will have significant future utility in aerospace vehicles. In order to plan for orderly, timely and expanded developments in AI, the Avionics Laboratory will be conducting a Workshop on Artificial Intelligence for Generic Avionics. The overall objective is to identify the "key basic research issues" that constrain the future expanded applicability of artificial intelligence technology to avionics applications and to outline what research should be pursued to remove the constraints. The workshop will be held at Georgia Tech Research Institute, Georgia Institute of Technology, Atlanta, Georgia on March 26-28, 1986. The workshop is planned to be an intensive 3-day work session involving 35-40 (maximum) of the best researchers in the field. Attendance will be by invitation only. If you feel you could contribute significantly to the objectives of the workshop and are interested in attending, please contact either of the following by 20 Jan 86: Lawrence E. Porter (513) 255-4415, AFWAL/GLXRA, Wright Patterson AFB, OH 45433-6543 or Michael Noviskey (513) 255-2713, same address. The mission of the Avionics Laboratory is broad and includes the primary areas of navigation, surveillance, reconnaissance, electromagnetic warfare, fire control, weapon delivery, communications, system architecture, information and signal processing and control, subsystem integration and supporting electronics, and software and electromagnetic device research and development. This mission spans the spectrum from basic research to advanced development. The emphasis of this workshop is being placed on the former. I encourage you to plan to attend the workshop and participate in a stimulating AI basic research exchange with your peers. I can assure you that the results will have a direct impact on future investment in AI basic research. Lawrence E. Porter Chairperson Artificial Intelligence Planning Avionics Laboratory ------------------------------ Date: Tue, 7 Jan 86 18:29:34 pst From: gluck@SU-PSYCH (Mark Gluck) Subject: Conference - Uncertainty and AI Workshop CALL FOR PARTICIPATION Second Workshop on: "Uncertainty in Artificial Intelligence" Philadelphia, PA. August 9-11, 1986 (preceeding AAAI conf.) Sponsored by: AAAI and RCA This workshop is a follow-up to the successful workshop in L.A., August 1985. Its subject is reasoning under uncertainty and representing uncertain information. The emphasis this year is on real applications, although papers on theory are also welcome. The workshop provides an opportunity for those interested in uncertainty in AI to present their ideas and participate in the discussions. Also panel discussions will provide a lively cross-section of views. Papers are invited on the following topics: *Applications--Descriptions of novel approaches; interesting results; important implementation difficulties; experimental comparison of alternatives etc. *Comparison and Evaluation of different uncertainty formalisms. *Induction (Theory discovery) under uncertainty. *Alternative uncertainty approaches. *Relationship between uncertainty and logic. *Uncertainty about uncertainty (Higher order approaches). *Other uncertainty in AI issues. Preference will be given to papers that have demonstrated their approach in real applications. Some papers may be accepted for publication but not presentation (except at a poster session). Four copies of the paper (or an extended abstract) should be sent to the arrangements chairman before 23rd. May 1986. Acceptances will be sent by the 20th. June and final (camera ready) papers must be received by 11th. July. Proceedings will be available at the workshop. General Chair: Program Chair: Arrangements Chair: John Lemmer Peter Cheeseman Lawrence Carnuccio KSC Inc. NASA-Ames Research Center RCA-Adv. Tech. Labs. 255 N. Washington St. Mail Stop 244-7 Mooretown Corp. Cntr. Rome, NY 13440 Moffett Field, CA 94035 Route 38, Mooretown, (315)336-0500 (415)694-6526 NJ 08057 (609)866-6428 Program Committee: P. Cheeseman, J. Lemmer, T. Levitt, J. Pearl, M. Yousry, L. Zadeh. ------------------------------ Date: Thu 9 Jan 86 15:12:41-CST From: CMP.LADAI@R20.UTEXAS.EDU Subject: Conference - User-System Interfaces Workshop USER-SYSTEM INTERFACES WORKSHOP When: January 31 - Febuary 1, 1986 Where: Austin South Plaza Hotel, Austin, Texas I-35 and Woodward What: A multidisciplinary conference addressing the problem of implementing effective communication between human and machine. The contributions of various fields such as Artificial Intelligence and Cognitive Psychology are considered. Participants: Brooks AFB SAM Burroughs IBM Lockheed MCC Rice University Southwest Research Institute Texas A&M University Texas Instruments University of Texas Registration: Mail to: Before Jan. 24 - $30.00 M. Sury Students - $15.00 Dept. T2-32, Bldg. 30E After Jan. 24 - $40.00 Lockheed Austin Division Students - $20.00 P.O. Box 17100 Includes lunch on Jan. 31. Austin, TX 78760 For additional info: Ron Grissell - [512]448-5154 Manda Sury - [512]448-5314 Diana Webster - [512]448-9186 ------------------------------ Date: Sun, 22 Dec 85 23:12:53 EST From: "John F. Sowa" Subject: Conference - IFIP Conference on Knowledge and Data IFIP INTERNATIONAL FEDERATION FOR INFORMATION PROCESSING ANNOUNCEMENT TC2 WORKING CONFERENCE organized by Working Group 2.6 Knowledge and Data (DS-2) November 3-7, 1986 in Albufeira (Algarve), Portugal Scope: Questions of meaning are more important for the design of a knowledge base than methods of encoding data in bits and bytes. As database designers add more semantic information to their systems, their conceptual schemata begin to look like AI systems of knowledge representation. In recognizing this convergence on issues of semantics, IFIP Working Group 2.6 is organizing a working conference on Knowledge and Data. It will address the issues and problems of knowledge representation from an interdisciplinary point of view. Topics: Design of a conceptual schema Knowledge and data modeling Database semantics Natural language semantics Expert database systems Logic, databases, and AI Methods of knowledge engineering Tools and aids for knowledge acquisition Invited speakers: Herve Gallaire, Germany Robert Meersman, Belgium J. Alan Robinson, USA Roger Schank, USA Dana Scott, USA An IFIP working conference is oriented towards detailed discussion of the topics presented. Participation is by invitation, with optional contribution of a paper that is refereed by the program committee. Anyone who is interested in participating should send an abstract of current research or a prospective paper to either of the program cochairmen. Abstracts are due March 14, 1986. Complete papers are due May 16, 1986. General Chairman: Amilcar Sernadas, Portugal Program cochairmen: John F. Sowa Robert Meersman IBM Systems Research Institute L.U.C. -- Dept. WNIF 500 Columbus Avenue Universitaire Campus Thornwood, NY 10594 B-3610 Diepenbeek U.S.A. Belgium CSNET: sowa.yktvmt@ibm ------------------------------ Date: 30 Dec 85 16:17:11 EST (Mon) From: Duke Briscoe Subject: Conference - 2nd Expert Systems in Government, Re-Revised Version This is yet another revision of the notice sent out several weeks ago, and is a revision of the revision sent out earlier today. I am sorry for the repetition, but there have been several foul-ups in the information being fed to me for the production of this announcement. CALL FOR PAPERS THE SECOND ANNUAL CONFERENCE ON EXPERT SYSTEMS IN GOVERNMENT Tyson's Westpark Hotel, McLean, VA in suburban Washington, D.C. October 20 - 24, 1986 The conference is sponsored by the IEEE Computer Society and the Mitre Corporation in cooperation with AIAA/NCS. The objective of the conference is to explore the following: - knowledge based applications and supporting technologies - implementation and impact of emerging application areas - future trends in available systems and required research Classified and unclassified papers which relate to the use of knowledge based systems are solicited. The topics of interest include, but are not limited to, the following applications: Professional: engineering, finance, law, management, medicine Office Automation: text understanding, intelligent DBMS, intelli- gent systems Command & Control: intelligence analysis, planning, targeting, communications, air traffic control, battle management Exploration: outer space, prospecting, archaeology Weapon Systems: adaptive control, electronic warfare, Star Wars, target identification Equipment: CAD/CAM, design monitoring, maintenance, repair Software: automatic programming, maintenance, verification and validation Architecture: distributed knowledge based systems, parallel com- puting Project Management: planning, scheduling, control Education: concept formation, tutoring, testing, diagnosis Imagery: photo interpretation, mapping Systems Engineering: requirements, preliminary design, critical design, testing, quality assurance Tools and Techniques: PROLOG, knowledge acquisition and represen- tation, uncertainty management Plant and Factory Automation Space Station Systems Human-Machine Interface Speech and Natural Language The program will consist of submitted and invited papers, which will provide an overview of selected areas. Contributed papers should be consistent with the following outline: 1. Introduction- state clearly the purpose of the work 2. Description of the actual work- must be new and significant 3. Results- discuss their significance 4. References Completed papers are to be no longer than 20 pages, including graphics. For classified papers, please submit a one page un- classified abstract. All classified papers must be releasable at the Secret level or below, and must be pre-approved by the author's cognizant security release authority. Papers to be presented by non-US citizens must be cleared through proper government to government channels. Four copies of the complete paper are to be submitted to: Dr. Kamal Karna, Conference Chairman IEEE Computer Society 1730 Massachusetts Ave., NW Washington, D.C. 20036-1903 Author's Schedule: Four copies of manuscript May 1, 1986 Acceptance letter June 15, 1986 Camera-ready copy July 15, 1986 Conference Chairman: Dr. Kamal Karna Washington AI Center Mitre Corporation Program Committee: Co-chairman: Classified Mr. Richard Martin Associate Director, Government Programs Software Engineering Institute Carnegie Mellon University Co-chairman: Unclassified Dr. Kamran Parsaye President Intelliware, Inc. ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Thu Jan 16 08:53:22 1986 Date: Thu, 16 Jan 86 08:53:18 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a012453; 15 Jan 86 22:19 EST Date: Wed 15 Jan 1986 10:03-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #6 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 16 Jan 86 05:26 EST AIList Digest Wednesday, 15 Jan 1986 Volume 4 : Issue 6 Today's Topics: Description - European Association for Theoretical Computer Science ---------------------------------------------------------------------- Date: 6 JAN 86 10:55-N From: ROZENBER%HLERUL5.BITNET@WISCVM.WISC.EDU Subject: Description - European Assoc. for Theoretical Computer Science [Forwarded from the SRI bboard by Laws@SRI-AI.] Dear collegue, I am taking advantage of this excellent communication medium, the "Theory Net", to send you information (actually the information leaflet) about the EUROPEAN ASSOCIATION FOR THEORETICAL COMPUTER SCIENCE (EATCS). Although our associa- tion is based in Europe, its membership is "intercontinental" - about 40% of our members comes from outside Europe. In our experience the only reason that a computer scientist who is either actively engaged or interested in theoretical computer science is not a member of EATCS is that she/he does not know about our organisation - just see how much we offer for so little!!! [...] If you have any questions do not hesitate to contact either myself (electronic address: ROZENBER@HLERUL5.BITNET) or the secretary of the association Th. Ottmann (electronic address: OTTMANN@GERMANY.CSNET). I take this opportunity to wish you the very best New Year. G. Rozenberg EATCS President =============================================================================== EUROPEAN ASSOCIATION FOR THEORETICAL COMPUTER SCIENCE (EATCS) COUNCIL OF EATCS BOARD President: G. Rozenberg, Leiden Vice President: W. Brauer, Munich Treasurer: J. Paredaens, Antwerp Secretary: Th. Ottmann, Karlsruhe Bulletin Editor: G. Rozenberg, Leiden TCS Editor: M. Nivat, Paris Past Presidents: M. Nivat, Paris (1972-1977) M. Paterson, Warwick (1977-1979) A. Salomaa, Turku (1979-1985) FURTHER COUNCIL MEMBERS G. Ausiello Rome J. De Bakker Amsterdam J. Diaz Barcelona F. Gecseg Szeged J. Gruska Bratislava Z. Manna Rehovot & Stanford H. Maurer Graz Ch.H. Papadimitriou Athens & Stanford A. Paz Haifa D. Perrin Paris E. Schmidt Aarhus D. Wood Waterloo EATCS HISTORY AND ORGANISATION EATCS is an international organisation founded in 1972. Its aim is to facilitate the exchange of ideas and results among theoretical computer scientists as well as to stimulate cooperation between the theoretical and the practical community in computer science. Its activities are coordinated by the Council of EATCS, out of which a President, a Vice President, a Treasurer and a Secretary are elected. Policy guidelines are determined by the Council and the General Assembly of EATCS. This assembly is scheduled to take place during the annual International Colloquium on Automata, Languages and Programming (ICALP), the conference of EATCS. MAJOR ACTIVITIES OF EATCS - Organization of ICALP's - Publication of the "Bulletin of the EATCS" - Publication of the "EATCS Monographs in Theoretical Computer Science" - Publication of the journal "Theoretical Computer Science" - Other activities of EATCS include the sponsorship of various more specialized meetings in theoretical computer science. Among such meetings are: CAAP (Colloquium on Trees in Algebra and Programming), TAPSOFT (Conference on Theory and Practice of Software Development), STACS (Symposium on Theoretical Aspects of Computer Science), Workshop on Graph Theoretic Concepts in Computer Science, European Workshop on Applications and Theory of Petri Nets, Workshop on Graph Grammars and their Applications in Computer Science. BENEFITS Benefits offered by EATCS include: - Receiving the "Bulletin of the EATCS" (about 600 pages per year) - Reduced registration fees at various conferences - Reciprocity agreements with other organisations - 25% discount in purchasing ICALP proceedings - 25% discount in purchasing books from "EATCS Monographs on Theoretical Computer Science" - About 70% (equals about 1000 Dutch guilders) discount per annual subscription to "Theoretical Computer Science". (1) THE ICALP CONFERENCE ICALP is an international conference covering all aspects of theoretical computer science and now customarily taking place during the third week of July. Typical topics discussed during recent ICALP conferences are: computability, automata theory, formal language theory, analysis of algorithms, computa- tional complexity, mathematical aspects of programming language definition, logic and semantics of programming languages, foundations of logic programming, theorem proving, software specification, computational geometry, data types and data structures, theory of data bases and knowledge based systems, cryptography, VLSI structures, parallel and distributed computing, models of concurrency and robotics. Sites of ICALP meetings: - Paris, France (1972) - Haifa, Israel (1981) - Saarbrucken, Germany (1974) - Aarhus, Denmark (1982) - Edinburgh, Great Britain (1976) - Barcelona, Spain (1983) - Turku, Finland (1977) - Antwerp, Belgium (1984) - Udine, Italy (1978) - Nafplion, Greece (1985) - Graz, Austria (1979) - Rennes, France (1986) - Noordwijkerhout, Holland (1980) - Karlsruhe, Germany (1987) (2) THE BULLETIN OF THE EATCS Three issues of the Bulletin are published annually appearing in February, June and October respectively. The Bulletin is a medium for rapid publication and wide distribution of material such as: - EATCS matters - Information about the current ICALP - Technical contributions - Surveys and tutorials - Reports on conferences - Calendar of events - Reports on computer science departments and institutes - Listings of technical reports and publications - Book reviews - Open problems and solutions - Abstracts of Ph.D. Theses - Information on visitors at various institutions - Entertaining contributions and pictures related to computer science. Contributions to any of the above areas are solicited. All written contributions should be sent to the Bulletin Editor: Prof.dr. G. Rozenberg Dept. of Mathematics and Computer Science University of Leiden P.O. Box 9512 2300 RA Leiden, The Netherlands Deadlines for submissions to reach the Bulletin Editor are: January 15, May 15 and September 15 for the February, June and October issue respec- tively. All pictures (preferably black and white) including text of what they are showing should be sent to the Picture Editor: Dr. P. van Emde-Boas University of Amsterdam Roeterstraat 15 1018 WB Amsterdam, The Netherlands Deadlines are 2 weeks before those for written contributions, indicated above. (3) EATCS MONOGRAPHS ON THEORETICAL COMPUTER SCIENCE This is a series of monographs published by Springer-Verlag and launched during ICALP 1984; within the first year six volumes appeared. The series includes monographs as well as innovative textbooks in all areas of theo- retical computer science, such as the areas listed above in connection with the ICALP conference. The volumes are hard-cover and ordinarily produced by type-setting. To ensure attractive prices other production methods are possible. The editors of the series are W. Brauer (Munich), G. Rozenberg (Leiden), and A. Salomaa (Turku). Potential authors should contact one of the editors. The advisory board consists of G. Ausiello (Rome), S. Even (Haifa), M. Nivat (Paris), C. Papadimitriou (Athens & Stanford), A. Rosenberg (Durham), and D. Scott (Pittsburgh). Updated information about the series can be obtained from the publisher, Springer-Verlag. (4) THEORETICAL COMPUTER SCIENCE The aim of the "Theoretical Computer Science" journal is to publish papers in the fast envolving field of theoretical computer science. The volume of research on theoretical aspects of computer science has increased enormously in the past. The classical theories of automata and formal languages still offer problems and results, while considerable attention is now being given to newer areas, such as the formal semantics of programming languages and the study of algorithms and their complexity. Behind all this lie the major problems of under- standing the nature of computation and its relation to computing methodology. While "Theoretical Computer Science" remains mathematical and abstract in spirit, it derives its motivation from the problems of practical computation. The editors intend that the domain covered by "Theoretical Computer Science" will increase and evolve with the growth of the science itself. The editor-in-chief of "Theoretical Computer Science" is: Prof. M. Nivat 162, Boulevard Malesherbes 75017 Paris, France. ADDITIONAL INFORMATION Please contact the Secretary of EATCS: Prof.dr. Th. Ottmann Institut fur Angewandte Informatik und Formale Beschreibungsverfahren Universitat Karlsruhe Postfach 6380 D-7500 Karlsruhe 1 West Germany DUES The dues are US $ 10.- for a period of one year. If the initial membership payment is received in the period December 21 - April 20, April 21 - August 20, August 21 - December 20, then the first membership year will start on June 1, October 1, February 1, respectively. Every continuation payment continues the membership for the same time period. An additional fee is required for ensuring the air mail delivery of the EATCS Bulletin outside Europe. The amounts are $ 7.- for USA, Canada, Israel, $ 10.- for Japan and $ 12.- for Australia per year. For information additonal fees for other destinations contact either the Secretary or the Treasurer. HOW TO JOIN EATCS To join send the annual dues, or a multiple thereof (to cover a number of years), to the Treasurer of EATCS: Prof.dr. J. Paredaens University of Antwerp, U.I.A. Department of Mathematics Universiteitsplein 1 B-2610 Wilrijk, Belgium The dues can be paid (in order of preference) by US $ bank cheques, other currency bank cheques, US $ cash, other currency cash. It cannot be paid by International Post Money Order. When submitting payment, please make sure to indicate complete name and address. For this purpose you may want to use the form below. You may also pay the membership fee via the following account: General Bank Antwerp Antwerp, Belgium Account number: 220-0596350-30 If a transfer is in US $ then the annual membership payment equals US $ 10.-. If a transfer (covering the membership for any number of years and/or addtitional air mail delivery for any number of years) is in a currency other than US $, then additional US $ 2.- for the transfer must be paid (the difference is used to cover the bank charges). Please remember to indicate your address clearly (since the Bulletin is send to the address you give). ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Thu Jan 16 08:53:15 1986 Date: Thu, 16 Jan 86 08:53:12 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a012832; 15 Jan 86 23:06 EST Date: Wed 15 Jan 1986 10:06-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #7 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 16 Jan 86 05:29 EST AIList Digest Wednesday, 15 Jan 1986 Volume 4 : Issue 7 Today's Topics: Seminars - Reasoning About Hard Objects (BBN) & LOGIN: A Logic Programming Language with Inheritance (MIT) & Temporal Reasoning and Default Logics (SU) & LISP/Prolog Memory Performance (Ames) ---------------------------------------------------------------------- Date: 9 Dec 1985 12:10-EST From: Brad Goodman Subject: Seminar - Reasoning About Hard Objects (BBN) [Forwarded from the MIT bboard by SASW@MIT-MC.] BBN Laboratories Science Development Program AI Seminars Speaker: Ernest Davis NYU Title: Issues in Reasoning about Hard Objects Date: Monday, December 16th, 10:30a.m. Place: BBN Labs, 10 Moulton Street, 3rd floor large conference room Abstract The physics of rigid solid objects raises two serious problems which have not been addressed in previous spatial and physical reasoning programs. Firstly, the physical properties of solid objects are sensitive to very slight variations in shapes. Therefore, when an ideal shape is used to approximate a real shape, the accuracy of the approximation must be tightly bounded. Secondly, the method of reasoning used by both Forbus and DeKleer of going from one critical point to the next is not, in general, appropriate. Frequently, as in reasoning about a ball going down a funnel, one is interested only in the final outcome (the ball goes out the funnel) and not in any of the intermediate critical points (collisions between the ball and the funnel). However, it is difficult to state axioms that assert global relationships of this sort in a way that allows them to be used in cases where additional objects enter the picture. ------------------------------ Date: Thu 9 Jan 86 13:37:32-EST From: Susan Hardy Subject: Seminar - LOGIN: A Logic Programming Language with Inheritance (MIT) [Forwarded from the MIT bboard by SASW@MC.LCS.MIT.EDU.] DATE: Thursday, January 16, 1986 TIME: 3:00 p.m. - Refreshments 3:15 p.m. - Lecture PLACE: NE43-512A LOGIN: A LOGIC PROGRAMMING LANGUAGE WITH BUILT-IN INHERITANCE Hassan Ait-Kaci A.I. Program MCC, Austin, Texas Since the early days of research in Automated Deduction, inheritance has been proposed as a means to capture a special kind of information; viz., taxonomic information. For example, when it is asserted that "whales are mammals", we understand that whatever properties mammals possess should also hold for whales. Naturally, this meaning of inheritance can be well captured in logic by the semantics of logical implication. However, this is not operationally satisfactory. Indeed, in a first-order logic deduction system realizing inheritance as implication, inheritance from "mammal" to "whale" is achieved by an inference step. But this special kind of information somehow does not seem to be meant as a deduction step---thus lengthening proofs. Rather, its purpose seems to be to accelerate, or focus, a deduction process---thus shortening proofs. In this talk, I shall argue that the syntax and operational interpretation of first-order terms can be extended to accommodate for taxonomic ordering relations between constructor symbols. As a result, I shall propose a simple and efficient paradigm of unification which allows the separation of (multiple) inheritance from the logical inference machinery of Prolog. This yields more efficient computations and enhanced language expressiveness. The language thus obtained, called LOGIN, subsumes Prolog, in the sense that conventional Prolog programs are equally well executed by LOGIN. I shall start with motivational examples, introducing the flavor of what I believe to be a more expressive and efficient way of using taxonomic information, as opposed to straight Prolog. Then, I shall give a quick formal summary of how first-order terms may be extended to embody taxonomic information as record-like type structures, together with an efficient type unification algorithm. This will lead to a technical proposal for integrating this notion of terms into the SLD-resolution mechanism of Prolog. With examples, I shall illustrate a LOGIN interpreter. Host: Rishiyur Nikhil (617)253-0237 Nikhil@mit-xx.arpa ------------------------------ Date: 13 Jan 86 1659 PST From: Vladimir Lifschitz Subject: Seminar - Temporal Reasoning and Default Logics (SU) Next nonmotonic reasoning meeting: A Review and Critique of: "Temporal Reasoning and Default Logics" by Steve Hanks and Drew McDermott Yale/CSD/RR #430 October 1985 by Benjamin Grosof, inquisitioner Thursday, January 16, 4pm MJH 252 Hanks and McDermott in their recent Yale Tech Report pose an example problem in temporal reasoning and claim that none of the leading formalisms for default reasoning (namely Reiter's Default Logic, McDermott and Doyle's modal Non-Monotonic Logic, and Circumscription) adequately capture the type of non-monotonic reasoning that is (what they claim is) desirable in the example. They give an algorithm which does. They go on to conclude rather pessimistically that there seems to be some inherent problem in the semantics of all three default formalisms. In this talk, I review their paper, including their temporal logic. I argue that their example in particular is interesting and suggestive, but that the semantical difficulty that they emphasize arises from an underspecification of the problem. I will go on to suggest how indeed to represent the additional CRITERION satisfied by their algorithm (but not by their formulations in default formalisms). I show how Vladimir's new circumscription presented in our fall sessions of the non-monotonic reasoning seminar can solve the representational problem they pose. I argue that circumscription, because it can incorporate certain kinds of preferences among competing extensions via prioritization, has an advantage over the other two default formalisms, and promises to be able to represent the CRITERION more generally than their algorithm does. I also discuss how their temporal formalism occupies an intermediate place between STRIPS and situation calculus. ------------------------------ Date: Tue, 14 Jan 86 21:47:09 pst From: eugene@AMES-NAS.ARPA (Eugene Miya) Subject: Seminar - LISP/Prolog Memory Performance (Ames) National Aeronautics and Space Administration Ames Research Center SEMINAR ANNOUNCEMENT Joint Ames AI Forum/RCR Branch SPEAKER: Evan Tick Computer Systems Laboratory Stanford University TOPIC: Memory Performance of Lisp and Prolog Programs ABSTRACT: This talk presents a comparison between Lisp and Prolog architectures based on memory performance. A subset of the Gabriel benchmarks was translated into Prolog, compiled into the Warren Abstract Machine instruction set and emulated. The programs were also measured with an instrumented Common Lisp targeted to a Series 9000/HP237. Memory usage statistics indicate how the two langauges do fundamental computations different ways with varying efficiency. DATE: 28 January 1986 TIME: 1030 AM BLDG: 172 ROOM: 233 Tuesday POINT OF CONTACT: E. Miya PHONE NUMBER: (415)-694-6453 emiya@ames-vmsb I am current attending a conference, please send mail or contact my office mate. VISITORS ARE WELCOME: Register and obtain vehicle pass at Ames Visitor Reception Building (N-253) or the Security Station near Gate 18. See map below. Do not use the Navy Main Gate. Non-citizens (except Permanent Residents) must have prior approval from the Director's Office one week in advance. Submit requests to the point of contact indicated above. Non-citizens must register at the Visitor Reception Building. Permanent Residents are required to show Alien Registration Card at the time of registration. ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Jan 16 23:16:52 1986 Date: Thu, 16 Jan 86 23:16:37 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a011104; 15 Jan 86 19:01 EST Date: Wed 15 Jan 1986 10:12-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #8 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 16 Jan 86 22:53 EST AIList Digest Wednesday, 15 Jan 1986 Volume 4 : Issue 8 Today's Topics: Queries - Macsyma & Symbolics Prolog & Speech Learning Machine, Definition - Paradigm, Intelligence - Computer IQ Tests, AI Tools & Applications - Expert Systems and Computer Graphics & Common Lisp for Xerox & Real-Time Process Control ---------------------------------------------------------------------- Date: 13 Jan 86 23:16 GMT From: dkb-amos @ HAWAII-EMH.ARPA Subject: Macsyma I would appreciate any help that could be supplied in locating a source for Macsyma. I'm looking for a version that will run under Franzlisp Opus 38.91. We do contract work for the Air Force but I have no immediate contract application for this package, I would just like to get familier with it and have it around for possible future applications. Thanks. -- Dennis Biringer ------------------------------ Date: Mon 13 Jan 86 16:07:10-PST From: Luis Jenkins Subject: Symbolics Prolog [Sorry if this topic has been beaten to death before many time ...] Here at Schlumberger Palo Alto Research (SPAR) we have been working for some time on large Prolog programs for Hardware Verification, first in Dec-20 Prolog and then in Quintus Prolog for Suns. Recently we have been interested in the possibility of using Symbolics Prolog for further R&D work, as the lab has a bunch of LispMs. Does anyone out there has first-hand (or n-hand, please specify) experience with the Prolog that Symbolics offers. Specifically, we want to hear praises/complaints about :- o DEC-10/Quintus Compatibility o Speed o Bugs o Extensions o Interface with the LispM environment o Mixing Prolog & Lisp code o Random User Comments Thanks, Luis Jenkins Schlumberger Palo Alto Research lej@sri-kl ...decwrl!spar!lej ------------------------------ Date: 13 Jan 86 11:22:01 EST From: kyle.wbst@Xerox.ARPA Subject: Johns Hopkins Learning Machine Does anyone have any more info on the following: I caught the tail end of a news item on the NBC Today Show this morning about someone at Johns Hopkins who has built a "Networked" computer consisting of 300 "elements" that has a speech synthesizer attached to it. The investigator claims that the thing learns to speak English the same way a human baby does. They played a tape recording which represented a condensation of several hours of "learning" by the device. The investigator claims he does not know how the the thing works. I didn't catch his name. Who is this person and what is the system configuration of the machine (which seemed to fit into one large rack of equipment). Earle Kyle ------------------------------ Date: Tue, 14 Jan 86 09:34:53 est From: Walter Hamscher Subject: Today Show Segment A friend of mine saw the Today Show this Monday morning, and said there was a particularly breathless segment that left the impression that somebody has solved `the AI problem'. It seems to have been a rather vague story about someone at Johns Hopkins who has built some sort of massively parallel machine that learns language. Sorry the details are so sketchy. Did anybody else see this segment or know the story behind the story? ------------------------------ Date: 14 Jan 86 22:05:47 EST From: Mike Tanner @ Ohio State Subject: Paradigm I've seen some discussion of paradigm in recent AILists and since I just audited a grad course in philosophy of science where we read Kuhn I thought I'd summarize what I remember of Kuhn's notion of paradigm. (Auditing a course certainly does not make me an expert, but it does mean that I've read Kuhn recently and carefully.) Several people have pointed out that the dictionary definition (e.g., Webster's 3rd New International) of `paradigm' is `example', `pattern', or `model'. But they further claim that this is not what Kuhn meant. However, I think that the way `paradigm' is used by Kuhn is (most of the time) perfectly compatible with the dictionary. In _The_Structure_of_Scientific_Revolutions_ Kuhn normally uses `paradigm' to mean `example of theory applied' or `example of how to do science'. (Sometimes he uses it to mean `theory', which is confusing and I think he later admits that it is just sloppiness on his part.) Ron Laymon, our prof in the philosophy of science course, suggested that it might be best to think of paradigm as `an uninterpreted book'. Everybody working in some field points to a book when asked what they do and says, "There, read that book and you'll know." Of course, once the book is opened there's likely to be a lot of disagreement about what it means. Another important characteristic of paradigms is that they suggest a lot of further research. If I were a cynical person I would say that the success of a paradigm depends on people's perceptions of funding prospects for research of the sort that it defines. I'm not sure that AI is mature enough to rate any paradigms. But I think that a case could be made for some things as "mini-paradigms", such as GPS, MYCIN, Minsky's frame paper, etc. That is, they defined some sub-discipline within AI where a lot of people did, and are doing, fruitful work. (I don't mean "mini" to be pejorative. I just think that a paradigm has to be a candidate for unifying research in the field, or maybe even defining the field, and these probably don't qualify. But then, I might be expecting too much of paradigms.) -- mike ARPA: tanner@Rutgers CSNet: tanner@Ohio-State Physically, I am at Ohio State but I have a virtual existence at Rutgers and can receive mail either place. ------------------------------ Date: Wed 15 Jan 86 09:42:58-CST From: David Throop Subject: Computers & IQ Tests There have been recent inquiries about how well computer programs can do on IQ tests. An article in the journal _Telicom_ (1) mentions a computer program for taking IQ tests. It seems to be aimed entirely at the kinds of math puzzles that fill in missing numbers in series. "The program (is) called HIQ-SOLVER 160 ... BASIC, less than 10 Kbytes... in July/August Dutch computer magazine _Sinclair_Gebruiker_ has the listing... The program has been tried on the numerical test in Hans Eysenck's _Check_Your_Own_IQ_ and it solved 36 out of 50 problems, corresponding with an IQ of about 160 (hence its name); as some items in the Eysenck test were of a type that had not been implemented one might argue that the program's raw score corresponds with an even higher IQ ..." He goes on to give the algorithm. I think this example highlights an example of the difficulty of applying human IQ tests to machines - the program scores very high on certain IQ tests because it does a very limited kind of pattern recognition very well. But it is completely brittle - it's helpless to recognize patterns that are only slightly off what it expects. Human intelligence tests do not measure human intelligence directly. They measure characteristics associated with intelligence. The underlying assumption is that this association is good enough that it will predict how well humans will do on tasks that cannot be given as standard tests, but evince intelligence. This is a dubious proposition for humans, but it breaks down completely on machines. Nonetheless, it shouldn't be too hard to CONS up some programs that do terribly well on some not too terribly well designed IQ tests. (1)Feenstra, Marcel "Numerical IQ - Tests and Intelligence" Telicom, Aug 85, Bx 141 San Francisco 94101 ------------------------------ Date: Sun 12 Jan 86 18:05:49-PST From: Ken Laws Subject: Expert Systems and Computer Graphics IEEE Computer Graphics and Applications, December 1985, pp. 58-59, has a review by Ware Myers of the 6th Eurographics conference. The key theme was integrating expert systems and computer graphics. Several of the papers discussed binding Prolog and the GKS graphical kernel standard. ------------------------------ Date: Sun 12 Jan 86 17:28:34-PST From: Ken Laws Subject: Common Lisp for Xerox Expert Systems, Vol. 2, No. 4, October 1985, p. 252, reports that Xerox will be implementing Common Lisp on its Lisp workstations. The first copies may be available in the second quarter of 1986. Xerox will continue to support Interlisp-D, and will be adding extensions and compatable features to both languages. A package for converting Interlisp-D programs to Common Lisp is being developed. Guy Steele said (Common Lisp, p. 3) that it is expected that user- level packages such as InterLisp would be built on top of the Common Lisp core. Perhaps that is now happening. Xerox is also offering CommonLoops as a proposed standard for object-oriented programming. ------------------------------ Date: Sun 12 Jan 86 18:00:24-PST From: Ken Laws Subject: Real-Time Process Control IEEE Spectrum, January 1986, p. 64, reports the following: The building of engineering expertise into single-loop controllers is beginning to bear fruit in the form of a self-tuning process controller. The Foxboro Co. in Foxboro, Mass., included self-tuning features in its Model 760 single-loop controller as well as in three other controller-based products. Common PID (proportional, integral, and derivative) controllers made by Foxboro now have a built-in microprocessor with some 200 production rules; the loop-tuning rules have evolved over the last 40 years both at Foxboro and elsewhere. The Foxboro self-tuning method is a pattern recognition approach that allows the user to specify desirable temporal response to disturbances in the controlled parameter or in the controlled set point. The controller then observes the actual shape of these disturbances and adjusts its PID values to restore the desirable response. Asea also makes a self-tuning controller, Novatune, but the current version requires substantial knowledge of stochastic control theory to install. Lisp Machine Inc. has now installed PICON, its expert system for real-time process control, at about a half-dozen sites. It has also announced support for GM's MAP communication protocol for factory automation. ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Tue Jan 21 11:05:31 1986 Date: Tue, 21 Jan 86 11:05:28 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a009705; 20 Jan 86 17:10 EST Date: Mon 20 Jan 1986 13:26-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #9 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 21 Jan 86 05:29 EST AIList Digest Monday, 20 Jan 1986 Volume 4 : Issue 9 Today's Topics: Queries - System V Franz & OPS5 & Address for Prof. Bouille & Knowledge-Engineering Software & Supercomputers and AI & AI and Process Control & What is a Symbol? ---------------------------------------------------------------------- Date: Wed, 15 Jan 1986 18:50 PLT From: George Cross Subject: System V Franz? Does anyone sell or distribute a version of FranzLisp that runs under Unix System V on a VAX? or another machine? ---- George - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - George R. Cross cross@wsu.CSNET Computer Science Department cross%wsu@csnet-relay.ARPA Washington State University faccross@wsuvm1.BITNET Pullman, WA 99164-1210 (509)-335-6319/6636 Acknowledge-To: George Cross ------------------------------ Date: Thu 16 Jan 86 10:42:00-PST From: Ted Markowitz Subject: OPS5 query I'd like to try a version of OPS5 on an IBM-PC for exploration (not necessarily system delivery) and would like some opinions of the various flavors I've seen advertised. A few I've noticed are TOPSI and OPS83. Any thoughts on price, speed, portability, etc. would be welcome. I can digest the responses and post them back to the list. Thanx muchly. --ted ------------------------------ Date: Fri, 17 Jan 86 11:06 IST From: Amir Toister Subject: help CAN ANYONE HELP ME LOCATE: PROF. F. BOUILLE LABORATOIRE D'INFORMATIQUE DES SCIENCE DE LA TERRE, UNIV. PIERRE ET MARIE CURIE. PARIS ------------------------------ Date: Fri, 17 Jan 86 15:04:24 est From: Tom Scott Subject: Two questions on knowledge-engineering software 1. Rick Dukes from Symbolics recently gave an interesting talk on AI/KE to the Northwest Ohio chapter of the ACM. He mentioned an expert-system-building tool, MRS, from Stanford. I ran across another reference to MRS in the Winter 1986 issue of "AI Magazine" (p. 107). Can anyone tell me about the system? What does it do? What representation and search techniques are available through it? Can it handle frames? Semantic networks? Certainty factors? How does it work as an expert-system development environment? Most importantly, how does a university acquire MRS? I think Rick told us that it was available to universities essentially for free. If that is true, then where can we send a tape? 2. Several good works have been published on Prolog, e.g., Clocksin & Mellish's "Programming in Prolog" and Lloyd's "Foundations of Logic Programming". It appears, however, that there is no book yet on "advanced" AI/KE programming techniques in Prolog. The Clocksin & Mellish text is good as an introduction, the Lloyd book as a theoretical discussion of logical foundations. A number of us would like to see a Prolog book that covers topics similar in scope to part II of Charniak, Riesbeck, and McDermott's "Artificial Intelligence Programming". Charniak et al. use Lisp; who does the same with Prolog? One hope along these lines is an MIT Press book, "The Art of Prolog" by Sterling and Shapiro. I first saw a reference to it in an advertisement on p. A-22 of "Communications of the ACM" (January 1986). Has the book been published yet or is it not supposed to come out until May? Does anyone know about it? What does it cover? ------------------------------ Date: Mon, 20 Jan 86 09:47:20 cet From: JOHND%IDUI1.BITNET@WISCVM.WISC.EDU Subject: Supercomputers and AI I would like to know if anyone has any references to AI projects being done on supercomputers. We have a class here on supercomputers that will be using a Cray XMP/24, an Intel Hypercube, and perhaps an MPP. I am interested in having a student do an AI related project, and I'd like it to relate to some current work. I am also interested in how much AI software (languages and systems) has been transported to these supercomputer. All references will be most appreciated. John Dickinson Univ. of Idaho JOHND%IDUI1 (on BITNET) ------------------------------ Date: Mon, 20 Jan 86 9:21:58 MET From: mcvax!delphi.UUCP!mdc@seismo.CSS.GOV Subject: AI and process control I am involved in a AI factory automation project. Can you give me any reference or material on this subject? Thanks Maurizio De Cecco DELPHI S.p.A. Via Della Vetraia, 11 55049 Viareggio Italy [Two magazine articles are Expert Systems, Vol. 1, No. 1, July 1984, and High Technology, May 1985. The first is a description of the CMU ISIS scheduling system, the latter a report on factory automation. -- KIL] ------------------------------ Date: 19 Jan 86 17:12:15 EST From: David.Plaut@K.CS.CMU.EDU Subject: What is a symbol? This is a request for help.... The idea of a symbol is found throughout AI and Cognitive Science, and seems to bear considerable theoretical weight. Newell and Simon's Physical Symbol System Hypothesis, that a machine that carries out processes operating on symbol structures has the necessary and sufficient means for general intelligent action, seems to be an expression of the underlying assumptions of the majority of work in AI. Yet it seems that no satisfactory definition/description (necessary and sufficient characteristics) of what is meant by a symbol (sorry about the pun) has ever been presented. The following rough description seems to be a standard attempt: A symbol is a formal entity whose internal structure places no restrictions on what it may represent in the domain of interest. Unfortunately, when combined with the Physical Symbol System Hypothesis, this notion of symbol creates a problem with regard to so-called "connectionist" systems. It is possible to design a connectionist system that exhibits, if not "general intelligent action", certainly "knowledge-level" behavior, without any processes operating on symbol structures. The formal, computational processes of the system are operating below the symbol level, in terms of the interaction of units representing non-symbolic "micro-features". A symbol level description of the system only applies to emergent patterns of micro-features. Unfortunately these patterns fail to qualify as symbols by the above account due to the fact that it is precisely their internal structure which determines what they represent. Thus we are left with a system capable of knowledge-level behavior apparently without symbols. It seems there are three ways out of this dilemma: (1) deny that connectionist systems are capable, in principle, of "true" general intelligent action; (2) reject the Physical Symbol System Hypothesis; or (3) refine our notion of a symbol to encompass the operation and behavior of connectionist systems. (1) seems difficult (but I suppose not impossible) to argue for, and since I don't think AI is quite ready to agree to (2), I'm hoping for help with (3) - Any suggestions? David Plaut (dcp@k.cs.cmu.edu) ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Tue Jan 21 11:06:33 1986 Date: Tue, 21 Jan 86 11:06:30 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a009493; 20 Jan 86 16:57 EST Date: Mon 20 Jan 1986 13:36-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #10 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 21 Jan 86 05:27 EST AIList Digest Monday, 20 Jan 1986 Volume 4 : Issue 10 Today's Topics: Machine Learning - Connectionist Speech Machine ---------------------------------------------------------------------- Date: Wed, 15 Jan 86 23:06 EST From: Tim Finin Subject: nettalk Several people inquired about the work of Terrence Sejnowski (of Johns Hopkins) which was reported on the Today show recently. This abstract is to a talk given by Sejnowski here at Penn in October '85: NETTALK: TEACHING A MASSIVELY-PARALLEL NETWORK TO TALK TERRENCE J. SEJNOWSKI BIOPHYSICS DEPARTMENT JOHNS HOPKINS UNIVERSITY BALTIMORE, MARYLAND Text to speech is a difficult problem for rule-based systems because English pronunciation is highly context dependent and there are many exceptions to phonological rules. A more suitable knowledge representation for correspondences between letters and phonemes will be described in which rules and exceptions are treated uniformly and can be determined with a learning algorithm. The architecture is a layered network of several hundred simple processing units with several thousand weights on the connections between the units. The training corpus is continuous informal speech transcribed from tape recordings. Following training on 1000 words from this corpus the network can generalize to novel text. Even though this network was not designed to mimic human learning, the development of the network in some respects resembles the early stages in human language acquisition. It is conjectured that the parallel architecture and learning algorithm will also be effective on other problems which depend on evidential reasoning from previous experience. (No - I don't have his net address. Tim.) ------------------------------ Date: 16 Jan 86 1225 PST From: Richard Vistnes Subject: John Hopkins learning machine: info See AIList Digest V3 #183 (10 Dec 1985) for a talk given at Stanford a little while ago that sounds very similar. The person is: Terrence J. Sejnowski Biophysics Department Johns Hopkins University Baltimore, MD 21218 (I didn't attend the talk). -Richard Vistnes ------------------------------ Date: Sun, 19 Jan 86 0:19:10 EST From: Terry Sejnowski Subject: Reply to Inquiries NBC ran a short segment last Monday, January 13, on the Today Show about my research on a connectionist model of text-to-speech. The segment was meant for a general audience (waking up) and all the details were left out, so here is an abstract for those who have asked for more information. A technical report is available (Johns Hopkins Electrical Engineering and Computer Science Technical Report EECS-8601) upon request. NETtalk: A Parallel Network that Learns to Read Aloud Terrence Sejnowski Department of Biophysics Johns Hopkins University Baltimore, MD 21218 Charles R. Rosenberg Department of Psychology Princeton Unviversity Princeton, NJ 08540 Unrestricted English text can be converted to speech by applying phonological rules and handling exceptions with a look-up table. However, this approach is highly labor intensive since each entry and rule must be hand-crafted. NETtalk is an alternative approach that is based on an automated learning procedure for a parallel network of deterministic processing units. After training on a corpus of informal continuous speech, it achieves good performance and generalizes to novel words. The distributed representations discovered by the network are damage resistant and recovery from damage is about ten times faster than the original learning starting from the same level of performance. Terry Sejnowski ------------------------------ Date: Thu, 16 Jan 86 12:53 EST From: Mark Beutnagel Subject: speech learning machine The speech learning machine referred to in a recent AIList is almost certainly a connection machine built by Terry Sejnowski. The system consists of a maybe 200 processing elements (or simulations of such) and weighted connections between them. Input is a small window of text (5 letters?) and output is phonemes. The system learns (i.e. modifies weights) based on a comparison of the predicted phoneme with the "correct" phoneme. After running overnight the output was recognizable speech--good but still slightly mechanical. Neat stuff but nothing mystical. -- Mark Beutnagel (The above is my recollection of Terry's talk here at UPenn last fall so don't quote me.) ------------------------------ Date: Sun 19 Jan 86 12:31:31-PST From: Ken Laws Subject: Speech Learning I'll have a try at summarizing Terry's talk at Stanford/CSLI: The speech learning machine is a three-layer "perceptron-like" network. The bottom layer of 189 "processing units" simply encodes a 7-character window of input text: each character (or space) activates one of 27 output lines and suppresses 26 other lines. The top, or output, layer represents a "coarse coding" of the phoneme (or silence) which should be output for the character at the center of the 7-character window. Each bit, or output line, of the top layer represents some phoneme characteristic: vowel/consonant, voiced, fricative, etc. Each legal phoneme is thus represented by a particular output pattern, but some output patterns might not correspond to legal phonemes. (I think they were mapped to silence in the recording.) The output was used for two purposes: to compute a feedback error signal used in training the machine, and to feed the output stage of a DecTalk speech synthesizer so that the output could be judged subjectively. The heart of the system is a "hidden layer" of about 200 processing units, together with several thousand interconnections and their weights. These connect the 189 first-level outputs to the small number of output processing units. It is the setting of the weight coefficients for this network that is the central problem. Input to the system was a page of a child's speech that had be transcribed in phonetic notation by a professional. Correspondence had been established between each input letter and the corresponding phoneme (or silence), and the coarse coding of the phonemes was known. For any possible output of the machine it was thus possible to determine which bits were correct and which were incorrect. This provided the error signal. Unlike the Boltzmann Machine or the Hopfield networks, Sejnowski's algorithm does not require symmetric excitory/inhibitory connections between the processing units -- the output computation is strictly feed-forward. Neither did this project require simulated annealing, although some form of stochastic training or of "inverse training" on wrong inputs might be helpful in avoiding local minima in the weight space. What makes this algorithm work, and what makes it different from multilayer perceptrons, is that the processing nodes do not perform a threshold binarization. Instead, the output of each unit is a sigmoid function of the weighted sum of its inputs. The sigmoid function, an inverse exponential, is essentially the same one used in the Boltzmann Machine's stochastic annealing; it also resembles the response curve of neurons. Its advantage over a threshold function is that it is differentiable. This permits the error signal to be propagated back through each processing unit so that appropriate "blame" can be attributed to each of the hidden units and to each of the connections feeding the hidden units. The back-propagated error signals are exactly the partial derivatives needed for steepest-descent optimization of the network. Subjective results: The output of the system for the page of text was originally just a few random phonemes with no information content. After sufficient training on the correct outputs the machine learned to "babble" with alternating vowels or vowel/consonants. After further training it discovered word divisions and then began to be intelligible. It could eventually read the page quite well, with a distinctly childish accent but with mechanical pacing of the phonemes. It was then presented with a second page of text and was able to read that quite well also. I have seen some papers by Sejnowski, Kienker, Hinton, Schumacher, Rumelhart, and Williams exploring variations of this machine learning architecture. Most of the work has concerned very simple, but difficult, problems, such as learning to compute exclusive OR or the sum of two two-bit numbers. More complex tasks involved detecting symmetries in binary matrices and computing figure/ground (or segmentation) relationships in noisy images with an associated focus of attention. I find the work promising and even exciting. -- Ken Laws ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Jan 23 02:30:59 1986 Date: Thu, 23 Jan 86 02:30:55 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a001031; 22 Jan 86 13:38 EST Date: Wed 22 Jan 1986 10:07-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #11 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 23 Jan 86 02:19 EST AIList Digest Wednesday, 22 Jan 1986 Volume 4 : Issue 11 Today's Topics: Query - LISP Language Standard, Correction - Spang Robinson Report on Reasoning Systems, AI Tools - AI and Supercomputers & MRS, Definitions - Paradigm & Symbol, Expert Systems & AI in the Media - Connectionist Speech Learning & Arthur Young's System for Financial Auditing ---------------------------------------------------------------------- Date: 21 Jan 86 01:18:00 PST From: sea.wolfgang@ames-vmsb.ARPA Reply-to: sea.wolfgang@ames-vmsb.ARPA Subject: LISP Language Standard I am currently involved in the definition of some loose LISP programming standards [loose LISP sink ships], has anyone given any thought to this, particularly as it applies to LISP environments, or does anyone know of any articles on the topic?. I will be happy to collect responses and send them back out on the List. Thank you, S. Engle, Informatics General Co. NASA/Ames Research Center MS 242-4 Moffet Field, CA 95035 SEA.WOLFGANG@AMES-VMSB.ARPA ------------------------------ Date: Wed, 15 Jan 86 04:18:25 cst From: Laurence Leff Subject: Correction [Joseph Rockmore, vice president of Reasoning Systems, says that the Spang Robinson report on his company's agreement with Lockheed was correct, but that the summary in AIList incorrectly identified his company's work with "USC Kestrel Institute". He points out that Reasoning Systems is associated with Kestrel, but that neither is associated with USC-ISI. Laurence Leff has provided the following additional summary in the course of resolving this matter. Contact rockmore@kestrel.ARPA for further information. -- KIL] In my abstracts of Spang Robinson Report, I reported parenthetically that Reasoning Systems is commercializing the work of [...] Kestrel Institute. That parenthetical statement was based on my own analysis of the situation and was not included in the Spang Robinson report. My apologies for any confusion created. Its was based on what I perceived to be a similarity between the work and the fact that one person has moved from that organization over to Reasoning Systems (as indicated in the address of authors section of IEEE Transactions on software Engineering). Also, quoting from "Software Environments at Kestrel Institute" in the November 1985 Volume Se-11 No 11, "One of the authors (G. B. Kotik) is currently with Reasoning Systems, a company founded in 1984 in order to apply the body of basic research in knowledge-based programing to commercial problems. Reasoning Systems develops special-purpose knowledge-based program generators and programming environments for various domains." and later in the same article "Toward these ends, Reasoning Systems has developed a system called REFINE," "Although REFINE derives its inspiration from many sources, it utilizes the principles and system structure laid out in the CHI project." ------------------------------ Date: Tue 21 Jan 86 13:52:27-CST From: CMP.BARC@R20.UTEXAS.EDU Subject: AI and Supercomputers On January 17, UCSD offered a one-day program, called "Capabilities and Applications of the San Diego Supercomputer Center", in conjunction with the opening of their new center. One of the talks was "AI and Expert Systems on Supercomputers" by Dr. Robert Leary, a Senior Staff Scientist at the San Diego Supercomputer Center. I didn't attend the course but heard that Leary's talk was preliminary and did not present any significant applica- tions. Further information can probably be obtained from SDSC on the UCSD campus or from UCSD Extension. The address of UCSD is La Jolla, CA 92903. Dallas Webster CMP.BARC@R20.UTexas.Edu ------------------------------ Date: Tue, 21 Jan 86 09:52:16 est From: Walter Hamscher Subject: MRS Date: Fri, 17 Jan 86 15:04:24 est From: Tom Scott Subject: Two questions on knowledge-engineering software 1. Rick Dukes from Symbolics recently gave an interesting talk on AI/KE to the Northwest Ohio chapter of the ACM. He mentioned an expert-system-building tool, MRS, from Stanford. * * * Can anyone tell me about the system? What does it do? What representation and search techniques are available through it? It's a logic programming system written in Lisp. The principal underlying inference engine is resolution, you can also do forward & backward chaining. The name means `Metalevel Reasoning System' because you can write meta-level axioms, axioms about the base level knowledge -- usually these meta axioms are used to guide the search-based inference procedures. I hear the latest version lets one write meta-meta-axioms, meta-meta-meta-axioms, etc ("Anything you can do, I can do Meta," as Brachman says). For background see "An Overview of Meta-Level Architecture" Genesereth AAAI-83. Stanford Heuristic Programming Project probably has some kind of MRS manual; there's also an `MRS Dictionary' but that's really more of a reference tool. Can it handle frames? Semantic networks? Certainty factors? It can `handle' anything you can write in lisp... does it provide any of these facilities, No, I don't think so. How does it work as an expert-system development environment? Good question. How does Lisp work as an expert-system environment? For applications to troubleshooting & test generation see Genesereth, AAAI-82; Yamada, IJCAI-83; Singh's PhD thesis from Stanford (1985); Genesereth in AI Journal V 24 #1-3 or `Qualitative Reasoning about Physical Systems', ed. Bobrow. It's NOT a traditional expert-system envirionment ala KEE, ART, S1, DUCK, etc. Most importantly, how does a university acquire MRS? Jane Hsu (HSU@SCORE) should be able to tell you all about this. I believe she's charge of maintenance & distribution. She may refer you on to Arthur Whitney, but try Jane first. I think Rick told us that it was available to universities essentially for free. If that is true, then where can we send a tape? For some reason the figure $500 sounds right, but don't quote me. ------------------------------ Date: Fri, 17 Jan 86 10:43:33 PST From: kube%cogsci@BERKELEY.EDU (Paul Kube) Subject: What's a paradigm? A classic attempt to figure out just what the devil Kuhn means by `paradigm' is Margaret Masterman's `The nature of a paradigm' (in _Criticism and the Growth of Knowledge_, I. Lakatos and A. Musgrave, eds.). She finds 21 ("possibly more, not less") senses of the term in the first edition of _The Structure of Scientific Revolutions_; take your pick. ------------------------------ Date: Wed, 22 Jan 86 02:16:17 PST From: kube%cogsci@BERKELEY.EDU (Paul Kube) Subject: Re: What is a symbol? >.... Newell and Simon's Physical Symbol >System Hypothesis, that a machine that carries out processes operating on >symbol structures has the necessary and sufficient means for general >intelligent action, seems to be an expression of the underlying assumptions >of the majority of work in AI. ... > A symbol is a formal entity whose internal structure > places no restrictions on what it may represent in the > domain of interest. > >Unfortunately, when combined with the Physical Symbol System Hypothesis, >this notion of symbol creates a problem with regard to so-called >"connectionist" systems. I think at least two concepts, not just one, need some work here: it would help to have a better idea not only of what symbols are, but also of what operating on a symbol is. Under what one might call the Turing conception of `operating on a symbol'-- a strong, agentive interpretation: symbols are objects that get manipulated by a processor, e.g. written on and erased from a tape, or shuffled from location to location--I think that it's probably true that connectionist systems do not `operate' on symbols that have interesting external referents. But I doubt that the majority of workers in AI believe that in this sense `operating on symbols' is necessary for the production of intelligent action, and so there is no conflict with connectionism; that construal of the PSSH is easy enough to give up. (That `operating on symbols' in the Turing sense be sufficient for the production of intelligent action is, however, pretty clearly an underlying assumption of work in the field; but of course this doesn't conflict with connectionism either.) On the other hand, a weaker interpretation of what operating on symbols amounts to gives a PSSH that is compatible with connectionism, not to mention being more likely to be true. Certainly what's important about symbols for theory construction in AI is that they have formal properties which determine their interactions with other symbols without regard to any semantic properties they might have, while being susceptible of being assigned semantic properties in a way that is dependent on these interactions. (Anyway I don't think it's helpful to require of a symbol that its `internal structure places no restrictions on what it may represent', at least without further specification of what counts as internal structure. Take an English word: `symbol', say. What's between the quotes is a symbol, I'd think, but intuitively its internal structure places pretty strong restrictions on what it represents: try composing it of six different letters, for example.) But then they don't need to be objects; symbols can be states, and the formal properties which determine their interaction (`operations' on them) can be identified with certain of their causal properties. Now, one way a system can be in symbolic states is to operate on symbols in the strong, Turing sense; but this is only one way. Symbolic states can also be emergent states of a connectionist system. Paul Kube Computer Science Division U.C. Berkeley Berkeley, CA 94720 kube@cogsci.berkeley.edu ucbvax!kube ------------------------------ Date: Mon, 20 Jan 86 16:57:43 mst From: ted%nmsu.csnet@CSNET-RELAY.ARPA Subject: today show segment I think that the work that was mentioned recently in the digest from the today show (which I didn't see) was the speech synthesis work which was described earlier on the aidigest (sketchily). I don't remember the contact (sejnowski??), but the machine was a neural analog network that modified it's own weights when given a training corpus of textual english with correct voice synthesizer outputs. Then, when given more english (it wasn't clear that this new text had not appeared in the original training corpus) the machine produced coherent control inputs for the voice synthesizer. Claims that ``it learns to speak the way that human babies do'' and so on are obviously bunk since people don't learn initially to read text and because people also have to derive the correlation between their motor stimulation (essentially the voice synthesizer control level), the sound thereby produced and the percepts that are returned via their ears. A measure of the comparative difficulty is that programs which do text to speech conversion extremely well have been existence for several years (DECtalk is the current avatar), but no program can yet even reproduce an infant's use of auditory language. Certainly, no-one can be claiming that a program that can learn to do the former must be able to consequently be able to learn to do the latter, much less that the acquisition method that would be used is the one used by human children. The most interesting thing is that my original contact with the author of the project in question (I think), is that he never mentioned this sort of comparison. sigh....the original work was interesting, possibly even progressive. But then here comes the today show interviewers looking for a BREAKTHROUGH. So they find (make) one and we hear about another case of ai-hype. Everybody get ready for another wave of flames. ------------------------------ Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: News Flash Source: January 16 Wall Street Journal FIRST Page "CPA firm Arthur Young unveils a computer system today that uses expert systems to help the auditor focus on areas where risk of error is greatest. The system could mean average savings of 10% in time and money, says Arthur Young's Robert Temkin" ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Thu Jan 23 02:46:37 1986 Date: Thu, 23 Jan 86 02:46:34 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a001519; 22 Jan 86 16:29 EST Date: Wed 22 Jan 1986 13:10-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #12 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 23 Jan 86 02:26 EST AIList Digest Thursday, 23 Jan 1986 Volume 4 : Issue 12 Today's Topics: Natural Language - Modulated Kitchens and Superior Borders, Humor - Pseudoscience Jargon, Logic & Humor - Proof that P = NP, Games - Othello Tournament Information Literature - New Text on Natural Language Processing ---------------------------------------------------------------------- Date: 22 Dec 1985 1822-PST (Sunday) From: Steven Tepper Subject: modulated kitchens and superior borders >From a recent issue of the Chronicle: "When you mount the cooker hood on a modulated kitchen, please care that the superior border of the caliber is on the inferior border of the incorporated board. When you fix the cooker hood to the incorporated board, please set this border on the wall up on the bottom of the incorporated board and use the unhooped holes." Instructions for fitting a stove hood made in Italy by the Zanussi company. The Plain English Campaign in London has awarded the directions its annual prize for the worst example of bureaucratic language, citing an "incompetent and baffling translation from an unknown language into sub-English." [This should give the machine translation people something to shoot for. -- KIL] ------------------------------ Date: Wed, 15 Jan 86 09:59 EST From: Sonny Crockett Subject: A good one on HAL I just got the videotape of 2010, and figured out what Dr. Chandra said about the reason HAL screwed up in the first mission. The major problem most SF authors have is trying to come up with ways to express advanced scientific things in a way that sounds very scientific...this is a great one: (Dr. Chandra has just finished explaining that HAL was given conflicting orders, and was only trying to interpret them the best he could) "...HAL was trapped, more precisely he got caught in an H. Mobius Loop, which is possible in autonomous call-seeking computers." I thought it was funny, anyway... -Chris PS If anyone (like me) enjoys laughing at these kind of "pseudo-science" phrases, I recommend watching Dr. Who (most famous for "Multi-dimensional Time/Space Vortex"), and Star Trek ("Hodgkins Theory for Parallel Planet Development," is one of my favorites). I'm sure there are many others as well. ------------------------------ Date: Fri, 27 Dec 85 16:11:46 pst From: Alain Fournier Reply-to: fournier@Navajo.UUCP (Alain Fournier) Subject: Logic & Humor - Proof that P = NP [Forwarded from the Stanford bboard by Laws@SRI-AI.] > From: Len > > $15 to anyone who can prove P = NP. > > #8^) > Len This is an old one, but what the hell, it's $15.00: ----------------------------------- | Exactly 2 of the statements | | in these 3 boxes are false | | | ----------------------------------- ----------------------------------- | | | P != NP | | | ----------------------------------- ----------------------------------- | The statement in the first | | box is true. | | | ----------------------------------- It is left to the reader to show that assuming statement 1 is true leads to a contradiction, so 1 is false, therefore 3 is false, and 2 has to be false. The same conclusion is reached if the truth value of 3 is examined. So 2 is false, and P=NP, QED. The $15 can be sent in my name to my favourite charity, the Douglas Hofstadter Home for the Terminally Self-Referential. An accompanying note should specify that I requested that my gift should have no accompanying note. ------------------------------ Date: January 17, 1986, 5:51 PM. From: <1gtlmkf%calstate.BITNET@WISCVM.WISC.EDU> Subject: Othello Tournament Information For anyone who might be interested in the upcoming Computer Othello Tournament at CSU, Northridge on February 15-16: Yoy may contact the tournament organizers over BITNET at the following addresses -- Brian Swift (AGTLBJS@CALSTATE.BITNET) Marc Furon (1GTLMKF@CALSTATE.BITNET) Any questions or requests for information about the tournament may be sent to either of us at the addresses above. We look forward to a successful tournament and hope to hear from any and all interested Othello programmers. Thanks to Kurt Godden for sending the announcement to AILIST. Marc Furon Yes, Othello is a trademark of CBS Toys. ------------------------------ Date: Tue, 14 Jan 86 09:52:55 EST From: "Richard E. Cullingford" Subject: new AI text This note is an announcement of a new AI book which may be of interest to the readers of this newsgroup. The book is "Natural Language Processing: A Knowledge Engineering Approach," and it will be available from Rowman & Allanheld, Publishers, of Totowa, NJ, early in the spring of 1986. The work is intended as a practical introduction to a theory and technology for building natural language text-processing interfaces to database management or expert reasoning systems. The text has been in use, in manuscript, in courses at Princeton University and Georgia Tech for the past two years, and extensive course materials have been developed. A software system, the NLP Toolkit, is also available, through the publisher, that runs all of the text's examples, and is suitable for experimentation by teachers and programmers. The Toolkit contains representation design tools, a conceptual analyzer, a conceptual generator, a large shared dictionary, and a knowledge-base management support package. Questions regarding the book and the programs can be addressed to the author, Richard E. Cullingford, at the School of Information & Computer Science, georgia Tech, Atlanta, GA 30332; at (404) 894-3227; or gatech!rec (uucp) or rec@gatech (csnet). The book's table of contents follows: Table Of Contents Natural Language Processing: A Knowledge Engineering Approach Preface Notes on the Use of This Book Acknowledgments Table of Contents Table of Diagrams Table of Figures Chapter 1: Natural Language Processing: An Overview 1.0 Introduction 1.1 Related Fields: An Overview 1.1.1 NLP, Artificial Intelligence, and Knowledge Engineering 1.1.2 NLP and the Sciences of Language 1.2 NLP Efforts in AI 1.2.1 Early Efforts 1.2.2 Second Generation Systems 1.2.3 Third Generation Systems: A Look into the Future 1.3 Outline of the Book Part I: A General-Purpose Language Processing Interface Chapter 2: An Introduction to Representation Design 2.0 The Representation Problem 2.1 The Need for a Formal Representational System 2.2 Requirements on a Representational System 2.3 Introduction to ERKS 2.3.1 The ISA-Hierarchy of the Core System 2.3.2 Criteria for Selection of the Primitive Types 2.4 ERKS in LISP 2.5 The Maximal Inference-Free Paraphrase 2.6 Building a Model Corpus 2.7 A Simple Corpus 2.8 Primitive Actionals and Statives 2.9 Conceptual Relationships 2.10 A Representational Case Study: CADHELP 2.10.1 The CADHELP Microworld 2.10.2 A Typical Command 2.10.3 Knowledge Representation Issues 2.11 Summary Chapter 3: Software Tools for Representation Design 3.0 Introduction 3.1 Navigating in an ISA-Hierarchy 3.2 Defining ERKS Types 3.3 Access and Updating Machinery 3.4 The def-wordsense Record Macro 3.5 Summary Chapter 4: Surface-Semantic Conceptual Analysis 4.0 Introduction: Lexicon-Driven Analysis 4.1 A Simple Model of Sentence Structure 4.2 Production Systems, Requests, and Processing Overview 4.3 Request Pool Consideration 4.3.1 Analysis Environment 4.3.2 Request Types 4.4 Requests in More Detail 4.5 Morphological Fragments and "to be" 4.6 A Processing Example 4.7 Summary Chapter 5: Problems in Conceptual Analysis 5.0 Introduction 5.1 Tri-Constituent Forms and Imbedded Sentences 5.1.1 Handling Indirect Objects 5.1.2 Infinitives and Gerunds 5.1.3 Relative Clauses 5.2 Prepositions and "to be," Revisited 5.3 Word Meaning Disambiguation 5.3.1 Pronominal Reference 5.4 Coordinate Constructions 5.5 Ellipsis Expansion 5.6 A Concluding Example 5.7 Summary Chapter 6: Generating Natural Language from a Conceptual Base 6.0 Introduction 6.1 Overview of Generation Process 6.2 Dictionary Entries 6.3 Morphology and the Verb Kernel 6.3.1 Plural and Possessive Morphology 6.3.2 Subject-Verb Agreement and Modals 6.3.3 Tensing 6.3.4 Subject-Auxiliary Inversion 6.4 "Advanced" English Syntax 6.4.1 The Infinitive Construction 6.4.2 The Possessive Sketchifier 6.4.3 The Entity-Reference Sketchifier 6.5 A Processing Example 6.6 Summary Part II: Building a Conversationalist Chapter 7: Summarizing Knowledge Bases 7.0 Introduction: What to Say versus How to Say It 7.1 Explanations as Summaries 7.2 Explanations in CADHELP 7.3 Representational Overview 7.4 Concept Selection 7.5 An Example 7.6 Summary Chapter 8: Knowledge-Base Management 8.0 Introduction 8.1 KB Organization 8.1.1 The Slot-Filler Tree 8.1.2 Slot-Filler Tree Construction 8.1.3 Index Quality 8.1.4 Best-First Ordering of KB Items 8.2 KB Search 8.2.1 The Tree Search Mechanism 8.3 Performance 8.4 Summary Chapter 9: Commonsense Reasoning 9.0 Introduction: The Need for Reasoning in Language Understanding 9.1 Deductive Retrieval 9.2 YADR, Yet Another Deductive Retriever 9.3 The YADR Interface 9.4 The YADR Top Level 9.5 Logical Connectives in Antecedent Forms 9.6 Summary Chapter 10: Putting It All Together: A Goal-Directed Conversationalist 10.0 Introduction 10.1 The ACE Microworld 10.2 A Model of Purposive Conversation 10.3 The Conversational Strategist 10.4 The Conversational Tactician 10.5 The Academic Scheduling Expert 10.6 More Problems in Language Understanding 10.6.1 Coordinate Constructions and Ellipses 10.6.2 Defining "And" for the Analyzer 10.6.3 Using Expectations during Analysis 10.7 More Problems in Language Generation 10.7.1 Asking Questions 10.7.2 Producing Coordinate Constructions 10.7.3 Generating Attributes, Absolute Times, Locales, and Names 10.8 Putting It All Together: A Session with ACE 10.9 Parting Words Appendix I: The ERKS Types Appendix II: Source for YADR, Yet Another Deductive Retriever Appendix III: Glossary of Terms Rich Cullingford ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Thu Jan 23 02:46:28 1986 Date: Thu, 23 Jan 86 02:46:24 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a001571; 22 Jan 86 16:40 EST Date: Wed 22 Jan 1986 13:17-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #13 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 23 Jan 86 02:28 EST AIList Digest Thursday, 23 Jan 1986 Volume 4 : Issue 13 Today's Topics: Seminars - Controlling Backward Inference (SRI) & Automata Approach to Program Verification (MIT) & Problem Solving for Distributed Systems (MIT) & Problem-Solving Languages (CSLI) & Pointwise Circumscription (SU) & Methodological Issues in Speech Recognition (Edinburgh) & Intuitionistic Logic Programming (UPenn) ---------------------------------------------------------------------- Date: Wed 15 Jan 86 15:22:41-PST From: LANSKY@SRI-AI.ARPA Subject: Seminar - Controlling Backward Inference (SRI) CONTROLLING BACKWARD INFERENCE Dave Smith (DE2SMITH@SUMEX-AIM) Stanford University 11:00 AM, MONDAY, January 20 SRI International, Building E, Room EJ228 (new conference room) Effective control of inference is a critical problem in Artificial Intelligence. Expert systems have made use of powerful domain-dependent control information to beat the combinatorics of inference. However, it is not always feasible or convenient to provide all of the domain-dependent control that may be needed, especially for systems that must handle a wide variety of inference problems, or must function in a changing environment. In this talk a powerful domain-independent means of controlling inference is proposed. The basic approach is to compute expected cost and probability of success for different backward inference strategies. This information is used to select between inference steps and to compute the best order for processing conjuncts. The necessary expected cost and probability calculations rely on simple information about the contents of the problem solvers database, such as the number of facts of each given form and the domain sizes for the predicates and relations involved. ------------------------------ Date: 01/16/86 17:18:02 From: LISA at MC.LCS.MIT.EDU Subject: Seminar - Automata Approach to Program Verification (MIT) [Forwarded from the MIT bboard by SASW@MC.LCS.MIT.EDU.] DATE: Thursday, January 23, 1986 TIME: 3:45 p.m......Refreshments 4:00 p.m......Lecture PLACE: NE43 - 512A "AN AUTOMATA-THEORETIC APPROACH TO AUTOMATIC PROGRAM VERIFICATION" MOSHE Y. VARDI IBM Almaden Research Center We describe an automata-theoretic approach to automatic verification of concurrent finite-state programs by model checking. The basic idea underlying the approach is that for any temporal formula PHI we can construct an automaton A(PHI) that accepts precisely the computations that satisfy PHI. The model-checking algorithm that results from this approach is much simpler and cleaner than tableaux-based algorithms. We also show how the approach can be extended to probabilistic concurrent finite-state programs. Albert Meyer Host ------------------------------ Date: Thu 16 Jan 86 11:49:34-EST From: John J. Doherty Subject: Seminar - Problem Solving for Distributed Systems (MIT) [Forwarded from the MIT bboard by SASW@MC.LCS.MIT.EDU.] Date: January 24, 1986 Place: NE43-512A Time: Refreshments 2:15 P.M. Seminar 2:30 P.M. Problem Solving for Distributed Systems: An Uplifting Experiment in Progress Herb Krasner Member of the Technical Staff MCC Software Technology Project This presentation describes the empirical studies efforts of the STP Design Process Group focusing on models of the design process. Preliminary findings of the "lift" experiment are reported, from which a model of expert designer behavior and high leverage characteristics is being derived. Goals of the pilot study, experimental setup, problem, data analysis technique, hypotheses and subsequent activities are discussed. The "lift" experiment was initiated to examine the early stages of design problem solving behavior prototypical of users of the futuristic software design environment LEONARDO. It addresses the large effect of individual differences on productivity data, and differs from previous studies in its focus on large-scale design problems. Host: Irene Greif ------------------------------ Date: Wed 15 Jan 86 16:52:56-PST From: Emma Pease Subject: Seminar - Problem-Solving Languages (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] CSLI ACTIVITIES FOR NEXT THURSDAY, January 23, 1986 2:15 p.m. CSLI Seminar Computer Problem Solving Languages, Programming Languages and Mathematics Curtis Abbott (Abbott@xerox) Computer Problem Solving Languages, Programming Languages and Mathematics by the Semantically Rational Computer Languages Group Programming languages are constrained by the requirement that their expressions must be capable of giving rise to behavior in an effective, algorithmically specified way. Mathematical formalisms, and in particular languages of logic, are not so constrained, but their applicability is much broader than the class of problems anyone would think of ``solving'' with computers. This suggests, and I believe, that formal languages can be designed that are connected with the concerns associated with solving problems with computers yet not constrained by effectiveness in the way programming languages are. I believe that such languages, which I call ``computer problem solving languages,'' provide a more appropriate evolutionary path for programming languages than the widely pursued strategy of designing ``very high level'' programming languages, and that they can be integrated with legitimate programming concerns by use of a transformation-oriented methodology. In this presentation, I will give several examples of how this point of view impacts language design, examples which arise in Membrane, a computer problem solving language I am in the process of designing. --Curtis Abbot ------------------------------ Date: 17 Jan 86 1639 PST From: Vladimir Lifschitz Subject: Seminar - Pointwise Circumscription (SU) Pointwise Circumscription Vladimir Lifschitz Thursday, January 23, 4pm, MJH 252 (I have a few copies of the paper in my office, MJH 362). ABSTRACT Circumscription is logical minimization, that is, the minimization of extensions of predicates subject to restrictions expressed by predicate formulas. When several predicates are to be minimized, circumscription is usually thought of as minimization with respect to an order defined on vectors of predicates, and different ways of defining this order correspond to different kinds of circumscription: parallel and prioritized. The purpose of this paper is to discuss the following principle regarding logical minimization: Things should be minimized one at a time. This means, first of all, that we propose to express the circumscription of several predicates by the conjunction of several minimality conditions, one condition for each predicate. The difference between parallel and prioritized circumscription will correspond to different selections of predicates allowed to vary in each minimization. This means, furthermore, that we propose to modify the definition of circumscription so that it will become an "infinite conjunction" of "local" minimality conditions; each of these conditions expresses the impossibility of changing the value of the predicate from True to False at one point. (Formally, this "infinite conjunction" will be represented by means of a universal quantifier). This is what we call "pointwise circumscription". We argue that this approach to circumscription is conceptually simpler than the traditional ``global'' approach and, at the same time, leads to generalizations with the additional flexibility and expressive power needed in applications to the theory of commonsense reasoning. Its power is illustrated, in particular, on a problem posed by Hanks and McDermott, which apparently cannot be solved using other existing formalizations of non-monotonic reasoning. ------------------------------ Date: Tue, 21 Jan 86 10:44:43 GMT From: Gideon Sahar Subject: Seminar - Methodological Issues in Speech Recognition (Edinburgh) Department of Artificial Intelligence, University of Edinburgh, and Artificial Intelligence Applications Institute Edinburgh Artificial Intelligence Seminars Speaker - Dr. Henry Thompson, Dept. of A.I., University of Edinburgh Title - Methodological issues in Speech Recognition Abstract - What methodological issues arise from the belief that fully automatic high quality unrestricted speech recognition is impossible, when one has overall technical responsibility for a multi-year multi-million pound Alvey Large Scale Demonstrator? I will give a brief overview of the overall structure of the project, and discuss at more length two basic issues: - Why top-down vs. bottom-up is the wrong question, and selectional vs. instructional interaction is the right question, and what the right answer is. - How giving up on fully automatic ... changes the way you do things in surprising ways. ------------------------------ Date: Tue, 21 Jan 86 23:27 EST From: Tim Finin Subject: Seminar - Intuitionistic Logic Programming (UPenn) From: Dale Miller UPenn Math-CS Logic Seminar An Intuitionistic Basis for Extending Logic Programming Dale Miller Tuesday 28 Jan 86, 4:30 - 6:00, 4E17 DRL There is a very natural extension to Horn clauses which involves extending the use of implication. This extension has a natural operational semantics which is not sound with respect to classical logic. We shall show that intuitionistic logic, via possible worlds semantics, provides the necessary framework to give a sound and complete justification of this operational semantics. This will be done by providing a least fix-point construct of a Kripke-model. We shall also show how this logic can be used to provide logic programming languages with a logical foundations for each of the following programming features: program modules, recursive call memo-izing, and local environments (permanent vs. temporary asserts). This extension to logic programming can also simulate various features of negation - not through logical incompleteness as in negation-by-failure, but through constructing proofs of a certain kind. ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Sat Jan 25 02:48:56 1986 Date: Sat, 25 Jan 86 02:48:50 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a028883; 24 Jan 86 23:02 EST Date: Fri 24 Jan 1986 10:12-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #14 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sat, 25 Jan 86 02:22 EST AIList Digest Friday, 24 Jan 1986 Volume 4 : Issue 14 Today's Topics: Conferences - Knowledge Acquisition for KB Systems & Symposium on Logic Programming & Office Information Systems '86 & Uncertainty and KBS, Course - Object-Oriented Programming ---------------------------------------------------------------------- Date: Thu, 16 Jan 86 11:20:57 pst From: bcsaic!john@uw-june.arpa Subject: Workshop - Knowledge Acquisition for KB Systems Call for Participation KNOWLEDGE ACQUISITION FOR KNOWLEDGE-BASED SYSTEMS WORKSHOP Sponsored by AMERICIAN ASSOCIATION FOR ARTIFICIAL INTELLIGENCE (AAAI) Banff, CANADA November 3-7, 1986 The bottleneck in the process of building knowledge-based systems is usally acquiring the appropriate problem solving knowledge. The objective of this workshop is to assemble theoriticians and practioners of AI who recognize the need for developing systems that assist the knowledge acquisition process. To encourage vigorous interaction and exchange of ideas the workshop will be kept small - about 30 participants. There will be individual presentations and ample time for technical discussions. An attempt will be made to define the state-of-the-art and the future research needs. Papers are invited for consideration in all aspects of knowledge acquisition for knowledge-based systems, including (but not restricted to): o Transfer of expertise - systems which interview experts to obtain and structure knowledge. o Transfer of expertise - manual knowledge engineering interviewing methods and techniques. o Induction of knowledge from examples. o Knowledge acquisition methodology. Four copies of an extended abstract (up to 8 pages, double-spaced) or a full-length paper should be sent to the workshop chairman before May 1, 1986. Acceptance notices will be mailed by July 1. Revised abstracts should be returned to the chairman by October 1, 1986, so that they may be bound together for distribution at the workshop. Potential attendees should also indicate their interest in chairing or participating in special topic discussion sessions. Co-Chairmen: John Boose (send papers here) Boeing Artificial Intelligence Center Boeing Computer Services M/S 7A-03 PO Box 24346 Seattle, Washington, USA, 98124 (206) 763-5811 and Brian Gaines Department of Computer Science University of Calgary 2500 University Dr. NW Calgary, Alberta, Canada, T2N 1N4 (403) 220-6015 Program and local arrangements comittee: Jeff Bradshaw, Boeing Computer Services William Clancey, Stanford University Cathy Kitto, Boeing Computer Services Janusz Kowalik, Boeing Computer Services John McDermott, Carnegie-Mellon University Ryszard Michalski, Univ. of Illinois (tentative) Art Nagai, Boeing Computer Services Mildred Shaw, University of Calgary John Boose, Boeing Artficial Intelligence Center 7A-03, PO Box 24346, Seattle, Wa., 98124, (206) 763-5811 ------------------------------ Date: Thu 9 Jan 86 09:04:20-MST From: "Robert M. Keller" Subject: Conference - Symposium on Logic Programming '86 SLP Call for Papers Third Symposium on Logic Programming Sponsored by the IEEE Computer Society September 21-25, 1986 Westin Hotel Utah Salt Lake City, UT The conference solicits papers on all areas of logic programming, including, but not confined to: Applications of logic programming Computer architectures for logic programming Databases and logic programming Logic programming and other language forms New language features Logic programming systems and implementation Parallel logic programming models Performance Theory Please submit full papers, indicating accomplishments of substance and novelty, and including appropriate citations of related work. The suggested page limit is 25 double-spaced pages. Send eight copies of your manuscript no later than 15 March 1986 to: Robert M. Keller SLP '86 Program Chairperson Department of Computer Science University of Utah Salt Lake City, UT 84112 Acceptances will be mailed by 30 April 1986. Camera-ready copy will be due by 30 June 1986. Conference Chairperson Exhibits Chairperson Gary Lindstrom, University of Utah Ross Overbeek, Argonne National Lab. Tutorials Chairperson Local Arrangements Chairperson George Luger, University of New Mexico Thomas C. Henderson, University of Utah Program Committee Francois Bancilhon, MCC William Kornfeld, Quintus Systems John Conery, University of Oregon Gary Lindstrom, University of Utah Al Despain, U.C. Berkeley George Luger, University of New Mexico Herve Gallaire, ECRC, Munich Rikio Onai, ICOT/NTT, Tokyo Seif Haridi, SICS, Sweden Ross Overbeek, Argonne National Lab. Lynette Hirschman, SDC, Paoli Mark Stickel, SRI International Peter Kogge, IBM, Owego Sten Ake Tarnlund, Uppsala University ------------------------------ Date: Tue, 21 Jan 86 19:22 EST From: Hewitt@MIT-MC.ARPA Subject: Conference - OIS-86 Because of the delay in the distribution of the call for papers for OIS-86 in the Newsletter, we have decided to postpone the deadline for paper submission from February 1 to March 1, 1986 in order to satisfy the requirements for broad distribution of the call. Enclosed please find the updated call for papers which reflects this change: ******************* C A L L F O R P A P E R S * * ---------------------------------------------- * * Third ACM Conference On * * OFFICE INFORMATION SYSTEMS * OIS-86 * * * October 6-8, 1986 * * Biltmore Plaza Hotel * * Providence, RI ******************* ------------------------------------------------- General Chair: Carl Hewitt, Topics appropriate for this MIT conference include (but are not restricted to) the following as they Program Chair: Stanley Zdonik, relate to OIS: Brown University Technologies including Display, Voice, Treasurer: Gerald Barber, Telecommunications, Print, etc. Gold Hill Computers Human Interfaces Local Arrangements: Andrea Skarra, Brown University Deployment and Evaluation An interdisciplinary conference on System Design and Construction issues relating to office information systems (OIS) sponsored Goals and Values by ACM/SIGOIS in cooperation with Brown University and the MIT Distributed Services and Applications Artificial Intelligence Laboratory. Submissions from the following Knowledge Bases and Reasoning fields are solicited: Distributed Services and Applications Anthropology Artificial Intelligence Indicators and Models Cognitive Science Computer Science Needs and Organizational Factors Economics Management Science Impact of Computer Integrated Psychology Manufacturing Sociology The program committee includes: Bob Allen Ray Panko Bellcore University of Hawaii Guiseppe Attardi Robert Rosin University of Pisa Syntrex James Bair Erik Sandewall Hewlett Packard Linkoping University Gerald Barber Walt Scacci Gold Hill Computers USC Peter de Jong Andrea Skarra MIT Brown University Irene Greif Susan Leigh Star MIT Tremont Research Institute Sidney Harris Luc Steels Georgia State University University of Brussels Carl Hewitt Sigfried Treu MIT University of Pittsburgh Heinz Klein Dionysis Tsichritzis SUNY University of Geneva Fred Lochovsky Eleanor Wynn University of Toronto Brandon Interscience Fanya Montalvo Aki Yonezawa MIT Tokyo Institute of Technology Naja Naffah Stanley Zdonik Bull Transac Brown University Margrethe Olson NYU Professor J.C.R. Licklider of MIT will be the keynote speaker. Unpublished papers of up to 5000 words (20 double-spaced pages) are sought. The first page of each paper must include the following information: title, the author's name, affiliations, complete mailing address, telephone number and electronic mail address where applicable, a maximum 150-word abstract of the paper, and up to five keywords (important for the correct classification of the paper). If there are multiple authors, please indicate who will present the paper at OIS-86 if the paper is accepted. Proceeedings will be distributed at the conference and will later be available from ACM. Selected papers will be published in the ACM Transactions on Office Information Systems. Please send eight (8) copies of the paper (which must arrive by March 1, 1986) to: Prof. Stan Zdonik OIS-86 Program Chair Computer Science Department Brown University P.O. Box 1910 Providence, RI 02912 DIRECT INQUIRIES TO: Margaret H. Franchi (401) 863-1839. IMPORTANT DATES Deadline for Paper Submission (postponed 1 mo.) March 1, 1986 Notification of Acceptance: April 30, 1986 Deadline for Final Camera-Ready Copy: July 1, 1986 Conference Dates: October 6-8, 1986 ------------------------------ Date: Thu 23 Jan 86 16:25:50-PST From: RUSPINI@SRI-AI.ARPA Subject: Conference - International Conference on Uncertainty and KBS ANNOUNCEMENT AND CALL FOR PAPERS INTERNATIONAL CONFERENCE ON INFORMATION PROCESSING AND MANAGEMENT OF UNCERTAINTY IN KNOWLEDGE-BASED SYSTEMS Paris, France June 30 - July 4 1986 Supported by: Ministere de la Recherche et de la Technologie AFCET Centre Nationale de la Recherche Scientiique Societe Francaise de Theorie de l'Informationn Chairpersons: Bernadette Bouchon (France) Ronald R. Yager (United States) Purpose of the Conference: The aim of this Conference is to bring together researchers working on information, uncertain data processing and related topics. The management of uncertainty is at the heart of many knowledge-based systems and a number of approaches have been developed for representing these types of information. It is hoped that this Conference will provide a useful exchange between practitioners and theoreticians using these methods. INTERNATIONAL PROGRAM COMMITTEE: J. Bezdek (U.S.A.) S.Ovchinnikov (U.S.A.) C. Carlsson (Finland) J. Pearl (U.S.A.) A. De Luca (Italy) B. Picinbono (France) Deng Ju-Long (China) J. Pitrat (France) H.J. Efstathiou (G.B.) D. Ralescu (U.S.A.) C. Gueguen (France) E. Ruspini (U.S.A.) S. Guiasu (Canada) A.P. Sage (U.S.A.) M.M. Gupta (Canada) G. Shafer (U.S.A.) J. Kacprzyk (Poland) J.C. Simon (France) J.L. Lauriere (France) M. Sugeno (Japan) G. Longo (Italy) E. Trillas (Spain) J. Lowrance (U.S.A.) R. Valee (France) H.T. Nguyen (U.S.A.) L.A. Zadeh (U.S.A.) H.J. Zimmermann (Germany) H. Akdag (France) M. Mugur-Schacter (France) G. Cohen (France) H. Prade (France) D. Dubois (France) E. Sanchez (France) P. Godlewski (France) M. Terrenoire (France) TOPICS: Knowledge Representation Uncertainty in Expert Systems Decision Making with Uncertainty Fuzzy Logic and Fuzzy Reasoning Representation of Commonsense Knowledge Possibility Measures Mathematical Theory of Evidence Combinatorial Information Theory Shannon Theory Questionnaire Theory Pattern Recognition and Image Processing Clustering and Classification Information Security Fuzzy Sets in Operations Research SUBMISSION INFORMATION Papers will be selected on the basis of a 500 word abstract. Communications will be in FRENCH or ENGLISH. All abstracts should be sent in triplicate to the Conference Secretary: Professor G. Cohen International Conference I.P.M.U. E.N.S.T. 46, rue Barrault 75013 PARIS, FRANCE SUBMISSION DEADLINE: FEBRUARY 15, 1986 ------------------------------ Date: Wed, 22 Jan 86 15:58:24 PST From: tektronix!mako.TEK!jans@ucbvax.berkeley.edu Subject: Course - Object-Oriented Programming Tektronix will be holding four day Introductory Smalltalk-80 classes at three locations in February, March and April, 1986. This class will introduce the student to object oriented programming, enable the student to extend the Smalltalk language by adding new methods and classes, and prepare the student to write simple applications, using the model-view-controller paradigm. Class notes, two textbooks, and four lunches are included. Participants should be experienced in at least one high level programming language. Also planned is an Advanced Smalltalk-80 class, which will enable the student to utilize advanced techniques of Smalltalk, including advanced model-view- controller concepts, project management and team programming, multi-process programming, and external processes and language interfaces. Participants will be expected to understand the major classes of Smalltalk and should have three to six months of Smalltalk programming experience. Schedule: Intorductory: Gaithersburg, Maryland 18-21 February 1986 Introductory: Dallas, Texas 17-20 March 1986 Advanced: Beaverton, Oregon 1- 4 April 1986 Introductory: Irvine, California 14-17 April 1986 Contact: Sandi Unger, (503)685-2941 for registration information, or Mary Wells, (503)685-2947 for information on course content. :::::: Artificial Intelligence Machines --- Smalltalk Project :::::: :::::: Jan Steinman Box 1000, MS 60-405 (w)503/685-2956 :::::: :::::: tektronix!tekecs!jans Wilsonville, OR 97070 (h)503/657-7703 :::::: ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Sat Jan 25 02:35:18 1986 Date: Sat, 25 Jan 86 02:35:12 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a027557; 24 Jan 86 19:46 EST Date: Fri 24 Jan 1986 10:20-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #15 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sat, 25 Jan 86 02:19 EST AIList Digest Friday, 24 Jan 1986 Volume 4 : Issue 15 Today's Topics: Query - RT/PC Common Lisp, Binding - Robert Leary @ San Diego Supercomputer Center, Corrections - "Meta" Quote & MRS, AI Tools - Representation of Uncertainty in MRS, Policy - Gatewaying of AIList from Usenet Net.AI & Relevance of Theoretical Computer Science to AI ---------------------------------------------------------------------- Date: 22 Jan 1986 19:14-EST From: NGALL@G.BBN.COM Subject: RT/PC Common Lisp Query Has anyone heard anything about a Common Lisp for the RT/PC (IBM's new Risc Engineering Workstation)? (By Lucid perhaps?) -- Nick ------------------------------ Date: 22 January 1986 1326-PST (Wednesday) From: west@nprdc.arpa (Larry West) Subject: Robert Leary @ San Diego Supercomputer Center In re Dallas Webster's short note about Dr. Robert Leary [AIList V4 #11]: First, a minor correction: UCSD's zip code is 92093, not 92903. Dr. Leary is with GA Technologies (San Diego) which operates the Supercomputer center for the University. You might be able to reach him thru UCSD, but I think GA Technologies would be a better bet. The phone book lists: GA Technologies, Inc. 10955 John Jay Hopkins Dr. and this is my guess: La Jolla, CA 92037 Phone (general info): 619-455-3000 An old, but possibly still valid, net address for him is: leary%gav@lll-mfe.ARPA Larry West, UCSD Institute for Cognitive Science, west@nprdc.ARPA ------------------------------ Date: Thu, 23 Jan 1986 08:31:26 From: rjb%allegra.btl.csnet@CSNET-RELAY.ARPA Subject: V4 #11: Quote about "meta" (Regarding W. Hamscher's response in V4 #11 to a query about MRS:) Please - let's give credit where it's deserved: "Anything you can do, I can do meta" should be attributed to David Levy (via Brian Smith). I merely used it in a talk at AAAI-80 (hopefully attributing it to David). Ron Brachman ------------------------------ Date: Thu, 23 Jan 86 17:13:58 CST From: veach%ukans.csnet@CSNET-RELAY.ARPA Subject: Correction. In a recent issue the full name MRS was incorrectly reported. MRS = "Modifiable Representation System" (source - "MRS Manual", Michael R. Genesereth, et. al. 1980, Stanford Heuristic Programming Project) ------------------------------ Date: Thu 23 Jan 86 16:30:46-PST From: Yung-Jen Hsu Subject: MRS distribution & maintenance Walter, Contrary to what you said in your recent message, which appeared in AILIST V4 #11, about the distribution and maintenace of MRS, I'm NOT the person in charge of the matter. If anyone would like to get a copy of the MRS system, I believe that the best person(s) to contact is Arthur Whitney (whitney@sumex) and/or Michael Genesereth (genesereth@sumex). Best regards. Jane Hsu ------------------------------ Date: 22 Jan 86 16:00:21 PST (Wed) From: whiting@sri-spam Subject: MRS info. Re: Date: Fri, 17 Jan 86 From: Tom Scott Subject: MRS Can it (MRS) handle ... Certainty factors? The MRS that is available for common distribution doesn't have the facility for dealing with uncertainty. An implementation of Dempster's Rule has been incorporated into a non-official version. There are some fairly strong restrictions on this version, but an application using this version has been implemented. It seems the situation is more "MRS's official release doesn't include the ability to deal with uncertainty, YET", than "MRS can't handle certainty factors". [Note: "Can it handle Certainty factors?" has been generalized to "Does MRS have the ability to deal with uncertainty?". Certainty factors are generally associated with MYCIN/E-MYCIN's methodology for dealing with uncertainty.] As an aside, Stuart Russell Esq., has put together a manual which is quite good, "The Compleat (sic) Guide to MRS", Stanford Knowledge Systems Laboratory Report No. KSL-85-12. Kevin Whiting ------------------------------ Date: Wed, 22 Jan 86 16:40:07 PST From: Kenneth I. Laws Subject: Resumption of AIList Gatewaying Eric Fair, the Berkeley Postmaster, has been handling the gatewaying of AIList to the Usenet mod.ai distribution. [The "mod" stands for "moderated".] He has offered to gateway net.ai submissions back to AIList if we wish. AIList used to have such an arrangement until our SRI-UNIX gateway broke. At that time AIList traffic dropped by about 50%, primarily through the loss of cross-net discussions (as opposed to seminar and conference announcements). I do not know whether net.ai continues to carry a great deal of non-AIList traffic, nor whether there would be an increase in useful interchanges if we again make it easy for academic/foreign Usenet readers to submit material to AIList. I do not know whether Usenet readers >>like<< having a "private" discussion channel in addition to the AIList stream that they get. I expect that the connection would increase my workload, but I am willing to take on the moderation as long as no one objects to my just ignoring net.ai comments that do not seem relevant. (Sending explict rejection notices involves numerous hassles, and hardly seems worth the effort since the submitter has already reached his net.ai audience and would be unaware of whether AIList also carried the text.) So, does anyone feel strongly one way or the other? The default is to go ahead with the connection, at least until it proves unmanageable. (I would rather drop seminar notices than lose personal interaction.) -- Ken Laws ------------------------------ Date: Wed, 22 Jan 86 15:57:16 -0200 From: mcvax!lifia!rit@seismo.CSS.GOV (Jean-Francois Rit ) Subject: Relevance of Theoretical Computer Science to AI To: AIList-REQUEST@SRI-AI.ARPA In article <8601151921.AA17286@ucbvax.berkeley.edu> you write: > Today's Topics: > Description - European Association for Theoretical Computer Science > .... > In our experience the only reason that a computer > scientist who is either actively engaged or interested in > theoretical computer science is not a member of EATCS... I found this message quite interesting...At least because it made me wonder why it was in mod.ai! Our Laboratory (LIFIA) clusters research groups in Theoretical CS and in AI. There are many CS labs in Grenoble, and AI could have as well been separated from TCS. However I find it quite hard to define any common interest other than "doing the soft for the future super-computer" which hardly leads to any tight cooperation (this is my personnal opinion only). Furthermore, one of the leaders of EATC is M. Nivat : >TCS Editor: M. Nivat, Paris >Past Presidents: M. Nivat, Paris (1972-1977) A semestrial course on CS of whom I attended, where I learnt much (:-) about automata and grammars but never heard the words AI. (it is said in my lab that he is not a strong supporter of AI but these are rumors that I could not personnally verify). So, are there any AI researchers who feel actively engaged or interested in TCS? ^^^^^^^^ For example, in working or publishing in one of the following fields : > Typical topics discussed during recent ICALP conferences are: > computability, automata theory, formal language theory, analysis of > algorithms, computational complexity, mathematical aspects of > programming language definition, logic and semantics of programming > languages, foundations of logic programming, theorem proving, software > specification, computational geometry, data types and data structures, > theory of data bases and knowledge based systems, cryptography, VLSI > structures, parallel and distributed computing, models of concurrency > and robotics. ^^^^^^^^ (oh oh! robotics indeed?) > ... Behind all this lie the major problems of under- > standing the nature of computation and its relation to computing > methodology. While "Theoretical Computer Science" remains mathematical > and abstract in spirit, it derives its motivation from the problems of > practical computation. I don't feel that a major problem for AI researchers is understanding the nature of computation, I think the AI point of view is much (maybe too much) broader or at least OPEN toward The "real" universe. I repeat I'm not opposed to TCS, I just wonder which real links bound TCS and AI. I'd like to know what other AI , and TCS (if they read mod.ai (-:) researchers think about that (that's why I submit this to the news). Jean-Francois Rit Laboratoire d'Informatique Fondamentale et d'Intelligence Artificielle BP 68 38402 Saint-Martin d'Heres cedex Disclaimer: This is only my postal address! UUCP: ...{mcvax,vmucnam}!lifia!rit [I was the one who forwarded the message to AIList -- perhaps I have been unduly influenced by the AI "neats" here at SRI-AI. I am a "scruffy" (or hacker or pragmatist), but there seem to be plenty of people in AI who hold that the problems will fall apart if and only if we solve the underlying difficult cases rigorously. There are those in the Representation and Reasoning Group here at the AI Center who consider automata theory an appropriate basis for robotic perception and action. Theorem proving is popular with our planning group and also underlies part of our natural language understanding effort. Grammar and formal language theories are used in NL understanding, although I don't know whether they are considered AI. Semantics of [certain] programming languages has been a topic on AIList and on the Prolog Digest, and may generate renewed interest when CommonLoops and other object-oriented languages become commonly available. Foundations of logic programming is an obvious match, and computational geometry is important to those of us in vision research. The theory of data bases is intermingled with data abstraction and conceptual modeling, as well as with practical development of efficient Prolog systems; it will become more important to AI as knowledge-based systems become larger. Parallel and distributed computing (or, at least, problem solving) are evidently of interest to the AIers on the PARSYM discussion list. Models of concurrency are important in multiagent planning. A Stanford professor has requested that I not forward any more articles from the "Theory Net" distribution. I will comply, but I do not agree that AI is (or should be) disjoint from CS theory. The results of CS research will be of use in AI, and the needs an theories of AI might well inspire further CS research. -- KIL] ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Jan 30 05:40:29 1986 Date: Thu, 30 Jan 86 05:40:20 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a022442; 29 Jan 86 23:59 EST Date: Wed 29 Jan 1986 20:41-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #16 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 30 Jan 86 05:27 EST AIList Digest Thursday, 30 Jan 1986 Volume 4 : Issue 16 Today's Topics: Journal Issue - Blackboard Models for AI in Engineering, Seminars - Naive Physics: Knowledge in Pieces (UCB) & Term Rewriting, Theorem Proving, Logic Programming (CSLI) & The Algebra of Time Intervals (SRI) & Machine Learning and Economics (RU) & Semi-Applicative Programming (UPenn) & Integrating Syntax and Semantics (Edinburgh) & Feature Structures in Unification Grammars (UPenn) ---------------------------------------------------------------------- Date: Mon, 27 Jan 86 23:21:15 est From: Michael Bushnell Subject: Call for Papers - Blackboard Models for AI in Engineering ====================================================================== | CALL FOR PAPERS for the | | | | INTERNATIONAL JOURNAL FOR | | | | ARTIFICIAL INTELLIGENCE IN ENGINEERING | | October, 1986 Special Issue | | | | Guest Editors: | | Pierre Haren, INRIA, France. | | Mike Bushnell, Carnegie-Mellon University | | | | Manuscripts in US should be sent to: | | Mike Bushnell | | Department of Electrical and Computer Engineering | | Carnegie-Mellon University | | Pittsburgh, PA 15213 | | USA | | (ARPAnet: mb@ohm.ece.cmu.edu) | | Deadline for receiving manuscripts: April 1st, 1986 | | | ====================================================================== We are soliciting papers for a special issue of the International Journal for AI in Engineering. This issue will focus on the AI Blackboard model, as applied to engineering problems. Papers describing the application of the Blackboard model to problems in the disciplines of Electrical Engineering, Computer Engineering, Chemical Engineering, Civil Engineering, Mechanical Engineering, Metallurgy, Materials Science, Robotics, and areas of Computer Science are appropriate. Papers describing applications to other disciplines may also be appropriate. In addition, papers discussing AI tools that are particularly appropriate for Engineering applications are most welcome, along with book reviews, letters to the editor, conference reports, and other relevant news. All submissions must be original papers written in English and will be refereed. The copyright of published papers will be vested with the publishers. Contributions will be classified as research papers and research notes, of up to 5000 equivalent words, or as review articles of up to 10,000 equivalent words. Authors wishing to prepare review articles should contact the editors in advance. Manuscripts should be typed double-spaced with wide margins, on one side of the paper only, and submitted in triplicate. The article should be preceded by a summary of not more than 200 words describing the entire paper. A list of key words is also required. The article title should be brief and stated on a separate page with the author's names and addresses. ------------------------------ Date: Wed, 22 Jan 86 16:47:34 PST From: admin%cogsci@BERKELEY.EDU (Cognitive Science Program) Subject: Seminar - Naive Physics: Knowledge in Pieces (UCB) BERKELEY COGNITIVE SCIENCE PROGRAM Spring 1986 Cognitive Science Seminar - IDS 237B Tuesday, January 28, 11:00 - 12:30 [NB. New Location] 2515 Tolman Hall Discussion: 12:30 - 1:30 [location TBA] ``Knowledge in Pieces'' Andrea A. diSessa Math Science and Technology, School of Education Abstract Naive Physics concerns expectations, descriptions and explanations about the way the physical world works that people seem spontaneously to develop through interaction with it. A recent upswing in interest in this area, particularly concern- ing the relation of naive physics to the learning of school physics, has yielded significant interesting data, but little in the way of a theoretical foundation. I would like to pro- vide a sketch of a developing theoretical frame together with many examples that illustrate it. In broad strokes, one sees a rich but rather shallow (in a sense I will define), loosely coupled knowledge system with elements that originate often as minimal abstractions of common phenomena. Rather than a "change of theory" or even a shift in content of the knowledge system, it seems that developing understanding of classroom physics may better be described in terms of a change in structure that includes selection and integration of naive knowledge elements into a system that is much less data-driven, less context dependent, more capable of "reliable" (in a technical sense) descriptions and explana- tions. In addition I would like to discuss some hypothetical changes at a systematic level that do look more like changes of theory or belief. Finally, I would like to consider the poten- tial application of this work to other domains of knowledge, and the relation to other perspectives on the problem of knowledge. ------------------------------ Date: Wed 22 Jan 86 17:32:26-PST From: Emma Pease Subject: Seminar - Term Rewriting, Theorem Proving, Logic Programming (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] CSLI ACTIVITIES FOR NEXT THURSDAY, January 30, 1986 2:15 p.m. CSLI Seminar Ventura Hall Term Rewriting Systems and Application to Automated Trailer Classroom Theorem Proving and Logic Programming Helene Kirchner (Kirchner@sri-ai) Term Rewriting Systems and Application to Automated Theorem Proving and Logic Programming Helene Kirchner Term rewriting systems are sets of rules (i.e. directed equations) used to compute equivalent terms in an equational theory. Term rewriting systems are required to be terminating and confluent in order to ensure that any computation terminates and does not depend on the choice of applied rules. Completion of term rewriting systems consists of building, from a set of non-directed equations, a confluent and terminating set of rules that has the same deductive power. After a brief description of these two notions, their application in two different domains are illustrated: - automated theorem proving in equational and first-order logic, - construction of interpretors for logic programming languages mixing relational and functional features. ------------------------------ Date: Thu 23 Jan 86 11:50:10-PST From: LANSKY@SRI-AI.ARPA Subject: Seminar - The Algebra of Time Intervals (SRI) THE ALGEBRA OF TIME INTERVALS Peter Ladkin (LADKIN@KESTREL) Kestrel Institute 11:00 AM, MONDAY, January 27 SRI International, Building E, Room EJ228 (new conference room) We build on work of James Allen (Maintaining Knowledge about Temporal Intervals, CACM Nov 1983), who suggested a calculus of time intervals. Allen's intervals are all convex (no gaps). We shall present a taxonomy of *natural* relations between non-convex [i.e., non-contiguous] intervals, and illustrate the expressiveness of this subclass, with examples from the domain of project management. In collaboration with Roger Maddux, we have new mathematical results concerning both Allen's calculus, and our own. We shall present as many of these as time permits. The talk represents work in progress. We are currently designing and implementing a time expert for the Refine system at Kestrel Institute, which will include the interval calculus. ------------------------------ Date: 22 Jan 86 09:18:06 EST From: Tom Subject: Seminar - Machine Learning and Economics (RU) [Forwarded from the Rutgers bboard by Laws@SRI-AI.] ML Colloquium talk Title: Market Traders: Intelligent Distributed Systems In an Open World Speaker: Prof. Spencer Star Laval University, Quebec Date: Friday, Jan 24 Time: 11 am Location: Hill 423 Professor Spencer Star is a computer scientist/economist who works on simulating economic markets. He will be spending the coming year on sabbatical at Rutgers to work on incorporating a machine learning component into his current market simulations. He is visiting now in order to meet the department and to get some feedback on his current research ideas on learning. Below is part of an abstract from his recent paper. [...] -Tom Mitchell Market Traders: Intelligent Distributed Systems In an Open World Although markets are at the heart of modern microeconomics, there has been relatively little attention paid to disequilibriun states and to the decision-making rules used by traders within markets. I am interested in the procedures that traders use to determine when and how much they will bid, how they adapt their behaviour to a changing market environment, and the effects of their adaptive behaviour on the market's disequilibrium path. This paper reports on research to study these questions with the aid of a computer program that represents a market with interacting and independent knowledge-based traders. The program is callled TRADER. In a series of experiments with TRADER I find that market efficiency requires a minimum number of intelligent traders with a capacity to learn, but when their knowledge is reflected in the market bids and asks, naive traders can enter the markets and sometimes do better than the expert traders. Moreover, the entrance of naive traders in a market that is already functioning efficiently does not degrade the market's performance. Since learning by independent agents appears to be a key element in understanding and using open systems, the focus of future research will be on studying learing and adaptive processes by intelligent agents in open systems. ------------------------------ Date: Tue, 28 Jan 86 15:41 EST From: Tim Finin Subject: Seminar - Semi-Applicative Programming (UPenn) SEMI-APPLICATIVE PROGRAMMING: AN EXAMPLE N.S Sridharan BBN Labs, AI Department, Cambridge MA 3pm Thursday, January 30, 1986 216 Moore, University of Pennsylvania Most current parallel programming languages are designed with a sequential programming language as the base language and have added constructs that allow parallel execution. We are experimenting with an applicative base language that has implicit parallelism everywhere, and then we introduce constructs that inhibit parallelism. The base language uses pure LISP as a foundation and blends in interesting features of Prolog and FP. Proper utilization of available machine resources is a crucial concern in functional programming. We advocate several techniques of controlling the behavior of functional programs without changing their meaning or functionality: program annotation with constructs that have benign side-effects, program transformation and adaptive scheduling. This combination yields us a semi-applicative programming language and an interesting programming methodology. In this talk we give some background information on our project, its aims and scope and report on work in progress in the area of parallel algorithms for context-free parsing. Starting with the specification of a context-free recognizer, we have been successful in deriving variants of the recognition algorithm of Cocke-Kasami-Younger. One version is the CKY algorithm in parallel. The second version includes a top-down predictor to limit the work done by the bottom-up recognizer. The third version uses a cost measure over derivations and produces minimal cost parses using a dynamic programming technique. In another line of development, we arrive at a parallel version of the Earley algorithm. ------------------------------ Date: Wed, 29 Jan 86 10:19:18 GMT From: Gideon Sahar Subject: Seminar - Integrating Syntax and Semantics (Edinburgh) EDINBURGH AI SEMINARS Date: 29th January l985 Time: 2.00 p.m. Place: Department of Artificial Intelligence Seminar Room - F10 80 South Bridge EDINBURGH. Dr. Ewan Klein, Centre for Cognitive Studies, University of Edinburgh will give a seminar entitled - "Integrating syntax and semantics : unification categorial grammar as a tool for a natural language processing". This talk will report on work carried out at the Centre for Cognitive Science By Henk Zeevat, Jo Calder and Ewan Klein as part of an ESPRIT project on natural language and graphics interfaces to a knowledge-base. In recent years there has been a surge of interest in syntactic parsers which exploit linguistically-motivated non-transformatinal grammar formalisms: instances are the GPSG chart parser at Hewlett-Packard, Palo Alto, and the PATR-II parser at SRI, Menlo Park. By contrast, progress in the development of tractable, truth-conditional semantic formalisms for parsing has lagged behind. Unification categorial grammar (UCG) employs three resources which significantly improve this situation. The first is Kamp's theory of Discourse Representation: this is essentially a first-order calculus which nevertheless provides a more elegant treatment of NL anaphora and quantification than standard first-order logic. Second, the grammar encodes both syntactic and semantic information in the same data structures, namely directed acyclic graphs, and manipulates them with same operation, namely unification. Third, the fundamental grammar rule is that of categorial grammar, namely functional application. Since the grammar objects contain both syntactic and semantic information, any rule application will simultaneously produce syntactic and semantic results. UCG translates readily into a PATR-like declarative formalism, for which Calder has written a Prolog implementation called PIMPLE. ------------------------------ Date: Tue, 28 Jan 86 15:41 EST From: Tim Finin Subject: Seminar - Feature Structures in Unification Grammars (UPenn) LOGICAL SPECIFICATIONS FOR FEATURE STRUCTURES IN UNIFICATION GRAMMARS William C. Rounds and Robert Kasper, University of Michigan 3pm Tuesday, February 4, 1986 216 Moore, University of Pennsylvania In this paper we show how to use a simple modal logic to give a complete axiomatization of disjunctively specified feature or record structures commonly used in unification-based grammar formalisms in computational linguistics. The logic was originally developed as a logic to explain the semantics of concurrency, so this is a radically different application. We prove a normal form result based on the idea of Nerode equivalence from finite automata theory, and we show that the satisfiability problem for our logical formulas is NP-complete. This last result is a little surprising since our formulas do not contain negation. Finally, we show how the unification problem for term-rewriting systems can be expressed as the satisfiability problem for our formulas. ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Jan 30 05:51:34 1986 Date: Thu, 30 Jan 86 05:51:31 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a022662; 30 Jan 86 0:32 EST Date: Wed 29 Jan 1986 20:54-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #17 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 30 Jan 86 05:31 EST AIList Digest Thursday, 30 Jan 1986 Volume 4 : Issue 17 Today's Topics: Queries - LISP-Based COBOL Parser or Compiler & AI Koans, AI Tools - Common Lisp for RT PC, Fiction - Pseudoscience Jargon in 2010, Policy - Theoretical CS, Games & Expert Systems - Hangman, Reports - MRS Manual & ISSCO Working Papers ---------------------------------------------------------------------- Date: Tue 28 Jan 86 21:06:57-CST From: John Hartman Subject: COBOL parser or compiler needed in Lisp environment Does anyone know of a Cobol parser or compiler that is written in LISP? (or PASCAL or will otherwise run on a LISP machine or DEC-20) [This is not a joke!] I'm working on a program understanding/program transformation system. The target language at the moment is Cobol because there are lots of unstructured Cobol programs and commercial systems that attempt to restructure them automatically. AI program understanding can improve the process. To demonstrate this, I need a Cobol parser, and would rather find one than build one. Does anyone have any pointers? Thanks, John Hartman ------------------------------ Date: Tue, 28 Jan 86 14:14:03 PST From: "Douglas J. Trainor" Subject: ai koans Has anyone heard any good ai koans over the past three years??? [][] Douglas J. Trainor [][] a pair of size 9 capri pants ------------------------------ Date: Sat, 25 Jan 1986 11:43 EST From: "Scott E. Fahlman" Subject: Common Lisp for RT PC In response to Nick Gall's query about Common Lisp for the RT PC: We at CMU have been working behind the scenes for some time to port our Spice/Accent operating system from the now-defunct Perq machine to the new IBM workstation, now dubbed the RT PC. As a part of that effort, we have ported the Spice Lisp implementation of Common Lisp, including the Hemlock editor. This port is mainly the work of Dave McDonald, with assists from Rob Maclachlan and Skef Wholey. Lisp and Hemlock are now running pretty well, with only a few finishing touches to be added and a lot of tuning to be done. There are still some holes in the Accent operating system for this machine, but we are working feverishly to patch them up. We are in the process of taking some benchmarks on the Lisp now. Early indications show the speed of the pre-tuning RT PC Lisp to be roughly in the ballpark (give or take a factor of two) of the Symbolics 3600 and the Sun 3, though you have to be careful with declarations and give up most of the runtime checking to go that fast. (Also necessary on other stock hardware like Sun, but not on Symbolics.) Please do not flood us with request for this system. The Lisp is not particularly to port over to any flavor of Unix, and Accent is not yet ready for use outside the friendly confines of CMU. At some point in the future, we may make the whole package available WITHOUT ANY SUPPORT, for users elsewhere who can tolerate unsupported university-quality software, but before we do that we will have to think very hard about how to minimize the hassles to all concerned. If we do that, I'll see that people reading this list hear about it. IBM has not announced any plans for introducing a supported Common Lisp product on the RT PC's officially sanctioned unix-based operating system. I believe that there would be great demand for such a product, but what their plans are I can't say. -- Scott ------------------------------ Date: Wed 22 Jan 86 18:57:57-PST From: Bill Poser Subject: Pseudoscience Jargon What exactly is a "call-seeking" computer? Anything to do with a "come-from" statement? ------------------------------ Date: Tue 28 Jan 86 07:26:53-PST From: Ted Markowitz Subject: 2010 and H-Mobius Loops I've checked my copy of 2010 and reviewed Dr. Chandra's explanation of HAL's paranoia. The H-Mobious Loop phenomenon often occurs in "autonomous GOAL-seeking programs". Not a bad lay-description of a program that got confused as to what to do next? --ted ------------------------------ Date: Mon, 27 Jan 86 12:52:16 CST From: veach%ukans.csnet@CSNET-RELAY.ARPA Subject: Comment on EATCS. Concerning the posting of the "European Association for Theoretical Computer Science" announcement in vol 4:6, I would like to make the following comments: 1) I agree whole-heartedly with the editorial comment which Ken Laws made at the end of vol 4:15 (except for his acquiescence to the Stanford professor's request that similar postings not be made in the future). 2) My reading of this Digest leads me to believe that the contributers and the readers as a whole, span a wide range of interests. This disparity of interests has been with AI since its begining and indeed is what makes AI what it is. We should recognize that with such a variety of research in AI (from vision to mathematical logic; design and fabrication of robotic limbs to analysis of cognitive processes; etc...) there is and should be a tremendous pool of resources which we individually draw from and collectively share. One does not have to look far to find common ground among researchers who delve into such distinct subjects (graph theory, predicate calculus, statistical analysis, etc.). In conclusion, rather than restrict the flow of information, I hope that as we see information which could benefit the community, we would share it. Glenn Veach (veach@ukans.csnet) ------------------------------ Date: 26 January 1986 1902-PST (Sunday) From: west@nprdc.arpa (Larry West) Subject: Theoretical CS vis-a`-vis AI In AIList V4 #15, Jean-Francois Rit said: ``I don't feel that a major problem for AI researchers is understanding the nature of computation, I think the AI point of view is much (maybe too much) broader or at least OPEN toward The "real" universe.'' I agree that those who are doing Expert Systems or similar kinds of programming need not worry too much about what a computation is nor how it is achieved. But those in Cognitive Science -- those interested in how brains do the things they do so well -- might well be interested in formalisms to help grasp the underlying processes of computation. On the other hand, my prejudice is that these are not yet understood in Theoretical Computer Science, either, and may not even be of interest to those in the field (TCS). Still, Parallel Distributed Processing or Connectionism seems to hold much promise for lower-level information processing, and perhaps higher-level as well, though that's harder to see at this point. See, e.g., Hinton & Anderson's *Parallel Models of Associative Memory* (Erlbaum, 1981), or Hinton's and Feldman's articles in the April 1985 BYTE magazine, or Minsky and Papert's *Perceptrons* or ... well, further references supplied on demand. My opinion would thus be not to exclude TCS out of hand, but don't go out of your way (KIL) looking for articles/ messages/seminar announcements relevant to AIList, either. Larry West (programmer) west@nprdc.ARPA UCSD Institute for Cognitive Science La Jolla, CA 92093 [That seems a fair summary of the feedback I've received, and of the general AIList screening policy. -- KIL] ------------------------------ Date: Wed, 29 Jan 86 22:19:21 EST From: Moorthy Subject: Hangman We have developed a computer program to play hangman by itself. Here the computer both guesses a word and tries to find what the guessed word is. This program is a variation of hangman available under unix 4.2. The program to guess the words is partly rule based (these rules are obtained by talking to an "expert") and partly searches the dictionary judiciously. The programs are written in C and uses system calls to AWK for searching various subsets of dictionary. We have tested the program fairly exhaustively and it plays reasonably well. If anyone is interested in knowing more about the program, you could contact moorthy@rpics. The developers of this program are Patrick Harubin, a junior in Computer Science at R.P.I and myself. Krishnamoorthy Department of Computer Science R.P.I., Troy NY 12181. ------------------------------ Date: Tue 28 Jan 86 12:28:22-PST From: Stuart Russell Subject: mrs manual The Compleat Guide to MRS is now available as a Stanford CS report, number STAN-CS-85-1080. To obtain a copy send mail to Kathy Berg (BERG@SCORE.ARPA) or write to her at Comp Sci Dept, Stanford, CA 94305. Stuart Russell (RUSSELL@SUMEX) ------------------------------ Date: Mon, 27 Jan 86 13:24:44 pst From: Mike Rosner Subject: ISSCO working papers Fondazione Dalle Molle Geneva ISSCO WORKING PAPERS No. 46 (1981) M Rosner Three Strategic Goals in Conversational Openings This paper tries to explain a short transcript of a conversational opening as completely as possible within the framework which takes conversational behaviour as defined by the operation of a sohisticated planning mechanism. It is argued that a critical role is played by the satifaction, for each participant, of three strategic goals relating to attention, identification, and greeting. Additional tactics for gaining information are also described as necessary to account for this transcript. No. 47 (1983) F di Primio & Th Christaller A Poor Man's Flavor System This paper is the result of an attempt to understand 'flavors', the object oriented programming system in Lispmachine Lisp. The authors argue that the basic principles of such systems are not easily accessible to the programming public, because papers on the subject rarely discuss concrete details. Accordingly, the authors' approach is pedagogical, and takes the form of a description of the evolution of their own flavor system. An appendix contains programming examples that are sufficienly detailed to enable an average Lisp programmer to build a flavor system, and experiment with the essential concepts of object-oriented programming. No. 48 (1984) Eric Wehrli A Government-Binding Parser for French This paper describes a parser for French based on an adaptation of Chomsky's Government and Binding theory. Reflecting the modular conception of GB-grammars, the parser consists of several modules corresponding to some of the subtheories of the grammar, such as X bar, binding, etc. Making an extensive use of lexical information and following strategies which attempt to take advantage of the basic properties of natural languages, this parser is powerful enough to produce all of the grammatical structures of sentences for a fairly substantial subset of French. At the same time, it is restricted enough to avoid a proliferation of alternative analyses, even with highly complex constructions. Particular attention has been paid to the problem of the grammatical interpretation of wh-phrases, to clitic constructions, as well as to the organisation and management of the lexicon. No 49 (1985) Patrick Shann AI Approaches to Machine Translation This paper examines some experimental AI systems that were specifically developed for machine translation (Wilks' Preference Semantics, the Yale projects, Salat and CONTRA). It concentrates on the different types of meaning representation used, and the nature of the knowledge used for the solution of difficult problems in MT. To explore particular AI approaches, the resolution of several types of ambiguity is discussed from the point of view of different systems. No. 50 (1985) Beat Buchmann & Susan Warwick Machine Translation: Pre-ALPAC history, Post-ALPAC overview This paper gives a historical overview of the field of Machine Translation (MT). The ALPAC report, the now well-known landmark in the history of MT, serves to delimit the two sections of this paper. The first section, Pre-ALPAC history, looks in some detail at the hopeful beginnings, the first euphoric developments, and the onsetting disillusionment in MT. The second section, Post-ALPAC overview, describes more recent developments on the basis of current prototype and commercial systems. It also reviews some of the basic theoretical and practical issues in the field. No 51 (1985) Rod Johnson & Mike Rosner Software Engineering for Machine Translation In this paper we discuss the desirable properties of a software environment for MT development, starting from the position that succesful MT depends on a coherent theory of translation. We maintain that such an environment should not just provide for the construction of instances of MT systems within some preconceived (and probably weak) theoretical framework, but should also offer tools for rapid implementation and evaluation of a variety of experimental theories. A discussion of some potentially interesting properties of theories of language and translation is followed by a description of a prototype software system which is designed to facilitate practical experimentation with such theories. Requests for these papers should be addressed to ISSCO working papers 54 route des Acacias 1227 Geneva Switzerland The price per paper, including air mail, is SFr 10 (or equivalent). Cheques should be made payable to "Institut Dalle Molle" ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Jan 30 05:51:23 1986 Date: Thu, 30 Jan 86 05:51:18 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a022792; 30 Jan 86 0:44 EST Date: Wed 29 Jan 1986 21:06-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #18 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 30 Jan 86 05:35 EST AIList Digest Thursday, 30 Jan 1986 Volume 4 : Issue 18 Today's Topics: Machine Learning - Self Knowledge & Perceptrons, Theory - Definition of Symbol ---------------------------------------------------------------------- Date: Thu 23 Jan 86 04:46:01-PST From: Bill Park Subject: Re: Speech Learning This stuff about Sejnowski's speaking reminds me eerily of the parts of Asimov's @i{I, Robot}, that tells how Susan Calvin's career began: >From "Robbie," where Susan is observing a little girl named Gloria trying to get some help during a tour of the Museum of Science and Industry ... "The Talking Robot was a @i{tour de force}, a thoroughly impractical device, possessing publicity value only. Once an hour, an escorted group stood before it and asked questions of the robot engineer in charge in careful whispers. Those the engineer decided were suitable for the robot's circuits were transmitted to the Talking Robot. "It was rather dull. It may be nice to know that the square of fourteen is one hundred ninety-six, that the temperature at the moment is 72 degrees Fahrenheit, and the air-pressure 30.02 inches of mercury, that the atomic weight of sodium is 23, but one doesn't really need a robot for that. One especially does not need an unwieldy, totaly immobile mass of wires and coils spreading over twnty-five square yards." ... ... ``There was an oily whir of gears and a mechanically timbred voice boomed out in words that lacked accent and intonation, `I -- am -- the -- robot -- that -- talks.' ``Gloria stared at it ruefully. It did talk, but the sound came from inside somewheres. There was no @i{face} to talk to. She said, `Can you help me, Mr. Robot, sir?' ``The Talking Robot was designed to answer questions, and only such questions as it could answer had ever been put to it. It was quite confident of its ability, therefore. `I -- can -- help -- you.' ``'Thank you, Mr. Robot, sir. Have you seen Robbie?' ``'Who -- is -- Robbie?' ```He's a robot, Mr. Robot, sir.'' She stretched to tip-toes. ``He's about so high, Mr. Robot, sir, only higher, and he's very nice. He's got a head, you know. I mean you haven't, but he has, Mr. Robot sir.' ``The Talking Robot had been left behind, `A -- robot?' ```Yes, Mr. Robot, sir. A robot just like you, except he can't talk, of course, and -- looks like a real person.' ```A -- robot -- like -- me?' ```Yes, Mr. Robot, sir.' ```To which the Talking Robot's only response was an erratic splutter and an occasional incoherent sound. The radical generalization offered it, i.e., its existence, not as a particular object, but as a member of a general group, was too much for it. Loyally, it tried to encompass the concept and half a dozen coils burnt out. Little warning signals were buzzing.' ``(The girl in her mid -teens left at that point. She had enough for her Physics-1 paper on `Practical Aspects of Robotics.' This paper was Susan Calvin's first of many on the subject.)'' ------------------------------ Date: 23-Jan-86 12:52:19-PST From: jbn@FORD-WDL1 Subject: Perceptrons-historical note Since Perceptron-type systems seem to be making a comeback, a historical note may be useful. The original Perceptron was developed in the early 1950s, and was a weighted-learning type scheme using electromechanical storage, with relay coils driving potentiometers through rachets being the basic learning mechanism. The original machine used to be on display at the Smithsonian's Museum of History and Technology, (now called the Museum of American History); it was a sizable unit, about the size of a VAX 11/780. But it is no longer on display; I've been checking with the Smithsonian. It has been moved out to their storage facility in Prince George's County, Maryland. It's not gone forever; the collection is rotated through the museum. If there's sufficient interest, they may put it back on display again. Another unit in the same collection has relevance to this digest; Parts of Reservisor, the first airline reservations system, built for American Airlines around 1954, are still on display; they have a ticket agent's terminal and the huge magnetic drum. Contrast this with Minksy's recent claims seen here that airline reservation systems were invented by someone at the MIT AI lab in the 1960s. John Nagle ------------------------------ Date: 22 Jan 86 14:41:45 EST From: Mark.Derthick@G.CS.CMU.EDU Subject: Re: What is a Symbol? This is a response to David Plaut's post (V4 #9) in which he maintains that connectionist systems can exhibit intelligent behavior and don't use symbols. He suggests that either he is wrong about one of these two points, or that the Physical Symbol System Hypothesis is wrong, and seeks a good definition of 'symbol. First, taking the PSSH seriously as subject to empirical confirmation requires that there be a precise definition of symbol. That is, symbol is not an undefined primitive for Cognitive Science, as point is for geometry. I claim no one has provided an adequate definition. Below is an admittedly inadequate attempt, together with particular examples for which the definition breaks down. 1) It seems that a symbol is foremost a formal entity. It is atomic, and owes its meaning to formal relationships it bears to other symbols. Any internal structure a [physical] symbol might posess is not relevant to its meaning. The only structures a symbol processor processes are symbol structures. 2) The processing of symbols requires an interpreter. The link between the physical symbols and their physical interrelationships on the one hand, and their meaning on the other, is provided by the interpreter. 3) Typically, a symbol processor can store a symbol in many physically distinct locations, and can make multiple copies of a symbol. For instance, in a Lisp blocks world program, many symbols for blocks will have copies of the symbol for table on their property lists. Many functionally identical memory locations are being used to store the symbols, and each copy is identical in the sense that it is physically the same bit pattern. I can't pin down what about the ability to copy symbols arbitrarily is necessary, but I think something important lurks here. The alternative to symbolic representations, analog (or direct) representations, do not lend themselves to copying so easily. For instance, on a map, distance relations between cities are encoded as distances between circles on paper. Many relations are represented, as in the case with the blocks world, but you can't make a copy of the circle representing a city. If it's not in the right place, it just won't represent that city. 4) Symbols are discrete. This point is where connectionist representations seem to diverge most from prototypical symbols. For instance, in Dave Touretzky's connectionist production system model (IJCAI 85), working memory elements are represented by patterns of activity over units. A particular element is judged to be present if a sufficiently large subset of the units representing the pattern for that element are on. Although he uses this thresholding technique to enable discrete answers to be given to the user, what is going on inside the machine is a continuum. One can take the pattern for (goal clear block1) and make a sequence of very fine grained changes until it becomes the pattern for (goal held block2). To show where my definition breaks down, consider numbers as represented in Lisp. I don't think they are symbols, but I'm not sure. First, functions such as ash and bit-test are highly representation dependent. Everybody knows that computers use two's complement binary representation for arithmetic. If they didn't, but used cons cells to build up numbers from set theory for instance, it would take all day to compute 3 ** 5. Computers really really have special purpose hardware to do arithmetic, and computer programmers, at least sometimes, think in terms of ALU's, not number theory, when they program. So the Lisp object 14 isn'sometimes t atomic, sometimes its really 1110. Its easy to see that the above argument is trying to expose numbers as existing at a lower level than real Lisp symbols. At the digital logic level, then, bits would be symbols, and the interpreter would be the adders and gates that implement the semantics of arithmetic. Similarly, it may be the case that connectionist system use symbols, but that they do not correspond to, eg working memory elements, but to some lower level object. So a definition of "symbol" must be relative to a point of view. With this in mind, it seems that confirmation of the Physical Symbol System Hypothesis turns on whether an intelligent agent must be a symbol processor, viewed from the knowledge level. If knowledge level concepts are represented as structured objects, and only indirectly as symbols at some lower level, I would take it as disconfirmation of the hypothesis. I welcome refinements to the above definition, and comments on whether Lisp numbers are symbols, or whether ALU bits are symbols. Mark Derthick mad@g.cs.cmu.edu ------------------------------ Date: 27 January 1986 1532-PST (Monday) From: hestenes@nprdc.arpa (Eric Hestenes) Subject: Re: What is a symbol? Article 125 of net.ai: In article <724@k.cs.cmu.edu>, dcp@k.cs.cmu.edu (David Plaut) writes: > It seems there are three ways out of this dilemma: > > (1) deny that connectionist systems are capable, in > principle, of "true" general intelligent action; > (2) reject the Physical Symbol System Hypothesis; or > (3) refine our notion of a symbol to encompass the operation > and behavior of connectionist systems. > > (1) seems difficult (but I suppose not impossible) to argue for, and since I > don't think AI is quite ready to agree to (2), I'm hoping for help with (3) > - Any suggestions? > David Plaut > (dcp@k.cs.cmu.edu) Symbol is unfortunately an abused word in AI. Symbol can be used in several senses, and when you mix them things seem illogical, even though they are not. Sense 1: A symbol is a token used to represent some aspect or element of the real world. Sense 2: A symbol is a chunk of knowledge / human memory that is of a certain character. ( e.g. predicates, with whole word or phrase size units ) While PDP / connectionist models may not appear to involve symbolic processes, meaning mental processes that operate on whole chunks of knowledge that consistute symbols they DO assign tokens as structures that represent some aspect or element. For instance, if a vision program takes a set of bits from a visual array as input, then at that point each of the bits are assigned a symbol and then a computation is performed upon the symbol. Given that pdp networks do have this primitive characterization in every situation, they fit Newell's definition of a Physical Symbol System [paraphrased as] "a broad class of systems capable of having and manipulating symbols, yet realizable in the physical world." The key is to realize that while the information that is assigned to a token can vary quite significantly, as in connectionist versus high level symbolic systems, the fact that a token has been assigned a value remains, and the manipulation of that newly created symbol is carried out in either kind of system. Many connectionists like to think of pdp systems as incorporating "microfeatures" or "sub-symbolic" knowledge. However, by this they do not mean that their microfeatures are not symbols themselves. Rather they are actively comparing themselves against traditional AI models that often insist on using a single token for a whole schema ( word, idea, concept, production ) rather than for the underlying mental structures that might characterize a word. A classical example is the ( now old ) natural language approach to thinking that parses phrases into trees of symbols. Not even the natural language people would contend that the contents of memory resembles that tree of symbols in terms of storage. In this case the knowledge that is significant to the program is encoded as a whole word. The connectionist might create a system that parses the very same sentences, with the only difference being how symbols are assigned and manipulated. In spite of their different approach, the connectionist version is still a physical symbol system in the sense of Newell. This point would be moot if one could create a connectionist machine that computed exactly the same function as the high-level machine, including manipulating high level symbols as whole. While both languages are Turing equivalent, one has yet to see a system that can compile a high-level programming language with a connectionist network. The problems with creating such a machine are many; however, it is entirely possible, if not probable. See the paper for a Turing <--> Symbol System proof. Reference: Newell, Allen. Physical Symbol Systems. Cognitive Science 4, 135-183 (1980). Copy me on replies. Eric Hestenes Institute for Cognitive Science, C-015 UC San Diego, La Jolla, CA 92093 arpanet: hestenes@nprdc.ARPA other: ucbvax!sdcsvax!sdics!hestenes or hestenes@sdics.UUCP ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Tue Feb 4 23:30:44 1986 Date: Tue, 4 Feb 86 23:30:33 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a006480; 3 Feb 86 13:40 EST Date: Mon 3 Feb 1986 10:13-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #19 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 4 Feb 86 00:52 EST AIList Digest Monday, 3 Feb 1986 Volume 4 : Issue 19 Today's Topics: Queries - Prolog for Compiler Writing & LISP Compilers & LISP Tutorial Source Code & Mathematical Structure of OOPL & Equation Solver, Binding - Supercomputer Center & Grenoble Labs, History - Airline Reservation Systems, Report - Calculus of Partially-Ordered Type Structures, Review - Technology Review Article ---------------------------------------------------------------------- Date: 31 Jan 86 08:02:00 EST From: "INFO1::ELDER" Reply-to: "INFO1::ELDER" Subject: Prolog for Compiler Writing Greg Elder ------------------------------ Date: Fri, 31 Jan 86 13:56:44 CST From: Al Gaspar Subject: LISP Compilers? A friend that doesn't have access to the net asked me to post this query. What brands of Common LISP would run best on a VAX 780 under UNIX Sys V.2? Any and all recommendations would be appreciated. Please reply to me directly as I don't subscribe to AILIST. If there are enough replies, I'll summarize to the net. Thanks in advance-- Al Gaspar USAMC ALMSA, ATTN: AMXAL-OW, Box 1578, St. Louis, MO 63188-1578 COMMERCIAL: (314) 263-5118 AUTOVON: 693-5118 seismo!gaspar@almsa-1.arpa ------------------------------ Date: 0 0 00:00:00 EST From: "Don Mcdougall" Reply-to: "Don Mcdougall" Subject: request for LISP source code [Interesting date on this message! -- KIL] I am teaching an AI course for the continuing education program at St. Mary's College in Southern Maryland. This is my first time teaching LISP and I would appreciate access to the source code for "project- sized" LISP programs or any other teaching aids or material. We are using the 2nd edition of both Winston's AI and Winston&Horne's LISP. I hate to ask for help, but we are pretty far from mainstream AI down here and my students and I all have full time jobs so any help we can get from the professional AI community would be greatly appreciated by all of us. Bob Woodruff Veda@paxrv-nes.arpa ------------------------------ Date: Fri, 31 Jan 86 8:58:50 EST From: "Srinivasan Krishnamurthy" <1438@NJIT-EIES.MAILNET> Subject: Mathematical Structure of OOPL I would like to hear about any definitive work on the mathematical structure of object oriented programming languages (eg. smalltalk). I am interested in the current status of the subject. Reference to a good review will be most helpful. Would also appreciate receiving papers or reports on the subject. My netaddress is : srini%NJIT-EIES.MAILNET@MIT-MULTICS.ARPA U.S Postal Address: Srinivasan Krishnamurthy COMSAT LABS, (NTD) RM:7142 22300 Comsat Drive Clarksburg, MD-20871 Tel: (301)428-4531(W) Thanks. Srini. ------------------------------ Date: 29 Jan 86 16:10:06 GMT From: ucdavis!lll-crg!topaz!harvard!cmcl2!philabs!dpb@ucbvax.berkeley. edu (Paul Benjamin) Subject: Re: Equation solver > I am looking for a program that can solve simple algebraic expressions > of the type: > > 10x - 15 = 5 > > This system would have the capability of SIMPLIFYING expressions, EXPANDING > expressions and SOLVING expressions (where possible). > Note that I am looking for simple solutions, I have no need of the extensive > capabilities of MACSYMA or some such thing. > It needs to work on fairly small (pdp-11, non-unix) machines. > It's purpose is to act a a simple but patient tutor in pre-algebra. > Consequently it must give hints, advice, etc. > Any help, pointers, suggestions, etc. from people is much appreciated. > > Dick Pierce > ucdavis!lll-crg!seismo!harvard!talcott!panda!teddy!rdp@ucbvax.berkeley.edu > Organization: GenRad, Inc., Concord, Mass. You may want to look at Sleeman's work, although it is more along the lines of simulating student's solutions to such tasks. It can be found with related work in "Intelligent Tutoring Systems", published by Academic Press in 1982. The editors are D. Sleeman and J. S. Brown. Good luck. Paul Benjamin ------------------------------ Date: 29 Jan 86 16:01:00 GMT From: pur-ee!uiucdcs!uiucdcsb!mozetic@ucbvax.berkeley.edu Subject: Re: Equation solver Some work on algebraic manipulation was done at the Edinburgh Univ. (Dept. of AI) by A.Bundy and others. I can give you few references: Bundy, Silver: Preparing Equations for Change in Unknown, IJCAI-81, and DAI research paper 159. Bundy, Sterling: Meta-level Inference in Algebra, DAI 164. Bundy, Welham: Using Meta-level Inference for Selective Application of Multiple Rewrite Rules in Algebraic Manipulation, Artificial Intelligence 16(2), 1981. You may also consult the book: Bundy: The Computer Modelling of Mathematical Reasoning, Academic Press, 1983. Good luck. ------------------------------ Date: Thu 30 Jan 86 15:26:58-CST From: CMP.BARC@R20.UTEXAS.EDU Subject: Re: Supercomputers and AI Sorry about the transposition of the zip code for UCSD. Maybe I can make up for it with the correct zip for GA Technologies. The mailing address they seem to giving out for Supercomputer Center communications is GA Technologies P.O. Box 85608 San Diego, CA 92138 Dallas Webster CMP.BARC@R20.UTexas.Edu ut-sally!batman!dallas ------------------------------ Date: 30 Jan 86 21:38:16 GMT From: "mcvax!vmucnam!imag!lifia!rit"@SEISMO.ARPA Subject: Grenoble labs Someone mailed me for enquiries about computer science Grenoble labs in response to an article in mod.ai. I lost his message, I'll answer him if he remails. Jean-Francois Rit Laboratoire d'Informatique Fondamentale et d'Intelligence Artificielle BP 68 38402 Saint-Martin d'Heres cedex Disclaimer: This is only my postal address! UUCP: ...{mcvax,vmucnam}!lifia!rit decvax!genrad!panda!talcott!harvard!seismo! mcvax!vmucnam!imag!lifia!rit@ucbvax.berkeley.edu ------------------------------ Date: Fri, 31 Jan 86 07:59:38 EST From: Alan Bawden Subject: Contrast Date: 23-Jan-86 12:52:19-PST From: jbn at FORD-WDL1 ... Contrast this with Minksy's recent claims seen here that airline reservation systems were invented by someone at the MIT AI lab in the 1960s. I decided to take a close look at this contrast. After searching through the recent archives, the only mention by Minsky of airline reservation systems that I can find is: And I'm pretty sure that the first practical airline reservation was designed by Danny Bobrow of the BBN AI group around 1966.! Now that I have refreshed my memory with what he actually said, I think the contrast is not quite as unflattering. Given the use of the adjective ``practical'', someone might even be able to make a case that he is right. ------------------------------ Date: Thu 30 Jan 86 15:15:53-CST From: AI.HASSAN@MCC.ARPA Subject: Calculus of Partially-Ordered Type Structures This message is an commmon answer to all those individuals (thanks for your interest) that have been asking me for copies of my Ph.D. Dissertation (A Lattice-Theoretic Approach to Computation Based on a Calculus of Partially-Ordered Type Structures). My thesis is being revised for publication as a book. I am out of copies the version I've been sending. You may: .write or call U.of Penn. CIS dpt. 215-898-8540 (Ph.D. 9/84) .write University Microfilms at Ann-Arbor, MI .get hold of one from a friend and ask a nice secretary to xerox it .steal one (no one will mind: it's a cheap value!). .or you can wait and bear with my slow work in translating a big Scribe mess into an even larger LaTeX mess(*)---send me another message in, oh, about 3 months. Hope that'll help. Thanks for your patience. Cheers, Hassan (*) By the way, any info of programs that do that is welcome! ------------------------------ Date: 31 Jan 86 17:18:00 GMT From: decvax!cca!ada-uts!richw@ucbvax.berkeley.edu Subject: Technology Review article Has anyone read the article about AI in the February issue of "Technology Review"? You can't miss it -- the cover says something like: "In 25 years, AI has still not lived up to its promises and there's no reason to think it ever will" (not a direct quote; I don't have the copy with me). General comments? -- Rich Wagner "Relax! They're just programs..." P.S. You might notice that about 10 pages into the issue, there's an ad for some AI system. I bet the advertisers were real pleased about the issue's contents... ------------------------------ Date: 3 Feb 86 14:25:24 GMT From: vax135!miles@ucbvax.berkeley.edu (Miles Murdocca) Subject: Re: Technology Review article The [Technology Review] article was written by the Dreyfuss brothers, who are famous for making bold statements that AI will never meet the expectations of the people who fund AI research. They make the claim that people do not learn to ride a bike by being told how to do it, but by a trial and error method that isn't represented symbolically. They use this argument and a few others such as the lack of a representation for emotions to support their view that AI researchers are wasting their sponsors' money by knowingly heading down dead-ends. As I recall ["Machine Learning", Michalski et al, Ch 1], there are two basic forms of learning: 'knowledge acquisition' and 'skill refinement'. The Dreyfuss duo seems to be using a skill refinement problem to refute the work going on in knowledge acquisition. The distinction between the two types of learning was recognized by AI researchers years ago, and I feel that the Dreyfuss two lack credibility since they fail to align their arguments with the taxonomy of the field. Miles Murdocca, 4G-538, AT&T Bell Laboratories, Crawfords Corner Rd, Holmdel, NJ, 07733, (201) 949-2504, ...{ihnp4}!vax135!miles ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Sat Feb 8 01:36:10 1986 Date: Sat, 8 Feb 86 01:36:04 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a009044; 7 Feb 86 13:47 EST Date: Fri 7 Feb 1986 10:29-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #20 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sat, 8 Feb 86 01:21 EST AIList Digest Friday, 7 Feb 1986 Volume 4 : Issue 20 Today's Topics: Seminars - Logics of Programmes (Edinburgh) & The Origins of Logic (UCB) & A Fuzzy Inference Engine (UPenn) & Intuitionistic Logic Programming Language (CMU) & Minsky and Dreyfus on AI (USantaClara), Conferences - Intelligent Robotic Systems & Cognitive Science Society ---------------------------------------------------------------------- Date: Wed, 5 Feb 86 12:20:32 GMT From: Gideon Sahar Subject: Seminar - Logics of Programmes (Edinburgh) EDINBURGH AI SEMINARS Date: 5th February 1986 Time: 2pm Place: Department of Artificial Intelligence Forrest Hill Seminar Room Dr. D.C. McCarty, Center for Cognitive Sciences, University of Edinburgh, will give a seminar entitled - `Logics of Programmes: Some Constructive Comments'. The talk will give an introduction to and overview of the applications of constructive logic to programme verification. Three topics will be of interest: the idea that functional interpretations of constructive set theory are `high level' compilers; the relations between constructive logic and Reynolds' `specification logic'; and the use of a constructive meta theory in giving completeness proofs for hoare-style logics. We will pre-suppose only a basic knowledge of mathematical logic; the requisite technicalities from constructive logic and programme verification will be explained in the talk. ------------------------------ Date: Wed, 5 Feb 86 15:40:45 PST From: admin%cogsci@berkeley.edu (Cognitive Science Program) Subject: Seminar - The Origins of Logic (UCB) BERKELEY COGNITIVE SCIENCE PROGRAM Cognitive Science Seminar - IDS 237B Tuesday, February 11, 11:00 - 12:30 2515 Tolman Hall Discussion: 12:30 - 1:30 3105 Tolman (Beach Room) ``The Origins of Logic'' Jonas Langer Department of Psychology, UCB I will try to show that logical cognition (1) originates during the first year of infancy and (2) begins to be represen- tational during the second year of infancy. This includes pro- posing some of its initial structural features. These claims imply that (a) a symbolic language is not necessary for the origins of logical cognition and (b) that ordinary language is not necessary for its initial representational development. Supporting data will be drawn from J. Langer, The Origins of Logic: Six to Twelve Months, Academic Press, 1980, and The Ori- gins of Logic: One to Two Years, Academic Press, 1986. ------------------------------ Date: Wed, 5 Feb 86 12:10 EST From: Tim Finin Subject: Seminar - A Fuzzy Inference Engine (UPenn) A VLSI IMPLEMENTATION OF FUZZY INFERENCE ENGINE: TOWARD AN EXPERT SYSTEM ON A CHIP Hiroyuki Watanabe, AT&T Bell Laboratories, Holmdel, New Jersey 3pm Tuesday, February 11, 1986 216 Moore, University of Pennsylvania This talk describes a VLSI implementation of an inference mechanism to cope with uncertainty and to perform approximate reasoning. Some details of VLSI layout design is presented. Design of an inference mechanism is based on the "max-min operation" of fuzzy set theory for an effective and real-time use. This inference mechanism can handle imprecise and uncertain knowledge; therefore, it can represent human expert knowledge and simulate his/her reasoning processes. An inference mechanism has been realized by using custom CMOS technology which emphasizes simplicity, extensibility and efficiency. For example, all rules are executed in parallel for efficiency. Result of preliminary tests indicates that the inference engine can perform approximately 80,000 Fuzzy Logical Inferences Per Second (FLIPS). This chip is designed for the application of rule-based expert system paradigm in real-time control. Potential application of such inference engine is real-time decision-making in the area of command and control, intelligent robotic system and chemical process control. ------------------------------ Date: 5 February 1986 1529-EST From: Theona Stefanis@A.CS.CMU.EDU Subject: Seminar - Intuitionistic Logic Programming Language (CMU) JOINT LOGIC COLLOQUIUM (CMU, U of Pgh) Dale Miller CIS Department, University of Pennsylvania Date: Thursday February 13 Time: 3 pm Place: 4605 Wean Hall A Logic Programming Language Based on Intuitionistic Higher-Order Logic. Dale Miller CIS Department, University of Pennsylvania In this talk, we present a programming language whose operational semantics can be understood as searching for proofs with in a subset of intuitionistic higher-order logic. Kripke-models over a universe of higher-order terms provide a model theoretic semantics for our programs. Such models can be computed as least fix points. This logical language is a natural extension to Horn clause logic and the programming language based on it has many features not available in simple Horn clause based programming languages. In particular, this programming language can manipulate higher-order functions in a manner similar to many functional programming languages. An interesting notion of parametric modules is also available by virtue of the behavior of implication within an intuitionistic logic. An interpreter for this language must perform unification of higher-order terms. If time permits, we illustrate how this feature makes possible the very clean implementation of certain kinds of program transformation algorithms. ------------------------------ Date: Tue 4 Feb 86 14:45:07-PST From: HOFFMANN@SRI-KL.ARPA Subject: Seminars - Minsky and Dreyfus on AI (USantaClara) Two talks on AI at Mayer Theater, University of Santa Clara; both talks are free, first come, first served. Marvin Minsky - "Intelligence and Creativity" Monday, February 10th, 8:00 PM Hubert Dreyfus - "Limits of AI" Thursday, February 20th, 8:00 PM For additional information call Mayer Theater, (408) 554-4015 ------------------------------ Date: 31 Jan 1986 10:40:22 EST From: Martin Marietta Subject: Conference - Intelligent Robotic Systems SPIE's Symposium on Advances in Intelligent Robotics Systems, including o Intelligent Robots and Computer Vision Conference o Mobile Robots Conference o Optics, Illumination, and Image Sensing for Machine Vision o Space Station Automation o Automated Inspection and Measurement The Conference(s) take place October 26-31, 1986, at the Hyatt Regency in Cambridge, MA. General Chairman is David Casasent, Carnegie-Mellon University. Abstract due date: 15 April (200-300 word abstract) Manuscript due date: 29 September For author application or further information, contact SPIE Technical Program Committee PO Box 10 Bellingham, WA 98227-0010 (206) 676-3290 ------------------------------ Date: Mon, 3 Feb 86 07:52:16 pst From: gluck@SU-PSYCH (Mark Gluck) Subject: Conference - Cognitive Science Society 8th Annual Cognitive Science Society Conference will be held at U. Mass/Amherst from August 15th to 17th. Submission Deadline: March 14, 1986 to: Charles Clifton Department of Psychology U. Mass. Amherst, MA 01003 Include: author's name, address, and telephone number up to four keywords four copies of abstract (100-250 words) four copies of paper (4K words for presentation; 2K for poster) ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Sat Feb 8 01:35:59 1986 Date: Sat, 8 Feb 86 01:35:54 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a009608; 7 Feb 86 14:46 EST Date: Fri 7 Feb 1986 11:19-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #21 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sat, 8 Feb 86 01:22 EST AIList Digest Friday, 7 Feb 1986 Volume 4 : Issue 21 Today's Topics: Queries - ISIS & BIB-Format AI References, Logic Programming - Prolog for Compiler Writing, Expert Systems & Reports - MRS, Theory - Dreyfus Article in Technology Review ---------------------------------------------------------------------- Date: Tue, 04 Feb 86 16:19:44 cet From: WMORTENS%ESTEC.BITNET@WISCVM.WISC.EDU Subject: Query -- ISIS From: Uffe K. Mortensen ESA ( The European Space Agency ) Does anybody here know what 'ISIS' is ? I have been told it is a commercial package for planning/scheduling problems, but I would like to have more detailed information ( vendor, etc ). -- Uffe. ------------------------------ Date: 3 Feb 86 18:59:00 GMT From: pur-ee!uiucdcs!uiucdcsb!mklein@ucbvax.berkeley.edu Subject: Bib Format AI References Request I am interested in getting references in bib format for the following topics, ordered with the stuff most important to me now on top: * distributed problem solving * machine learning * planning * vision If you have any references available, please send them to: mklein@uiucdcsb Thanks! Mark Klein ------------------------------ Date: 06 Feb 86 09:37:54 +1100 (Thu) From: Isaac Balbin Subject: Re: Prolog for Compiler Writing I have not added compilers for prolog written in prolog, nor stuff on compiling techniques for prolog. %A H. Derby %T Using Logic Programming for Compiling APL %R Technical Report 84-5134 %I Department of Computer Science %I California Institute of Technology %C Los Angeles, California %D 1984 %A G.A. Edgar %T A Compiler Written in Prolog %J Dr. Dobbs Journal %D May, 1985 %A Harald Ganzinger %A Michael Hanus %T Modular Logic Programming of Compilers %J Proceedings of the 2nd IEEE International Symposium on Logic Programming %C Boston, USA %D July, 1985 %A D.H.D. Warren %T Logic for Compiler Writing %J Software Practice and Experience %V 10 %N 1 %P 97-125 %D 1980 %O Also available as DAI Research Paper 44 from Department of Artificial Intelligence, University of Edinburgh Isaac Balbin =========================== UUCP: {seismo,mcvax,ukc,ubc-vision}!munnari!isaac ARPA: isaac%munnari.oz@seismo.css.gov CSNET: isaac%munnari.oz@australia ------------------------------ Date: Tue, 4 Feb 86 15:46:28 EST From: munnari!goanna.oz!wjb@seismo.CSS.GOV (Warwick Bolam) Subject: Correction to correction to name of MRS >From: veach%ukans.csnet@CSNET-RELAY.ARPA > >In a recent issue the full name MRS was incorrectly reported. > > MRS = "Modifiable Representation System" > > (source - "MRS Manual", Michael R. Genesereth, et. al. > 1980, Stanford Heuristic Programming Project) In the bibliography of the paper "Partial Programs", Michael R Genesereth, 1984, Stanford HPP: M. R. Genesereth, R. Greiner, D. E. Smith: "MRS - A Meta-Level Representation System", HPP-83-27, Stanford University HPP, 1983. Is there anyone who REALLY knows what MRS stands for? I have a number of MRS documents and NONE of them says "MRS stand for ..." Warwick Bolam, Computing Dept, Royal Melbourne Institute of Technology, Melbourne, Victoria, Australia. ------------------------------ Date: Mon 3 Feb 86 17:15:26-PST From: Stuart Russell Subject: MRS manual I have been asked to point out to those who have requested copies of the MRS manual, or who intend to do so, that a nominal fee of $6.00 (plus tax if in CA) is suggested. At .004 cents per exqusitely chosen word, it's a bargain. Stuart Russell (RUSSELL@SUMEX) ------------------------------ Date: 4 Feb 86 13:30:31 GMT From: Bob Stine Subject: re: Dreyfus article "Why Computers May Never Think Like People," a recent diatribe by the brothers Dreyfus, has several problems. First and foremost is that AI research is implicitly identified as the development of rule-based systems. All of the well known limitations of rule-based systems are inappropriately attributed to AI research as a whole. There is a deeper problem with the article, that perhaps springs from a misguided humanism. The article claims that machines will never duplicate human performance in cognitive tasks, because humans have "intuition." These passages would read very much the same if 'magic' were substituted for 'intuition' - "Human begins have a magic intelligence that reasoning machines simply cannot match." "... a boxer seems to recognize the moment to begin an attack... ... the boxer is using his magic". The Dreyfus brothers claim that they are not "Luddites," that they are not opposed to technology per se, but just to wasting time and money on AI research. The basis of their position is that some aspect of human intelligence is inherently beyond human comprehension. There certainly are things that humans will never know. But no one thing is inherently unknowable. ------------------------------ Date: Tue, 4 Feb 86 09:58:53 CST From: sandon@ai.wisc.edu (Pete Sandon) Subject: Knowledge Aquisition -vs- Skill Refinement This is not to defend the Dreyfus brothers, since I have yet to read their books. On the other hand, I think they make a good point, though with a bad example, in emphasizing learning as a process of refinement. The example related in Miles Murdocca's submission is that of learning to ride a bike through trial and error. The reason the example is a bad one, is that it fits into the category of skill refinement as AI researchers would use the term. This leads to the argument that Dreyfus and Dreyfus are missing the critical distinction between knowledge acquisition and skill refinement. My feeling is that too much is made of this distinction. Had the example been one of learning to distinguish fruits from vegetables, or one of learning the symptoms of a class of diseases well enough to diagnose them, this argument would not have arisen. Clearly these involve knowledge acquisition rather than skill refinement. And yet, it could be argued, and perhaps is argued by the Dreyfus's, that what the AI researchers consider to be knowledge acquisition should be just as much a refinement process guided by trial and error as learning to ride a bike. Whereas AI considers concept formation to occur as the acquisition of discrete chunks of knowledge, an alternative is to use the gradual acquisition of evidence to support one concept definition over another, in a manner similar to skill refinement. Of course, if this criticism of AI is correct, AI has already answered it. The use of connectionist models, and the corresponding learning mechanisms currently being studied, provide just the sort of cognitive models that support this refinement type of learning through trial and error. --Pete Sandon ------------------------------ Date: 3 Feb 86 17:24:42 GMT From: nike!caip!im4u!milano!pcook@ucbvax.berkeley.edu Subject: Re: Technology Review article In article <7500002@ada-uts.UUCP>, richw@ada-uts.UUCP writes: > > Has anyone read the article about AI in the February issue of > "Technology Review"? You can't miss it -- the cover says something > like: "In 25 years, AI has still not lived up to its promises and > there's no reason to think it ever will" (not a direct quote; I don't > have the copy with me). General comments? > This article is a plug for a book and use of a current topic to get back at the AI community for an imagined snub. Hubert Dreyfus was stood up by John McCarthy of Stanford at a debate on a third echelon public tv station in the bay area, and is still mad. First, the premise: AI, expert systems, and knowledge-rule based systems have been overly optimistic in their promises and stand short of delivered results. Probably true, but many of the systems, once implemented, lose their mystical qualities, and look a lot like other computer applications. It's the ones that are in the buliding process which seem to present extravagant claims. As presented, however, the article is a shrill cry rather than a reasoned response. It leans heavily on proof by intense assertion. As a pilot I find examples which range from dubious to incorrect. As a scientist I object to the gee whiz Reader's Digest tone. As a retired Air Force Officer I object to the position that the commander's common sense is the ideal form of combat decision making. And as a philosopher (albeit not expert) I object to the muddy intellectual approach, rife with questionable presuppositions, faulty dilemmas, and illogical conclusions. I agree that the topic is worthy of discussion- our work to realize the potential of computers must not degenerate into a fad which will fade from the scene. But I object to a diatribe where advances in the field are dismissed as trivial because current systems do not equal human performance. -- ...Pete Peter G. Cook Lt. Colonel pcook@mcc.arpa Liaison, Motorola, Inc. USAFR(Ret) ut-sally!im4u!milano!pcook MCC-Software Technology Program 512-834-3348 9430 Research Blvd. Suite 200 Austin, Texas 78759 [There are, of course, two sides to the McCarthy incident. As I recall from an old SU-BBoard message, McCarthy had agreed to an interview under the impression that he would be on the program alone. At the last moment it was mentioned that Dreyfus had also been invited. Viewing this as "ambush journalism" -- my words -- McCarthy declined to participate in the impromptu debate. No doubt the station was just trying to schedule a lively evening, but they should have checked with McCarthy and given him time to prepare. He and Dreyfus have sufficient visibility that a poorly stated remark, on >>any<< radio station, could affect the future of AI funding. -- KIL] ------------------------------ Date: Thu, 6 Feb 86 08:46 EST From: Ken Haase Subject: Re: Technology Review article Date: 3 Feb 86 14:25:24 GMT From: vax135!miles@ucbvax.berkeley.edu (Miles Murdocca) Subject: Re: Technology Review article To: AIList@SRI-AI The [Technology Review] article was written by the Dreyfuss brothers, who are famous for making bold statements that AI will never meet the expectations of the people who fund AI research. They make the claim that people do not learn to ride a bike by being told how to do it, but by a trial and error method that isn't represented symbolically. They use this argument and a few others such as the lack of a representation for emotions to support their view that AI researchers are wasting their sponsors' money by knowingly heading down dead-ends. I don't think the Dreyfus brothers accuse AI researches of knowingly heading down dead-ends. They just claim that most of ``what people do'' cannot be captured by the ``abstracted representations'' of nearly all current AI research. I don't agree with this claim, but can't deny that we (in AI) may be all wrong about our central hypothesis. We just have to make our hypothesis clear and explicit. I think that most high level intellectual processes have effective symbolic representations (and I'm working to find out what such representations might be). That is an explicit hypothesis of my research. On the other hand, I do not think that there is anything like a symbolic representation of ``how to ride a bike''. What happens in such cases is that our intellect ``trains'' the animal that is the rest of us to ride the bicycle. As I recall ["Machine Learning", Michalski et al, Ch 1], there are two basic forms of learning: 'knowledge acquisition' and 'skill refinement'. The Dreyfuss duo seems to be using a skill refinement problem to refute the work going on in knowledge acquisition. The distinction between the two types of learning was recognized by AI researchers years ago, and I feel that the Dreyfuss two lack credibility since they fail to align their arguments with the taxonomy of the field. The alchemists could have made the same argument against arguments for the periodic table; what the Dreyfus brothers are arguing for is the need for just such a ``paradigm shift'' in cognitive science. The fact that this shift will disrupt the foundations of most current AI technology (most of which is not well proven anyway) should not effect scientific judgements at all (though, pessimistically, it certainly will). In any case, the dichotomy between skill refinement and knowledge acquisition is even suspect; outside of rote learning of facts, most gained knowledge is gained by appropriating the knowledge as skills (in a broad sense of skills, which includes responses, perceptual skills, etc). Ken ------------------------------ End of AIList Digest ******************** From vtvax3::foxea@vtcs1 Sat Feb 8 14:51:53 1986 Date: Sat, 8 Feb 86 14:51:47 est From: vtvax3::foxea@vtcs1.VT To: fox@vtopus (VTCS1::FOX) Subject: ailist etc. game msg Status: RO From: ARPA%"MEGIDDO@IBM-SJ.ARPA" 7-FEB-1986 18:35 To: FOXEA Subj: Received: From WISCVM(SMTPUSR1) by VTVAX3 with RSCS id 4205 for FOXEA@VTVAX3; Fri, 7-FEB-1986 18:34 EST Received: from BRL-AOS.ARPA by wiscvm.wisc.edu on 02/07/86 at 17:20:00 CST Received: from ibm-sj.arpa by AOS.BRL.ARPA id a019630; 7 Feb 86 14:15 EST Date: 7 Feb 86 10:08:45 PST From: MEGIDDO@IBM-SJ.ARPA To: THEORY@wisc-rsr0ch, AILIST@sri-ai.ARPA, ARMS-D@mit-mc.ARPA, ARPANET-BBOARDS@mit-mc.ARPA, EVOLUTION@kestrel.ARPA, MsgGroup@BRL.ARPA, NA@su-score.ARPA, PHIL-SCI@mit-mc.ARPA, POLI-SCI@rutgers.ARPA, PROLOG@su-score.ARPA, MEGIDDO@ibm-sj.ARPA First Announcement of a COMPUTER PROGRAMS TOURNAMENT (of the Prisoners' Dilemma game) 1. INTRODUCTION _______________ This is a first announcement of a tournament for computer programs, playing the famous Prisoners' Dilemma game. Detailed instructions and some background information are provided below. The tournament is organized for the purpose of research and no prizes are offered. It is intended however that the results and winners' names will be published with permission from the persons involved. One of the goals is to see what will happen during a SEQUENCE of tournaments in which information about the participating programs will be released, so that participants will be able to revise their programs. The tournament is open to everyone. However, notice the warnings below. If you have access to electronic mail then you can participate by submitting a FORTRAN program according to the instructions below. By doing so you will also release and waive all your copyright rights and any other intellectual property rights to your program. It will also be assumed that you have not violated any rights of any third party. This announcement also includes some programs that will help you prepare for the tournament. 2. BACKGROUND _____________ The so-called prisoners' dilemma game has drawn the attention of researchers from many fields: psychology, economics, political science, philosophy, biology, and mathematics. Computer scientists are also interested in this game in the context of fundamentals of distributed systems. The game is simple to describe, does not require much skill and is yet extremely interesting from both the theoretical and practical points of view. By the (one-shot) Prisoners' Dilemma game we refer to a game as follows. The game is played by two players with symmetric roles. Each has to choose (independently of the other) between playing action C ("cooperate") or action D ("defect"). The scores to the two players, corresponding to the four possible combinations of choices of actions, are as shown in the following table: Player 2 C D --------------- | 3 | 4 | C | | | | 3 | 0 | Player 1 |-------|-------| | 0 | 1 | D | | | | 4 | 1 | --------------- Thus, both players score 3 if both play C. Both score 1 if both play D. If one plays C and the other one plays D then the one who plays C scores 0 while the other one scores 4. The prisoner's dilemma game has been the subject of many experiments. A tournament was organized several years ago by R. Axelrod who later published a book on it under the title "The evolution of cooperation" (Basic Books, Inc., New York, 1984). Following is some discussion for the benefit of readers who are not familiar with the fundamental considerations of how to play the game. One should be careful to distinguish the one-shot game from the REPEATED game in which the (one-shot) game is played many times, and after each round both players are informed of each other's actions. Furthermore, one should distinguish between the infinitely repeated game and the finitely repeated one. These seem to be quite different from the point of view of equilibrium. An equilibrium in a 2-person game is a pair (S1,S2) of strategies (one for each player) such that, given that player i (i=1,2) is playing Si , the other player, j=3-i, scores the maximum if he plays Sj . We are interested here in the finitely repeated game where the number of rounds is known in advance. We first consider the one-shot game. The analysis of the one-shot game is obvious. Each of the players realizes that no matter what his opponent does, it is always better for him to play D rather than C. Thus, under a very weak assumption of rationality (namely, players do not choose actions that are strictly dominated by other actions), the pair of actions (D,D) remains the only rational choice. The resulting score of (1,1) is inferior to (3,3), which is possible if the choices are (C,C), and this is the source of the "dilemma". To get some insight into the more general case, consider first the 2-round game. After the first round (in which the players choose independently C or D) each player is informed of the choice of the other one and then, once again, the players choose independently C or D. In this game each player has EIGHT strategies that can be coded in the form XYZ where each of X,Y and Z equals either C or D. The interpretation of this notation is as follows. (1) Play X in round 1. (2) In round 2, play Y if the opponent played C and play Z if the opponent played D. It is easy to verify that any strategy XYZ is strictly dominated by XDD (that is, regardless of what was done in round 1, and regardless of what the opponent does in round 2, it is better to play D rather than C in round 2. However, there is no domination relation between the strategies CDD and DDD: if player 2 plays DDD then player 1 is better off playing DDD rather than CDD, whereas if player 2 plays DDC, player 1 is better off playing CDD rather than DDD. Of course, strategy DDC for player 2 is dominated by DDD, but in order for player 1 to deduce that player 2 will not play DDC, he has to assume that player 2 is capable of discovering this domination. Under such an assumption player 1 can eliminate 2's DDC. Thus, if both players are "rational" they are left only with strategy DDD as a reasonable choice. A similar process of repeatedly eliminating dominated strategies applies to the general N-round game. It is dominant for both players to defect in the last round. Therefore (after we drop all strategies that play C in the last round), it becomes dominant to defect in round N-1, and so on. This eventually leaves both players only with the strategy of always playing D. The winner in both tournaments run by R. Axelrod was the simple strategy called "Tit-for-Tat". It starts by playing C and in round i+1 plays whatever the opponent played in round i. It seems like a very good strategy for playing the repeated dilemma for an indefinite number of rounds. In the N-round game it is obvious that an improvement over Tit- for-Tat would be to play Tit-for-Tat except for the last round in which the optimal play is always to defect. 3. HOW TO PARTICIPATE IN THE TOURNAMENT? ________________________________________ If you think you understand the dilemma quite well and would like to participate in this tournament then please act according to the following instructions: 1. Design a strategy of how to play the game when the number of rounds is known in advance. The strategy should specify what to do in round 1 and at any point of the game, knowing what has been done so far and the number of rounds left, specify what to do in the next round. 2. Write a FORTRAN subroutine with the following specifications. Give it a six-letter name, for example, the first four letters of your last name followed by two initials. Suppose you picked the name JONERJ for your subroutine. Then the first line of your program should look as follows. SUBROUTINE JONERJ (N,J,I,M) The arguments are defined as follows. N - This is the total number of rounds to be played. Whenever your program is called it is told the total number of rounds and this will not change during a single game. J - This is the serial number of the round you are supposed to play in the current call. I - When J is greater than 1, this argument tells you what your opponent has played in the previous round. If I=1 it means your opponent has played C. If J=2 then he played D. Any other value is an error. M - This is what you return as your play in the current round. M=1 means you play C. M=2 means you play D. Any other value will result in an error. Your subroutine may compute anything you wish. In particular, it may keep track of the entire history of a single (N-round) game. However, it will not be able to record past games against any opponent since it will be unloaded at the end of a single N-round game. Please be reasonable with respect to the space and time you intend your program to use. Unreasonable programs will have to be dropped from the tournament at the discretion of the organizers. Also, if your program ever returns a faulty play, that is, it returns an M which is neither 1 nor 2, then it will be dropped from the tournament automatically. 3. Fill in the following information (to be transmitted only by electronic mail): NAME:_____________________________________________________________ AFFILIATION:______________________________________________________ STREET:___________________________________________________________ CITY:___________________ STATE:_____________ Zip:_______________ COUNTRY:__________________________________________________________ TELEPHONE:________________________________________________________ ELECTRONIC MAIL ADDRESS:__________________________________________ 4. Important notice! _________________________________________________________ | By sending your program to any one of the following | | addresses you agree to waive and release, to the extent | | permitted by law, all your copyright rights and other | | intellectual property rights in your computer program. | | You also warrant that no portion of your program or its | | use or distribution, violates or is protected by any | | copyright or other intellectual property right of any | | third party. You also warrant you have the right to, | | and hereby do, grant to IBM a royalty-free license to | | use your program. If any contestant is a minor under | | the laws of the state in which contestant resides, at | | least one of the contestant's parents should sign this | | warranty and license. IBM may elect to publish the | | results of the contest; names of participants or their | | submissions will not be published without the written | | approval and signature of the individual authors. | |_________________________________________________________| Please transmit your program by March 31, 1986, along with the filled questionnaire to one of the following addresses: CSNET or ARPANET: megiddo@ibm-sj VNET or BITNET : megiddo at almvma 4. TRAINING PROGRAM ___________________ For your convenience, we include here an interactive program that lets you play the game with another "player". While playing this interactive program please remember that your goal is actually to SCORE high and not necessarily to BEAT the other player. In the tournament, your ability to affect the player's total score is limited since he plays against many other players besides you. Thus you will benefit if you will create "confidence" so that both of you end up playing C very often. You have the option of either playing yourself or using the subroutine that represents you. If you use a subroutine then you have to name it MINE and follow the instructions in Section 3. Simply append it the following program. It is advised that you use this option to test your own program before submitting it to the tournament. ------------------------------------------------------------------------- INTEGER SCORE,SCORE2,CH1,CH2,PRE1,PRE2,CC,DD,CD,DC C DATA CC,DD,CD,DC/3,1,0,4/ 20 SCORE = 0 SCORE2 = 0 PRE1=1 PRE2=1 WRITE(6,102) 102 FORMAT(' ENTER NUMBER OF ROUNDS YOU WISH TO PLAY (0=END)') 103 FORMAT (I6) READ (5,*) NR IF (NR.LE.0) STOP 118 FORMAT(' WILL YOU (1) PLAY OR WILL YOUR SUBROUTINE (2) DO? (1/2)') 430 WRITE (6,118) READ (5,*) II IF (II.EQ.2) GO TO 420 IF (II.NE.1) GO TO 430 420 DO 30 JR = 1, NR 104 FORMAT(' ROUND NO.',I6,' OF',I6,' ROUNDS. PLEASE ENTER 1 OR 2') IF (II.EQ.2) GO TO 440 WRITE (6,104) JR,NR 40 CONTINUE READ (5,*) CH1 GO TO 450 440 CALL MINE(NR,JR,PRE2,CH1) IF ((CH1-1)*(CH1-2)) 470,71,470 470 WRITE (6,117) 117 FORMAT (' YOUR SUBROUTINE RETURNED A FAULTY PLAY') GO TO 20 450 IF ((CH1-1)*(CH1-2)) 70,71,70 70 IF (CH1.EQ.0) GO TO 20 105 FORMAT(' PLEASE ENTER EITHER 1 OR 2 . (0=END)') WRITE (6,105) GO TO 40 71 IF (JR-1) 220,220,230 220 CH2 = 1 IF (NR.EQ.1) CH2 = 2 GO TO 300 230 IF (JR-NR) 250,260,260 250 CH2 = PRE1 GO TO 300 260 CH2 = 2 107 FORMAT(' PLAY WAS: YOU=',I3,' OPPONENT=',I3) 300 WRITE(6,107) CH1,CH2 IF (CH1-1) 110,110,120 110 IF (CH2-1) 130,130,140 130 SCORE = SCORE + CC SCORE2 = SCORE2 + CC GO TO 35 140 SCORE = SCORE + CD SCORE2 = SCORE2 + DC GO TO 35 120 IF (CH2-1) 150,150,160 150 SCORE = SCORE + DC SCORE2 = SCORE2 + CD GO TO 35 160 SCORE = SCORE + DD SCORE2 = SCORE2 + DD 35 WRITE (6,106) SCORE,SCORE2 106 FORMAT (' NEW TOTAL SCORE: YOU=',I5,' OPPONENT=',I5) PRE1=CH1 PRE2=CH2 30 CONTINUE GO TO 20 END ------------------------------------------------------------------------- 5. SAMPLE PROGRAMS __________________ For your convenience we include here copies of two sample programs. The first subroutine, called TIFRTA, plays Tit-for-Tat (see Section 2) except that it always defects in the last round. The second, called GRIM, starts playing C but switches to D the first time th opponent has played D. It also always defects in the last round. ------------------------------------------------------------------------- SUBROUTINE TIFRTA (N,J,IHE,MY) C C THIS IS THE TIT-FOR-TAT RULE. IN ROUND 1 PLAY 1. IN ROUND N C PLAY 0. OTHERWISE, PLAY WHAT THE OPPONENT PLAYED IN THE PRECEDING C ROUND. C C N = TOTAL NUMBER OF ROUNDS C J = CURRENT ROUND C IHE = THE CHOICE OF THE OPPONENT IN THE PRECEDING ROUND (1 OR 2) C MY = MY CHOICE FOR THE CURRENT ROUND (1 OR 2) C IF (J-1) 20,20,30 20 MY = 1 IF(N.EQ.1) MY=2 RETURN 30 IF (J-N) 50,60,60 50 MY = IHE RETURN 60 MY = 2 RETURN END C------------------------------------------------------------------------ SUBROUTINE GRIM (N,J,IHE,MY) C C THIS IS THE GRIM STRATEGY: START WITH C AND SWITCH TO D C AS SOON AS THE OPPONENT DOES C IF (J-1) 10,10,20 10 IX = 1 20 IF (IHE.EQ.2) IX = 2 IF (J.EQ.N) IX = 2 MY = IX RETURN END ------------------------------------------------------------------------- From csvpi@vtcs1 Mon Feb 10 04:04:34 1986 Date: Mon, 10 Feb 86 04:04:30 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a004827; 10 Feb 86 2:22 EST Date: Sun 9 Feb 1986 22:57-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #22 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Mon, 10 Feb 86 03:52 EST AIList Digest Monday, 10 Feb 1986 Volume 4 : Issue 22 Today's Topics: Queries - AI Society Information & J. of AI, Cognitive Science and Applied Epistemology & Natural Language Interfacing & 3D-package for Xerox 1108 & Psychological Knowledge Structures & ICAI for Physically/Mentally Impaired, Symbolic Math - PDP-11 Equation Solvers, Logic Programming - Bibliography Correction & Quick Summary of NAIL, AI Tools - MIRANDA Functional Programming System ---------------------------------------------------------------------- Date: Sun, 9 Feb 86 23:25:15 est From: walker@mouton.ARPA (Don Walker at mouton.ARPA) Subject: NEED INFORMATION ON AI SOCIETIES; PLEASE HELP I am preparing a short article on associations, societies, and related organizations in artificial intelligence. For each, I would appreciate receiving the following kind of information: name; purpose; date of establishment; principal people involved in getting it started; important events in its history; publications, conferences, and other activities; current membership (if relevant); and any other items of special interest. I would like to put the set of organizations in some historical perspective, if possible. Pointers to other places where something like this has already been done would be particularly helpful, and copies of same would be even more so. Needless to say, net transmission is most efficient, as the deadline is uncomfortably close. And I would particularly value finding someone who would be interested in helping put all this information together! I would expect to include SIGART, ACL, ICCL, AISB, IJCAII, AAAI, CSS, CSCSI, ECCAI, and as many other national and regional groups as possible. Please help if you can; share with me what you have available, even if you think you may not be the most appropriate person to do so; and help get this message out to the people who should know. Net messages to walker@mouton.arpa, walker%mouton@csnet-relay, or ucbvax(or ihnp4, etc.)!bellcore!walker; mail to Don Walker (EAI) Bell Communications Research 445 South Street, MRE 2A379 Morristown, NJ 07960, USA I am sending this notice to publications as well as bboards, digests, and people, but note that the time is too short to justify actually printing it in most of them. Instead, the editors should respond themselves or route it to those most likely to have the information. ------------------------------ Date: Thu 6 Feb 86 13:23:43-PST From: C.S./Math Library Subject: J. of AI, Cognitive Science and Applied Epistemology [Forwarded from the Stanford bboard by Laws@SRI-AI.] I received a message a while ago about the introduction of a new journal titled Journal for the Integrated Study of Artificial Intelligence, Cognitive Science and Applied Epistemology from Ghent, Belgium. However I have not been able to verify that such a journal has been published or is being planned. Does anyone have more information about it? Harry Llull ------------------------------ Date: Wed, 5 Feb 86 13:25 ??? From: Sonny Crockett Subject: Natural Language Interfacing A couple of my students are interested in doing some work on a natural language interface to an Operating System. I'm not really well versed in this particular field. Can someone point me towards a few good papers on this topic? They don't necessarily have to be specifically on Natural language interfaces to OS, generic ones will do. Thanks, Christopher A. Welty RPI/CIE Systems Manager ------------------------------ Date: 4 Feb 86 15:09:52 GMT From: ucdavis!lll-crg!seismo!mcvax!diku!daimi!fleckner@ucbvax.berkeley .edu (Kurt Fleckner) Subject: 3D-package for Xerox 1108 I'm working on a Xerox 1108, and would like to get information about a 3D-package for it. I am designing an expert system to draw the 3D structure of a RNA-molecule. If anyone has any knowledge of such a system, I would be glad if you could mail it to me. If you know about an expert system in that area, I'm interested too. Thanks, Kurt Fleckner Dept. of Comp. Science University of Aarhus Denmark {seismo!mcvax!diku!daimi!fleckner} [Check the last issue (or two) of IEEE Computer Graphics and Applications for some beautiful graphics of DNA molecules in various conformations and at several scales. I was enlightened by the sequence showing DNA twisting to form a chromosome. Ken Knowleton and several others have also developed molecular display software. (I've seen examples in the SIGGRAPH proceedings.) It would be a pity if all this had to be reinvented. -- KIL] ------------------------------ Date: Mon, 3 Feb 86 11:16 EST From: THOMPSON%umass-cs.csnet@CSNET-RELAY.ARPA Subject: Cognitive Psychology - Knowledge Structures I am looking for information about the knowledge structure differences of people who have different levels of expertise in a subject. For example, what is the difference in the knowledge structure of an "apprentice", a "journeyman",or a "master". I will be happy to collect these references and repost them. Please send them directly to me (via csnet). Roger Thompson Thompson@UMASS ------------------------------ Date: Sat, 8 Feb 86 21:49:29 est From: Walter Maner Subject: ICAI for Physically/mentally Impaired Could anyone point me to recent research in the development of intelligent tutoring/training systems for the physically/mentally impaired? My interest is on the software engineering side, not the hardware side. What kinds of unsolved problems exist which might be addressable by ICAI software methods? My impression is that, while there is much activity on the hardware frontier for impaired learners, there has been little innovative work on the software side. So much for my impressions :-). Please reply by mail directly to me. If there are enough responses, I will post a response summary back to mod.ai. Thank you. Walter Maner, Computer Science Department BEST CSNet maner@bgsu : ARPANet maner%bgsu@csnet-relay : UUCP ...cbosgd!osu-eddie!bgsuvax!maner : Mail BGSU, Bowling Green, OH 43403 : CompuServe 73157,247 WORST Phone (419) 372-8719 or -2337 ------------------------------ Date: Mon, 3 Feb 1986 23:10 EST From: Jonathan Cohn Subject: PDP-11 Equation Solvers I believe that such work was being done at Stevens Institute of Tech. in Hoboken NJ in 1982-3 on a Pro-350 (PC version of PDP-11) at the math department you might want to try in get in touch with Larry Levine there, he is in the math department, and I think lead that project. He has a computer address on bitnet of LLEVINE@SITVXB. Jonathan Cohn JC595C@GWUVM.BITNET COHN@NSFVAX.BITNET COHN@NSFVAX.CSNET ------------------------------ Date: Fri, 7 Feb 86 16:28:28 PST From: newton@vlsi.caltech.edu (Mike Newton) Subject: small correction A small correction to last digest's bibliography: %A H. Derby %T Using Logic Programming for Compiling APL ... %C Los Angeles, California to: %A H. Derby %T Using Logic Programming for Compiling APL %R Technical Report 84-5134 %I Department of Computer Science %I California Institute of Technology %C Pasadena, California 91125 %D 1984 - mike ------------------------------ Date: Wed, 29 Jan 86 10:20:22 pst From: Allen VanGelder Subject: Quick summary of NAIL [Forwarded from the Stanford bboard by Laws@SRI-AI.] NAIL is a research project one of whose goals is to determine what degree of expressiveness and efficiency can be obtained by a logic based language without resorting to certain "undesirable" non-logical mechanisms such as cut, assert and retract, rule order, and subgoal order. Jeff Ullman, the PI, likes to draw the analogy: "NAIL is to Prolog as Relational DBMS is to CODASYL." NAIL is in a preliminary stage of development at Stanford CSD. An overview, "Design overview of the Nail! System" is available from Professor Ullman. NAIL! is an acronym for "Not Another Implementation of Logic!" ------------------------------ Date: Fri, 31 Jan 86 23:39:35 GMT From: dat%ukc.ac.uk@cs.ucl.ac.uk Subject: MIRANDA Functional Programming System MIRANDA This is to inform anyone who may be interested that a UNIX implementation of the Miranda functional programming system is now available for the following machines: VAX (under 4.2 BSD), ORION, and SUN workstations. It will be ported to a number of other UNIX machines in the near future. The rest of this message contains a brief description of the Miranda system, followed by information about how to obtain it. What is Miranda? Miranda is an advanced functional programming language designed by David Turner of the University of Kent. It is based on the earlier languages SASL, KRC and ML. A program in Miranda is a set of equations describing the functions and data structures which the user wishes to compute. Programs written in Miranda are typically ten to twenty times shorter than the equivalent programs in a conventional high level language such as PASCAL. The main features of Miranda are: 1) Purely functional - no side effects 2) Higher order - functions can be treated as values 3) Infinite data structures can be described and used 4) Concise notation for sets and sequences ("zf expressions") 5) Polymorphic strong typing The basic types of the language are numbers (integer and double precision floating point), characters, booleans, lists, tuples, and functions. In addition a rich variety of user-defined types may be introduced by writing appropriate equations. A more detailed discussion of the language may be found in "Miranda: a non-strict functional language with polymorphic types", in Springer Lecture Notes in Computer Science, vol 201. The Miranda system is a self contained sub-system, running under UNIX. The Miranda compiler works in conjunction with a screen editor (normally this is `vi', but it is easy to arrange for this to be another editor if preferred). Programs are automatically recompiled in response to source edits and any syntax or type errors signalled immediately. The type system enables a high proportion of semantic errors to be detected at compile time. There is an online reference manual, which documents the system at a level appropriate for someone already familiar with the main ideas of functional programming (more tutorial material is in preparation). Execution is by a fast interpreter, using an intermediate code based on combinatory logic. The Miranda system is a powerful tool, enabling complex applications to be developed in a fraction of the time required in a conventional programming system. Applications which have been developed in Miranda include - compilers, theorem provers, and digital circuit simulation. It is envisaged that the main uses of Miranda will be: 1) Teaching the concepts of functional programming 2) Rapid prototyping 3) As a specification language 4) For further research into functional programming 5) As a general purpose programming language Release Information The Miranda system has been developed by Research Software Ltd. It is distributed in object code form and is currently available for the following machines - VAX (under 4.2BSD), ORION, SUN 2, SUN 3. The license fee, per cpu, is 300 pounds for an educational license and 975 pounds for a commercial license (US prices: $450, $1450, respectively). If you think you may be interested in obtaining a copy of the Miranda system please send your name and (postal) address to the following electronic mail address, and you will be sent further information and a copy of the license form etc: USENET: ...!mcvax!ukc!mira-request JANET: mira-request@ukc.ac.uk ARPANET: mira-request%ukc@ucl-cs Or telephone Research Software on: 0227 471844 (omit the initial `0' if calling from outside England) If you are interested in obtaining Miranda on a different machine, or a different version of Unix, from those listed above, it is also worth mailing details of your situation, since future porting policy will be largely determined by perceived demand. ((NB - UNIX systems only, please.)) David Turner ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Wed Feb 12 23:29:59 1986 Date: Wed, 12 Feb 86 23:29:55 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a002143; 12 Feb 86 13:29 EST Date: Wed 12 Feb 1986 09:31-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #23 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Wed, 12 Feb 86 23:21 EST AIList Digest Wednesday, 12 Feb 1986 Volume 4 : Issue 23 Today's Topics: Seminars - Systems of Actors (USC) & Artificial Concept Formation (Edinburgh) & Parallelism in Production Systems (SU) & A Storage Manager for Prolog (SU) & Statistical Theory of Evidence (SRI), Conference - Compcon Spring 86 ---------------------------------------------------------------------- Date: 7 Feb 1986 08:21-EST From: gasser@usc-cse.usc.edu Subject: Seminar - Systems of Actors (USC) USC DISTRIBUTED PROBLEM SOLVING GROUP MEETING "Formalizing the Development of Systems of Actors" Ed Ipser Ph.D Student, USC A formalization of the process of specifying and developing distributed systems is presented, with the emphasis on the description of multiple robot environments. The general scheme is a recursive reduction of behaviors with constraints to actors with pre-determined behaviors by showing that the behaviors of the actors satisfy the behavior and constraint requirements of the system. Possible applications of this scheme are presented, including automatic programming, planning, theorem proving, and the description of non-computable functions. This work is based on the work of Goldman and Wile on GIST, and Georgeff's work on the theory of processes. Time: 3:00 PM Wednesday, Feb 12, 1986 Place: Seaver Science Bldg., Room 319, USC Questions: Dr. Les Gasser, CS Dept., USC (213) 743-7794 ------------------------------ Date: Mon, 10 Feb 86 14:56:48 GMT From: Gideon Sahar Subject: Seminar - Artificial Concept Formation (Edinburgh) EDINBURGH AI SEMINARS Date: Wednesday, 12th February l986 Time: 2.00 p.m. Place: Department of Artificial Intelligence Seminar Room - F10 80 South Bridge EDINBURGH. Professor Donald Michie, The Turing Institute, Glasgow will give a seminar entitled - "Artificial Concept Formation". The approach develops from a position taken in the 1950's by H.A. Simon. He proposed, in essence, a new criterion for the adequacy of a theory (he considered economic theory), namely that in explaining the flux of transactions a theory must take full account of the resource-limited nature of the calculations performed by the participating agents. Is economic man rational in the sense of making fully rational choices whatever the computational cost (as in the von Neumann and Morgenstern theory of economic behaviour), or does he exhibit at most the level of rationality which human brains can feasibly compute in the time available for each choice? By implication Simon also requires that such a theory should be feasibly interpretable by its human user: runnability on the machine is not enough. This leads to the idea that what is run on the machine should be human-oriented in a very strong sense, unprecedented in conventional software technology even as an aspiration: if a program is to be not just an operationally effective description or prescription, but a machine representation of a concept and hence an eligible component of a Simon-type theory, it must be not only human-intelligible but also human-interpretable. This entails that the human expert skilled in the given area must be able mentally to check it against trial data in his head, just as he can in the case of his own professionally acquired concepts. ------------------------------ Date: Mon 10 Feb 86 09:28:13-PST From: Sharon Gerlach Subject: Seminar - Parallelism in Production Systems (SU) On Friday, Feb 21, Anoop Gupta, a CSL faculty candidate from CMU, will be speaking on "Parallelism in Production Systems" in MJH 352 at 3:15. Parallelism in Production Systems Anoop Gupta Department of Computer Science Carnegie-Mellon University Pittsburgh, PA 15213 Production systems (or rule-based systems) are widely used in Artificial Intelligence for modeling intelligent behavior and building expert systems. Most production system programs, however, are extremely computation intensive and run quite slowly. The slow speed of execution has prohibited the use of production systems in domains requiring high performance and real-time response. The talk will elaborate on the role of parallelism in the high-speed execution of production systems. On the surface, production system programs appear to be capable of using large amounts of parallelism -- it is possible to perform match for each production in a program in parallel. Our research shows that in practice, however, the speed-up obtainable from parallelism is quite limited, around 10-fold as compared to initial expectations of 100-fold to 1000-fold. The main reasons for the limited speed-up are: (1) there are only a small number of productions that are affected (require significant processing) as a result of a change to working memory and (2) there is a large variation in the processing requirement of these productions. Since the number of affected productions is not controlled by the implementor of the production system interpreter (it is governed mainly by the author of the program and the nature of the problem), the solution to the problem of limited speed-up is to somehow decrease the variation in the processing cost of affected productions. We propose a parallel version of the Rete algorithm which exploits parallelism at a very fine grain to reduce this variation. We further suggest that to exploit the fine-grained parallelism, a shared-memory multiprocessor with 32-64 high performance processors should be used. For scheduling the fine-grained tasks consisting of about 50-100 instructions, a hardware task scheduler is proposed. The results presented in the talk are based on simulations done for a large set of production systems exploiting different sources of parallelism. The simulation results show that using the suggested multiprocessor architecture (with individual processors performing at 2 MIPS), it is possible to obtain execution speeds of 5000-27000 working memory element changes per second. This corresponds to a speed-up of 5-fold to 27-fold over the best known sequential implementation using a 2 MIPS processor. This performance is also higher than that obtained by other proposed parallel implementations of production systems. ------------------------------ Date: Tue 11 Feb 86 16:30:05-PST From: Karin Scholz Subject: Seminar - A Storage Manager for Prolog (SU) this is a correction to the colloquium notice for this week: Database Seminar CS 545, Friday Feb 14, 3:15pm, mjh352 Persistent Prolog: A Secondary Storage Manager for Prolog Peter M D Gray University of Aberdeen, Scotland ABSTRACT OF TALK The talk will describe a general purpose "tight coupling" system based on a C-Prolog interpreter interfaced to a "Persistent Heap" database, which can store a wide variety of data types and objects. We are currently extending Prolog to allow definitions of modules and Abstract Data Types. This provides a disciplined way of accessing frame structures, bit maps, attached procedures and other non-Prolog objects. With this system we are able to use Prolog to maintain an evolving knowledge base on disc. Prolog clauses and data structures are manipulated in memory in the usual way, but migrate to disc on a "commit" step. This work is part of the U.K. "Alvey" program in IKBS ------------------------------ Date: Wed 12 Feb 86 08:55:19-PST From: FIRSCHEIN@SRI-AI.ARPA Subject: Seminar - Statistical Theory of Evidence (SRI) Bob Hummel will be giving a talk on Tuesday, Feb. 18 at 10:30, Conf room EK242 (the "old" conf room). An abstract of his talk follows: A Statistical Viewpoint on the Theory of Evidence Robert Hummel Courant Institute, New York University Abstract The Dempster/Shafer "Theory of Evidence" can be regarded as an alge- braic space with a combination formula that combines the opinions of "experts". This viewpoint, which is really the origin of the theory, will be explained by introducing spaces with simple binary operations, giving these spaces intuitive interpretations, relating them to Bayesian updating, and showing that the spaces are (in a homomorphic sense) equivalent to the Dempster/Shafer theory of evidence space. The viewpoint allows us to remark on limitations of the theory. By making compromises in a different manner, an alternative combination method can be introduced. This representation of states of belief by "Parameter- ized Statistics of Experts" will be described. ------------------------------ Date: Fri, 31 Jan 86 07:40:54 pst From: Doug Coffland Subject: Conference - Compcon Spring 86 Register for Compcon Spring 86 now and attend the year's only broad based computing conference sponsored by the IEEE Computer Society. Compcon will be held in San Fran- cisco, March 3-6, 1986. Key topics include: supercomputers, SDI software reliability, AI applications, Japanese software practices, RISC vs. CISC, and more. Four full day tutorials will be given on Monday, March 3. Topics include silicon compilation, issues in expert systems, complex computer graphics, and high performance com- puting. The advanced registration deadline is February 14. For further information, contact Robert M. Long, Lawrence Livermore National Labratory, P. O. Box 808, MS L130, Livermore, Ca. 94550. The telephone number is 415-422-8934. Telephone registrations will be accepted with Visa, MasterCard, or American Express. ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Feb 13 03:53:16 1986 Date: Thu, 13 Feb 86 03:53:12 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a007348; 13 Feb 86 0:32 EST Date: Wed 12 Feb 1986 09:47-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #24 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 13 Feb 86 03:46 EST AIList Digest Wednesday, 12 Feb 1986 Volume 4 : Issue 24 Today's Topics: Queries - Literature Search & Distributed Databases, AI Tools - LISP Source Code, Applications - ISIS, Journals - Belgian AI/CogSci/Epistemology Journal, Re: Cognitive Psychology - Knowledge Structures Humor - Animated Computer Personalities & Paranoid Computers & Koans ---------------------------------------------------------------------- Date: Mon, 10 Feb 86 14:52:34 CST From: veach%ukans.csnet@CSNET-RELAY.ARPA Subject: Literature search. I am beginning a research project on the control of multiple expert systems in a single package/environment. If anyone has any bibliographies and/or references to literature on the control/scheduling/implementation of multiple expert systems and would kindly share it with me I would appreciate it. Thanks Glenn O. Veach Artificial Intelligence Laboratory Department of Computer Science University of Kansas Lawrence, KS 66044 (913) 864-4482 ------------------------------ Date: 6 Feb 86 15:13:34 GMT From: ulysses!mhuxr!mhuxt!houxm!mtuxo!drutx!druky!krahl@ucbvax.berkeley.edu (R.H. Krahl) Subject: Distributed Databases Anyone having any articles or information in regards to distributed databases with expert systems would be very much appreciated. Thanks-in-advance. Rich Krahl @ AT&T-ISL, Denver EMAIL: {allegra, cbosgd, ihnp4}!druky!krahl 11900 N. Pecos Denver, CO. 80234. ------------------------------ Date: Sat, 8 Feb 86 14:07:34 pst From: sdcsvax!sdcrdcf!hplabs!oblio!paf@ucbvax.berkeley.edu (Paul Fronberg) Subject: Re: request for LISP source code Try Scheme from the GNU emacs distribution. This is the version of LISP utilized in "Structure and Interpretation of Computer Programs". The source is ~ $150 and includes GNU emacs + Scheme + Bison (as of 7/85). There was no problem in getting Scheme to build on either BSD 4.2 or USG V.2 (slight modification of build files necessary in the last case). ------------------------------ Date: Mon 10 Feb 86 15:52:39-CST From: CMP.BARC@R20.UTEXAS.EDU Subject: Re: ISIS ISIS is a factory scheduling KBS developed by Mark Fox and Stephen Smith at the Intelligent Systems Laboratory of the Robotics Institute at CMU, in conjunction with Westinghouse. It constructs job-shop schedules, monitors performance and avoids production bottlenecks, by evaluating and resolving conflicting factors such as productivity goals, resource requirements and machine preferences. References: Fox and Smith, "ISIS -- a KBS for factory scheduling", Expert Systems, v. 1, n. 1, July 1984, pp. 25-49. Fox, Smith, et al, "ISIS: A Constraint-Directed Reasoning Approach to Job Shop Scheduling", Proc IEEE Conf. on Trends and Applications 83, Gaithers- berg, MD, May 1983. Dallas Webster Burroughs Austin Research Center CMP.BARC@R20.UTexas.Edu {ihnp4, seismo, ctvax}!ut-sally!batman!dallas ------------------------------ Date: 10 FEB 86 17:04-N From: KEMPEN%HNYKUN52.BITNET@WISCVM.WISC.EDU Subject: Info on Belgian AI journal (AIList) The journal is called: Title: CC-AI Subtitle: The journal for the integrated study of Artificial Intelligence, Cognitive Science and Applied Epistemology. Editorial Address: CC-AI Blandijnberg 2 B-9000 Ghent, Belgium tel. +32 (91) 257571, ext. 4522 TELEX RUGENT 12.754 Publisher: Communication & Cognition (Same address) Gerard Kempen ------------------------------ Date: Mon, 10 Feb 86 18:23:59 pst From: sdcsvax!sdcrdcf!ucla-cs!koen@ucbvax.berkeley.edu (Koenraad Lecot) Subject: Re: J. of AI, Cognitive Science and Applied Epistemology The journal had a couple of issues last year. Papers cover a wide variety of topics within AI. Not too technical stuff. Have not received any issues this year yet. -- Koenraad Lecot ------------------------------ Date: Tue, 11 Feb 86 09:36:10 cst From: bulko@SALLY.UTEXAS.EDU (Bill Bulko) Reply-to: bulko@sally.UUCP (Bill Bulko) Subject: Re: Cognitive Psychology - Knowledge Structures My attempted mail reply to thompson@umass-cs.csnet failed, so I'm posting this instead. The request was for pointers to articles dealing with how varying levels of expertise could be represented. My research is related to problem solving in physics, and so I have read several papers dealing with the way people learn how to solve problems in technical fields. Below is an excerpt from my proposal containing the related (annotated) references; I hope that they prove helpful. Bhaskar, R., and H. A. Simon, "Problem Solving in Semantically Rich Domains: An Example from Engineering Thermodynamics." Cognitive Science, Vol. 1, No. 2, April 1977. This is a study of the processes used by people to solve problems in semantically rich domains, and how these processes compare with those in general problem-solving domains. The authors choose the field of thermodynamics, and use a protocol-encoding program called SAPA, which they theorize corresponds to their subject's problem-solving behavior. Chi, M. T. H., P. Feltovich, and R. Glaser, "Categorization and Representation of Physics Problems by Experts and Novices." Cognitive Science, Vol. 5, No. 2, April-June 1981. The authors compare the ways experts and novices categorize physics problems and form physical models of the problems based on the categories created. Studies are presented which investigate the implications of the differences found for problem solving in general. Larkin, J., J. McDermott, D. Simon, and H. A. Simon, "Models of Competence in Solving Physics Problems." Cognitive Science, Vol. 4, No. 4, October- December 1980. This article discusses how a person's experience and expertise in solving physics problems determine the process by which he solves them. The authors describe a set of two computer programs which they claim are accurate models of "expert" and "novice" problem-solving protocols. Larkin, J., and H. A. Simon, "Learning Through Growth of Skill in Mental Modeling." Proceedings of the Third Annual Conference of the Cognitive Science Society, p. 106. The authors study how people develop the ability to take physical situations and re-represent them in terms of scientific entities. They present a program called ABLE, which models the performance of human experts and novices as they solve physics problems, from this learning point of view. Luger, G., "Mathematical Model Building in the Solution of Mechanics Problems: Human Protocols and the MECHO Trace." Cognitive Science, Vol. 5, No. 1, January-March 1981. Luger describes an automatic problem solver, MECHO, and describes how it can be used for model building and manipulation in solving problems in physics. He compares traces of MECHO with the problem-solving protocols of several human subjects, and hypothesizes that these traces are similar to the model-building techniques that people in general use. Hope these help, Bill "In the knowledge lies the power." -- Edward A. Feigenbaum "Knowledge is good." -- Emil Faber Bill Bulko Department of Computer Sciences The University of Texas {ihnp4,harvard,gatech,ctvax,seismo}!sally!bulko ------------------------------ Date: 27 Jan 86 16:33:00 GMT From: pur-ee!uiucdcs!uicsl!pollack@ucbvax.berkeley.edu Subject: Re: Two AI software packages RE: Mom There was an article by Thomas Friedman in the NYT a couple of months ago on two programs for the Atari ST written by "the Israeli equivalent of Garry Trudeau": "MOM" and "MURRAY" are animated computer personalities, They sit in comfortable chairs on the screen and talk to you. Murray is a raconteur, with supposedly an ever-expanding database of humor, and a memory for the jokes he already told you, and MOM is a typical mother figure, who can make you feel guilty for anything, even spending the $49 to buy her. Their dialog appears in white bubbles above their heads, and the user gets to type in their name and answer yes/no questions. ------------------------------ Date: Fri, 31 Jan 86 14:16:32 GMT From: gcj%qmc-ori.uucp@cs.ucl.ac.uk Subject: noaK Re: 2010 and H-Mobius Loops and HAL's paranoia (Vol 4 # 17). Why not give HAL (an intelligent system) the Rorschach inkblot test, "to show intelligence, personality and mental state"? Another psychological test, the IQ test, was proposed by in volume 3, number 164. Gordon Joly aka The Joka ARPA: gcj%qmc-ori@ucl-cs.arpa UUCP: {...!seismo!mcvax}!ukc!kcl-cs!qmc-ori!gcj ------------------------------ Date: Mon, 10 Feb 86 14:08:31 EST From: decwrl!decvax!sunybcs!colonel@ucbvax.berkeley.edu (Col. G. L. Sicherman) Subject: Re: ai koans A P.I. who was trying to meet a deadline said to his assistant: "Excuse me, I couldn't help noticing that you're not working!" "The computer isn't working," the assistant replied. PASK, overhearing them, commented: "Not the assistant, not the computer. The man-machine interface isn't working." ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Wed Feb 12 23:40:49 1986 Date: Wed, 12 Feb 86 23:40:45 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a005723; 12 Feb 86 20:35 EST Date: Wed 12 Feb 1986 10:45-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #25 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Wed, 12 Feb 86 23:29 EST AIList Digest Wednesday, 12 Feb 1986 Volume 4 : Issue 25 Today's Topics: Journals - New Journal on Applied AI & CACM Invitation to Authors, Conference - NCAI Exhibit Program, Theory - Technology Review Article & Taxonomizing in AI ---------------------------------------------------------------------- Date: Sat, 25 Jan 86 14:21:41 est From: FOXEA%VTVAX3.BITNET@WISCVM.WISC.EDU Subject: New Journal on Applied AI [Forwarded from the IRList Digest.] New Journal: Applied Artificial Intelligence, An International Journal Publication Information: published quarterly starting March 86 Rates: $55/volume indiv ($88 institutional) plus $24 air mail postage Contacts: order with check or money order to - Hemisphere Publishing Corporation, Journals Dept., 79 Madison Ave. New York, New York 10016 Information: Elizabeth D'Costa, Circulation Mgr. (212) 725-1999 Aims and Scope: Applied Artificial Intelligence is intended to help exchange information about advances and experiences in this field among AI researchers. Furthermore, it will aid decision makers in industry and management to understand the accomplishments and limitations of the state-of-the-art of artificial intelligence. Research to be presented will focus on methodology, time-schedules, problems, work force strength, new tools, transfer of theoretical accomplishements to application problems, information exchange among concerned AI researchers and decision makers about the potential impact of their work on their decisions. ------------------------------ Date: Mon 10 Feb 86 22:49:05-PST From: Peter Friedland Subject: Invitation to Authors I have recently been named to the Editorial Panel of Communications of the ACM (CACM) with responsibility for artificial intelligence. CACM is by far the widest-read computing publication with a current circulation of over 75,000. I would like to encourage submissions to CACM in one of several forms: articles of general interest (surveys, tutorials, reviews), research contributions (original, previously-unpublished reports on significant research), and reports on conferences or committee meetings. In particular, manuscripts which act to bridge the gap between artificial intelligence research and traditional computing methodologies are welcome. All contributions will be fully reviewed with authors normally notified of acceptance or rejection within 3 months of receipt. In addition, CACM intends to devote substantial amounts of space to special collections of related, high-quality, "Scientific American-like" articles. For examples, see the September 1985 issue on "Architectures for Knowledge-Based Systems" or the November 1985 issue on "Frontiers of Computing in Science and Engineering." These special sections are usually composed of invited papers selected by a guest editor from the community. Professional editors at ACM headquarters devote on the order of man-weeks per article in developing graphics and helping to make the articles readable by a wide cross-section of the computing community. I welcome suggestions (and volunteers) from anybody in the AI community for such special sections. Articles and research contributions should be submitted directly to: Janet Benton Executive Editor, CACM 11 West 42nd St. New York, NY 10036 Ideas for articles or special sections, and volunteers for helping in the review process to insure the highest quality of AI publication in CACM should be sent to me as FRIEDLAND@SUMEX (or call 415-497-3728). Peter Friedland ------------------------------ Date: Mon 10 Feb 86 11:39:47-PST From: AAAI Subject: Special Invitation The AAAI would like to extend a special invitation to academic institutions and non-profit research laboratories to participate in this year's Exhibit Program at the National Conference on Artificial Intelligence, August 11-15, 1986 in the Philadelphia Civic Center. It's important to communicate what universities and labortories are doing in AI by demonstrating their different research projects to our conference attendees. The AAAI will provide one 10' x 10' booth free of charge, describe your demonstration in the Exhibit Guide, and assist you with your logistical arrangements. Although we can not provide support equipment (e.g. phone, lines or computers), we can direct you to different vendors who may be able to assist you with your equipment needs. If you and your department are interesting in participating, please call Ms. Lorraine Cooper at the AAAI (415) 328-3123. ------------------------------ Date: 3 Feb 86 19:46:53 GMT From: ulysses!burl!clyde!watmath!utzoo!utcsri!utai!lamy@ucbvax.berkeley.edu (Jean-Francois Lamy) Subject: Re: Technology Review article In article <7500002@ada-uts.UUCP> richw@ada-uts.UUCP writes: >like: "In 25 years, AI has still not lived up to its promises and >there's no reason to think it ever will" Still thinking that fundamental breakthroughs in AI are achievable in such an infinitesimal amount of time as 25 years is naive. I probably was not even born when such claims could have been justified by sheer enthousiasm... Not that we cannot get interesting and perhaps even useful developments in the next 25 years. >P.S. You might notice that about 10 pages into the issue, there's > an ad for some AI system. I bet the advertisers were real > pleased about the issue's contents... Nowadays you don't ask for a grant or try to sell a product if the words "AI, expert systems, knowledge engineering techniques, fifth generation and natural language processing" are not included. Advertisement is about creating hype, and it really works -- for a while, until the next "in" thing comes around. Jean-Francois Lamy Department of Computer Science, University of Toronto, Departement d'informatique et de recherche operationnelle, U. de Montreal. CSNet: lamy@toronto.csnet UUCP: {utzoo,ihnp4,decwrl,uw-beaver}!utcsri!utai!lamy CDN: lamy@iro.udem.cdn (lamy%iro.udem.cdn@ubc.csnet) ------------------------------ Date: Fri, 7 Feb 86 20:51:58 PST From: larry@Jpl-VLSI.ARPA Subject: Sparklers from the Tech Review I haven't read the Tech Review article; perhaps I shall just to see how different will be my interpretation of it from the opinions heard here. The discussion has made me want to offer some ideas of my own. What we lump under AI is several different fields of research with often very different if not contradictory approaches. As a dilletante in the AI field I perceive the following: COGNITIVE PSYCHOLOGY (a more restricted area than Cognitive Science) attempts to understand biologically based thinking using behavioral and psychiatric concepts and methods. This includes the effects emotional and social forces exert on cognition. This group is increasingly borrowing from the following groups. COGNITIVE SCIENCE attempts to broaden the study to include machine-based cognition. CS introduces heavy doses of metaphysics, logic, linguistics, and information theory. My impression is that this area is too heavily invested in symbol-processing research and could profitably spend more time on analog computation and associative memories. These may better model humans' near- instantaneous decision-making, which is more like doing a vector-sum than doing massively parallel logical inferences. PATTERN RECOGNITION, ROBOTICS, ETC. attempts to engineer cognition into machines. Many workers in this field have a strong "hard-science" background and a pragmatic approach; they often don't care whether they reproduce or whether they mimic biological cognition. EXPERT SYSTEMS, KNOWLEDGE ENGINEERING is more software engineering than hardware engineering. Logic, computer science, and database theory are strong here. Some of the simpler expert systems are imminently practical and have been around for decades--though "programmed" into trouble-shooting books and the like rather than a computer. (And while we're on this, most of what now passes for rule-based programming could be done in BASIC or assembly language, including self-modifying code, using fairly simple table-driven techniques.) And perhaps several more groups could be distinguished. Of course, there are plenty of exceptions to these categories, but humans do self-select into groups and distill ideas and techniques into a rudimentary group persona. If I were to characterize myself, I'd probably say that I'm less interested in AI than IA--Intelligence Amplification. I'm interested by attempts to create machine versions of human intelligence and I have little doubt that all the vaunted "mystical" abilities of humans will eventually be reproduced, including self-awareness. Some of these abilities may be much easier to reproduce than we suppose: intuition, for instance. I'm an artist in several media and use intuition routinely. I've spent a lot of time introspecting about what happens when I "solve" artistic problems, and I've learned how to "program" my undermind so that I can promise solutions with considerable reliability. I believe I could build an intuitive computer. But what fascinates me is the idea of building systems which combine the best capabilities of human and machine to overcome the limits of both. I think it's much more economical, practical, and probably even humane to, say, make a language-translation system that uses computers to do rapid, rough transla- tions of 99% of a text and uses human sensitivities and skills to polish and validate the translations. (Stated like that it sounds like two batch jobs with a pipe between them. My concept is an interactive system with both human and computer collaborating on the job, with the human doing continuous shaping and scheduling of the entire process.) Now I'll go back to being an interested by-stander for another six months! Larry @ JPL-VLSI.arpa ------------------------------ Date: 3 Feb 86 18:04:58 GMT From: amdcad!lll-crg!seismo!rochester!lab@ucbvax.berkeley.edu (Lab Manager) Subject: Re: Technology Review article In article <7500002@ada-uts.UUCP> richw@ada-uts.UUCP writes: > >Has anyone read the article about AI in the February issue of >"Technology Review"? You can't miss it -- the cover says something >like: "In 25 years, AI has still not lived up to its promises and >there's no reason to think it ever will" (not a direct quote; I don't >have the copy with me). General comments? They basically say that things like blocks world doesn't scale up, and AI can't model intuition because 'real people' aren't thinking machines. An appropriate rebuttal to these two self-styled philosophers: "In 3000 years, Philosophy has still not lived up to its promises and there's no reason to think it ever will." Brad Miller Arpa: lab@rochester.arpa UUCP: rochester!lab (also miller@rochester for non-lab stuff) Title: CS Lab Manager Snail: University of Rochester Computer Science Dept. 617 Hylan Building Rochester NY 14627 ------------------------------ Date: Sun, 9 Feb 86 16:38:38 est From: "Marek W. Lugowski" Subject: Taxonomizing in AI: neither useful or harmless > [Stan Shebs:] In article <3600036@iuvax.UUCP> marek@iuvax.UUCP writes: > > Date: 4 Feb 86 19:55:00 GMT > From: ihnp4!inuxc!iubugs!iuvax!marek@ucbvax.berkeley.edu > > ha ha ha! "taxonomy of the field" -- the latest gospel of AI? Let me be > impudent enough to claim one of the most misguided AI efforts to date is > taxonomizing a la Michalski et al: setting up categories along arbitrary > lines dictated by somebody or other's intuition. If AI does not have > the mechanism-cum-explanation to describe a phenomenon, what right does it > have to a) taxonomize it and b) demand that its taxonomizing be recognized > as an achievement? > -- Marek Lugowski > > I assume you have something wonderful that we haven't heard about? I assume that you are intentionally jesting, equating that which I criticize with all that AI has to offer. Taxonomizing is a debatable art of empirical science, usually justified when a scientist finds itself overwhelmed with gobs and gobs of identifiable specimens, e.g. entymology. But AI is not overwhelmed by gobs and gobs of tangible singulars; it is a constructive endeavor that puts up putatative mechanisms to be replaced by others. The kinds of learning Michalski so effortlessly plucks out of the thin air are not as incontrovertibly real and graspable as instances of dead bugs. One could argue, I suppose, that taxonomizing in absence of multitudes of real specimens is a harmless way of pursuing tenure, but I argue in Indiana U. Computer Science Technical Report No. 176, "Why Artificial Intelligence is Necessarily Ad Hoc: Your Thinking/Approach/Model/Solution Rides on Your Metaphors", that it causes grave harm to the field. E-mail nlg@iuvax.uucp for a copy, or write to Nancy Garrett at Computer Science Department, Lindley Hall 101, Indiana University, Bloomington, Indiana 47406. > Or do you believe that because there are unsolved problems in physics, > chemists and biologists have no right to study objects whose behavior is > ultimately described in terms of physics? > > stan shebs > (shebs@utah-orion) TR #176 also happens to touch on the issue of how ill-formed Stan Shebs's rhetorical question is and how this sort of analogizing has gotten AI into its current (sad) shape. Please consider whether taxonomizing kinds of learning from the AI perspective in 1981 is at all analogous to chemists' and biologists' "right to study the objects whose behavior is ultimately described in terms of physics." If so, when is the last time you saw a biology/chemistry text titled "Cellular Resonance" in which 3 authors offered an exhaustive table of carcinogenic vibrations, offered as a collection of current papers in oncology?... More constructively, I am in the process of developing an abstract machine. I think that developing abstract machines is more in the line of my work as an AI worker than postulating arbitrary taxonomies where there's neither need for them nor raw material. -- Marek Lugowski an AI graduate student Indiana U. CS ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Fri Feb 14 05:26:51 1986 Date: Fri, 14 Feb 86 05:26:47 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a019609; 14 Feb 86 1:11 EST Date: Thu 13 Feb 1986 21:48-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #26 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Fri, 14 Feb 86 05:12 EST AIList Digest Friday, 14 Feb 1986 Volume 4 : Issue 26 Today's Topics: Queries - Automatic Testing of Parsers & Baseball Expert Systems, Literature - AI in Engineering & Business Week on Expert Systems, AI Tools - LISP Compilers, Education - ICAI for the Physically/Mentally Impaired, Games - Artificial Animals & Software Robots ---------------------------------------------------------------------- Date: 13 Feb 86 10:52:00 EST From: "CUGINI, JOHN" Reply-to: "CUGINI, JOHN" Subject: automatic testing of parsers Do any systems exist which can accept a body of BNF (or some other syntactic production rules), and then generate or enumerate test cases to be run against an alleged parser of that BNF? Thanks in advance for any help... John Cugini National Bureau of Standards ------------------------------ Date: 8 Feb 86 00:31:08 GMT From: sdcsvax!noscvax!priebe@ucbvax.berkeley.edu (Carey E. Priebe) Subject: baseball expert systems **************************************************************** i need pointers to or information about expert systems that have been developed for the baseball domain. i would be interested in research or incomplete programs as well as mature systems. i believe there was some related work ongoing at yale recently, per- haps focusing on natural language, but my information is sketchy. reply directly to me or through the net. thanx in advance. cp ***************************************************************** ------------------------------ Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: AI in Engineering As a news editor for "Artificial Intelligence in Engineering", I request that people send me information on new applications of artificial intelligence to engineering problems, whether they be products, research efforts, industrial applications or related items such as conferences or new bindings. Please send the information to me at: Laurence L. Leff Computer Science and Engineering Southern Methodist University Dallas, Texas 75275 bitnet: E1AR0002 at SMUVM1 Arpanet, CSNET leff%smu@csnet-relay UUCPnet ihnp4!convex!smu!leff ------------------------------ Date: Thu 13 Feb 86 10:58:34-PST From: Ken Laws Subject: Business Week on Expert Systems Check the February 10 issue of Business Week, pp. 94, 98-99, for a discussion of the funding and prospects of Intellicorp, Teknowledge, Inference Corp., and the Carnegie Group. They are described as the Gang of Four in AI. -- Ken Laws ------------------------------ Date: Wed, 12 Feb 86 09:00:25 est From: sdcsvax!dcdwest!ittatc!decvax!linus!raybed2!gxm@ucbvax.berkeley.edu (GERARD MAYER) Subject: Re: LISP Compilers? Get in touch with Franz Inc., 2920 Domingo Ave, Suite 203, Berkeley, CA 94705 (415) 540-1224 for common lisp product running on unix. Gerard Mayer Raytheon Research Division uucp ..linus!raybed2!gxm ------------------------------ Date: Wed, 12 Feb 86 17:59:06 mst From: ulysses!ihnp4!alberta!arms@ucbvax.berkeley.edu (Bill Armstrong) Subject: Re: ICAI for Physically/mentally Impaired There is a softcover book: Microcomputer Resource Book for Special Education by Dolores Hagen published by Reston in 1984. It deals with questions of the learning impaired, deaf, blind, and physically handicapped, but points out that a lot of software is useful to the handicapped even if it isn't so labelled. The ISBN numbers are 0-8359-4345-3 and 0-8359-4344-5 (paperback) Call number LC4019.H33 1984. I don't know whether it satisfies the ICAI criterion or is just CAI. The person to talk to about ICAI would be Greg Kearsley, Courseware, Inc., 10075 Carroll Canyon Road, San Diego, California 92131. I hope this helps you. ------------------------------ Date: Thu 23 Jan 86 10:44:55-PST From: Mark Richer Subject: Artificial Animals [Excerpted from the AI-Ed distribution.] Computer Currents, 22-oct-85 [a computer newspaper] Strehlo: What's the nature of the research? Kay: It's yet another attempt to try and understand the thin edge of the long wedge. At PARC, the children used Smalltalk on the interim Dynabook to build their own application programs, their own editors and animation and stuff like that. In this case, we're sort of upping the ante to try and do a system in which the children can create little mentalities, animal level mentalities that can be put into a simulated environment where they have to survive. If you will, it's like creating a little Disney character that you then put out into a big world." Strehlo: We see this kind of thing on a simple level in adventure games where the player has to give characters the traits needed to achieve some goal. Kay: Right, exactly. Strehlo: And this just goes further? How would it go further? Kay: "It goes a lot further. We're shooting for something that will be dynamically animated and will actually learn things. The idea is to get kids to be more thoughtful about thinking by getting them to try to think about how animals think, and by taking the results of these comtemplations and actually building animal-like creatures that work. It's exciting. There's very little in existing AI or computer graphics that really serves this project, which is nice. We get to invent it." [AI-ED editor: If you are familiar with Doug Lenat's work, you might not be surprised to learn that Doug and Alan are friends. When Alan was at Atari, Doug consulted on the KNOESPHERE project along with ALan Borning, David McDonald, Craig Taylor & Stephen Weyer ... in alphabetical order. See IJCAI proceedings #8, p.167-169 if you are interested .. it's a bit vague and far out though] Strehlo: Who do you have working with you on this project? Kay: I've got Marvin Minsky helping on the AI stuff, I've got Seymour Papert helping on some of the curriculum design, I've got the visual language lab at MIT helping on the graphics for the animals and stuff. All different kinds of disciplines, different kinds of students, are working on it. If we can anchor the place over the next couple of years, and there's every reason to believe it's going to happen, Project Vivarium is going to be the most exciting place in the world to work. [...] ------------------------------ Date: 5 Feb 86 13:32:05 GMT From: decwrl!pyramid!pesnta!phri!greenber@ucbvax.berkeley.edu (Ross Greenberg) Subject: A contest in 'C'... There is a game making the rounds on some of the MS-DOS BBS's called CROBOTS. An interesting game that can allow those that respond to determine just how good their 'C' programming is. In this game, you program your "robot" to seek out and destroy other robots that have been programmed by someone else. Each robot has the capability of movement, sensor detection of other robots, and the ability to fire a cannon at a given direction and range. Typical robots might use programs that allow the robot to scan the playfield, locate any one of four opponents, fire a cannon at that opponent, and start zig-zagging towards that opponent while firing a cannon. If you are interested in determining how *your* robot stands up to other robots, then here are the contest rules: 1) Get a copy of the program from a local MS-DOS machine. There may be a UNIX version out, but I'm not aware of it 2) Create a robot that will (2 out of 3 times), destroy the preconfigured robots that come in the .ARC package. 3) Document your robot's code and send it off to me at the below address. Entries accepted until March 1, 1986. 4) You may enter no more than two robots. The way I'll run the contest should work, although comments are welcomed: For every four robots that come in, I'll send them off to battle. I'll run the simulation twice for each four, or until a have a clear consensus of which two out of the robots make it to the next round. This process will be repeated until there are finally only four top robots. They'll slug it out until I can determine which are the top two. From that, of course, I can determine which is the robot that deserves the applause. The top four robots will be posted to the net. Each losing robot will be returned to its designer, along with the code for the robots which destroyed it. Consider this first contest the beginning round. The next round will be in about three months. And I forgot to tell you where some of these boards are.... Two that I know of are: NYACC (New York Amateur Computer Club) at 1-718-539-3338 and my board at 1-212-889-6438, login with 'demo' and 'demo'. Happy Robot Designing.... Good Luck! Ross ross m. greenberg ihnp4!allegra!phri!sysdes!greenber [phri rarely makes a guest-account user a spokesperson. Especially not me.] ------------------------------ Date: 8 Feb 86 15:54:14 GMT From: ulysses!mhuxr!mhuxt!houxm!mtuxo!npois!npoiv!bad@ucbvax.berkeley.edu (Bruce Dautrich) Subject: Re: A contest in 'C'... This games sounds like a game called bolo which to my knowledge was first written by Peter Langston who also wrote empire. ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Fri Feb 14 05:26:16 1986 Date: Fri, 14 Feb 86 05:26:09 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a020528; 14 Feb 86 3:58 EST Date: Thu 13 Feb 1986 22:15-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #27 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Fri, 14 Feb 86 05:15 EST AIList Digest Friday, 14 Feb 1986 Volume 4 : Issue 27 Today's Topics: Query - OPS5 Demo, Cognitive Psychology - Knowledge Structures, Games & Logic - Prisoners' Dilemma Computer Programs Tournament ---------------------------------------------------------------------- Date: 6 Feb 86 22:19:00 GMT From: pur-ee!uiucdcs!convex!ctvax!kerry@ucbvax.berkeley.edu Subject: OPS5 Demo Needed Does anyone know where I can get a good production system demo that will run on the FRANZ LISP version of OPS5 (VPS)? ------------------------------ Date: Fri, 14 Feb 86 00:56:33 EST From: Mark Weiser Reply-to: mark@maryland.UUCP (Mark Weiser) Subject: Re: Cognitive Psychology - Knowledge Structures In article <8602111536.AA15674@sally.UTEXAS.EDU> sally!bulko (Bill Bulko) writes: > Chi, M. T. H., P. Feltovich, and R. Glaser, "Categorization and > Representation of Physics Problems by Experts and Novices." Cognitive > Science, Vol. 5, No. 2, April-June 1981. > The authors compare the ways experts and novices categorize physics problems > and form physical models of the problems based on the categories created. > Studies are presented which investigate the implications of the differences > found for problem solving in general. A related paper is : Mark Weiser and Joan Shertz. "Programming problem representation in novice and expert programmers." International Journal of Man-Machine Studies. December 1983. pp. 391-398. This paper is an application of some of the Chi, Feltovich, and Glaser methodology to the problem space of programming, with generically similar results. Differences in detail include categories of problem-solving used and not used by experts (algorithms yes, data-structures no), and differences between expert programmers and expert former programmers now programming managers. -mark Spoken: Mark Weiser ARPA: mark@maryland Phone: +1-301-454-7817 CSNet: mark@umcp-cs UUCP: {seismo,allegra}!umcp-cs!mark USPS: Computer Science Dept., University of Maryland, College Park, MD 20742 ------------------------------ Date: 7 Feb 86 10:08:45 PST From: MEGIDDO@IBM-SJ.ARPA Subject: Prisoners' Dilemma Computer Programs Tournament First Announcement of a COMPUTER PROGRAMS TOURNAMENT (of the Prisoners' Dilemma game) 1. INTRODUCTION This is a first announcement of a tournament for computer programs, playing the famous Prisoners' Dilemma game. Detailed instructions and some background information are provided below. The tournament is organized for the purpose of research and no prizes are offered. It is intended however that the results and winners' names will be published with permission from the persons involved. One of the goals is to see what will happen during a SEQUENCE of tournaments in which information about the participating programs will be released, so that participants will be able to revise their programs. The tournament is open to everyone. However, notice the warnings below. If you have access to electronic mail then you can participate by submitting a FORTRAN program according to the instructions below. By doing so you will also release and waive all your copyright rights and any other intellectual property rights to your program. It will also be assumed that you have not violated any rights of any third party. This announcement also includes some programs that will help you prepare for the tournament. 2. BACKGROUND The so-called prisoners' dilemma game has drawn the attention of researchers from many fields: psychology, economics, political science, philosophy, biology, and mathematics. Computer scientists are also interested in this game in the context of fundamentals of distributed systems. The game is simple to describe, does not require much skill and is yet extremely interesting from both the theoretical and practical points of view. By the (one-shot) Prisoners' Dilemma game we refer to a game as follows. The game is played by two players with symmetric roles. Each has to choose (independently of the other) between playing action C ("cooperate") or action D ("defect"). The scores to the two players, corresponding to the four possible combinations of choices of actions, are as shown in the following table: Player 2 C D --------------- | 3 | 4 | C | | | | 3 | 0 | Player 1 |-------|-------| | 0 | 1 | D | | | | 4 | 1 | --------------- Thus, both players score 3 if both play C. Both score 1 if both play D. If one plays C and the other one plays D then the one who plays C scores 0 while the other one scores 4. The prisoner's dilemma game has been the subject of many experiments. A tournament was organized several years ago by R. Axelrod who later published a book on it under the title "The evolution of cooperation" (Basic Books, Inc., New York, 1984). Following is some discussion for the benefit of readers who are not familiar with the fundamental considerations of how to play the game. One should be careful to distinguish the one-shot game from the REPEATED game in which the (one-shot) game is played many times, and after each round both players are informed of each other's actions. Furthermore, one should distinguish between the infinitely repeated game and the finitely repeated one. These seem to be quite different from the point of view of equilibrium. An equilibrium in a 2-person game is a pair (S1,S2) of strategies (one for each player) such that, given that player i (i=1,2) is playing Si , the other player, j=3-i, scores the maximum if he plays Sj . We are interested here in the finitely repeated game where the number of rounds is known in advance. We first consider the one-shot game. The analysis of the one-shot game is obvious. Each of the players realizes that no matter what his opponent does, it is always better for him to play D rather than C. Thus, under a very weak assumption of rationality (namely, players do not choose actions that are strictly dominated by other actions), the pair of actions (D,D) remains the only rational choice. The resulting score of (1,1) is inferior to (3,3), which is possible if the choices are (C,C), and this is the source of the "dilemma". To get some insight into the more general case, consider first the 2-round game. After the first round (in which the players choose independently C or D) each player is informed of the choice of the other one and then, once again, the players choose independently C or D. In this game each player has EIGHT strategies that can be coded in the form XYZ where each of X,Y and Z equals either C or D. The interpretation of this notation is as follows. (1) Play X in round 1. (2) In round 2, play Y if the opponent played C and play Z if the opponent played D. It is easy to verify that any strategy XYZ is strictly dominated by XDD (that is, regardless of what was done in round 1, and regardless of what the opponent does in round 2, it is better to play D rather than C in round 2. However, there is no domination relation between the strategies CDD and DDD: if player 2 plays DDD then player 1 is better off playing DDD rather than CDD, whereas if player 2 plays DDC, player 1 is better off playing CDD rather than DDD. Of course, strategy DDC for player 2 is dominated by DDD, but in order for player 1 to deduce that player 2 will not play DDC, he has to assume that player 2 is capable of discovering this domination. Under such an assumption player 1 can eliminate 2's DDC. Thus, if both players are "rational" they are left only with strategy DDD as a reasonable choice. A similar process of repeatedly eliminating dominated strategies applies to the general N-round game. It is dominant for both players to defect in the last round. Therefore (after we drop all strategies that play C in the last round), it becomes dominant to defect in round N-1, and so on. This eventually leaves both players only with the strategy of always playing D. The winner in both tournaments run by R. Axelrod was the simple strategy called "Tit-for-Tat". It starts by playing C and in round i+1 plays whatever the opponent played in round i. It seems like a very good strategy for playing the repeated dilemma for an indefinite number of rounds. In the N-round game it is obvious that an improvement over Tit- for-Tat would be to play Tit-for-Tat except for the last round in which the optimal play is always to defect. 3. HOW TO PARTICIPATE IN THE TOURNAMENT? If you think you understand the dilemma quite well and would like to participate in this tournament then please act according to the following instructions: 1. Design a strategy of how to play the game when the number of rounds is known in advance. The strategy should specify what to do in round 1 and at any point of the game, knowing what has been done so far and the number of rounds left, specify what to do in the next round. 2. Write a FORTRAN subroutine with the following specifications. Give it a six-letter name, for example, the first four letters of your last name followed by two initials. Suppose you picked the name JONERJ for your subroutine. Then the first line of your program should look as follows. SUBROUTINE JONERJ (N,J,I,M) The arguments are defined as follows. N - This is the total number of rounds to be played. Whenever your program is called it is told the total number of rounds and this will not change during a single game. J - This is the serial number of the round you are supposed to play in the current call. I - When J is greater than 1, this argument tells you what your opponent has played in the previous round. If I=1 it means your opponent has played C. If J=2 then he played D. Any other value is an error. M - This is what you return as your play in the current round. M=1 means you play C. M=2 means you play D. Any other value will result in an error. Your subroutine may compute anything you wish. In particular, it may keep track of the entire history of a single (N-round) game. However, it will not be able to record past games against any opponent since it will be unloaded at the end of a single N-round game. Please be reasonable with respect to the space and time you intend your program to use. Unreasonable programs will have to be dropped from the tournament at the discretion of the organizers. Also, if your program ever returns a faulty play, that is, it returns an M which is neither 1 nor 2, then it will be dropped from the tournament automatically. 3. Fill in the following information (to be transmitted only by electronic mail): NAME:_____________________________________________________________ AFFILIATION:______________________________________________________ STREET:___________________________________________________________ CITY:___________________ STATE:_____________ Zip:_______________ COUNTRY:__________________________________________________________ TELEPHONE:________________________________________________________ ELECTRONIC MAIL ADDRESS:__________________________________________ 4. Important notice! _________________________________________________________ | By sending your program to any one of the following | | addresses you agree to waive and release, to the extent | | permitted by law, all your copyright rights and other | | intellectual property rights in your computer program. | | You also warrant that no portion of your program or its | | use or distribution, violates or is protected by any | | copyright or other intellectual property right of any | | third party. You also warrant you have the right to, | | and hereby do, grant to IBM a royalty-free license to | | use your program. If any contestant is a minor under | | the laws of the state in which contestant resides, at | | least one of the contestant's parents should sign this | | warranty and license. IBM may elect to publish the | | results of the contest; names of participants or their | | submissions will not be published without the written | | approval and signature of the individual authors. | |_________________________________________________________| Please transmit your program by March 31, 1986, along with the filled questionnaire to one of the following addresses: CSNET or ARPANET: megiddo@ibm-sj VNET or BITNET : megiddo at almvma 4. TRAINING PROGRAM For your convenience, we include here an interactive program that lets you play the game with another "player". While playing this interactive program please remember that your goal is actually to SCORE high and not necessarily to BEAT the other player. In the tournament, your ability to affect the player's total score is limited since he plays against many other players besides you. Thus you will benefit if you will create "confidence" so that both of you end up playing C very often. You have the option of either playing yourself or using the subroutine that represents you. If you use a subroutine then you have to name it MINE and follow the instructions in Section 3. Simply append it the following program. It is advised that you use this option to test your own program before submitting it to the tournament. INTEGER SCORE,SCORE2,CH1,CH2,PRE1,PRE2,CC,DD,CD,DC C DATA CC,DD,CD,DC/3,1,0,4/ 20 SCORE = 0 SCORE2 = 0 PRE1=1 PRE2=1 WRITE(6,102) 102 FORMAT(' ENTER NUMBER OF ROUNDS YOU WISH TO PLAY (0=END)') 103 FORMAT (I6) READ (5,*) NR IF (NR.LE.0) STOP 118 FORMAT(' WILL YOU (1) PLAY OR WILL YOUR SUBROUTINE (2) DO? (1/2)') 430 WRITE (6,118) READ (5,*) II IF (II.EQ.2) GO TO 420 IF (II.NE.1) GO TO 430 420 DO 30 JR = 1, NR 104 FORMAT(' ROUND NO.',I6,' OF',I6,' ROUNDS. PLEASE ENTER 1 OR 2') IF (II.EQ.2) GO TO 440 WRITE (6,104) JR,NR 40 CONTINUE READ (5,*) CH1 GO TO 450 440 CALL MINE(NR,JR,PRE2,CH1) IF ((CH1-1)*(CH1-2)) 470,71,470 470 WRITE (6,117) 117 FORMAT (' YOUR SUBROUTINE RETURNED A FAULTY PLAY') GO TO 20 450 IF ((CH1-1)*(CH1-2)) 70,71,70 70 IF (CH1.EQ.0) GO TO 20 105 FORMAT(' PLEASE ENTER EITHER 1 OR 2 . (0=END)') WRITE (6,105) GO TO 40 71 IF (JR-1) 220,220,230 220 CH2 = 1 IF (NR.EQ.1) CH2 = 2 GO TO 300 230 IF (JR-NR) 250,260,260 250 CH2 = PRE1 GO TO 300 260 CH2 = 2 107 FORMAT(' PLAY WAS: YOU=',I3,' OPPONENT=',I3) 300 WRITE(6,107) CH1,CH2 IF (CH1-1) 110,110,120 110 IF (CH2-1) 130,130,140 130 SCORE = SCORE + CC SCORE2 = SCORE2 + CC GO TO 35 140 SCORE = SCORE + CD SCORE2 = SCORE2 + DC GO TO 35 120 IF (CH2-1) 150,150,160 150 SCORE = SCORE + DC SCORE2 = SCORE2 + CD GO TO 35 160 SCORE = SCORE + DD SCORE2 = SCORE2 + DD 35 WRITE (6,106) SCORE,SCORE2 106 FORMAT (' NEW TOTAL SCORE: YOU=',I5,' OPPONENT=',I5) PRE1=CH1 PRE2=CH2 30 CONTINUE GO TO 20 END 5. SAMPLE PROGRAMS For your convenience we include here copies of two sample programs. The first subroutine, called TIFRTA, plays Tit-for-Tat (see Section 2) except that it always defects in the last round. The second, called GRIM, starts playing C but switches to D the first time th opponent has played D. It also always defects in the last round. SUBROUTINE TIFRTA (N,J,IHE,MY) C C THIS IS THE TIT-FOR-TAT RULE. IN ROUND 1 PLAY 1. IN ROUND N C PLAY 0. OTHERWISE, PLAY WHAT THE OPPONENT PLAYED IN THE PRECEDING C ROUND. C C N = TOTAL NUMBER OF ROUNDS C J = CURRENT ROUND C IHE = THE CHOICE OF THE OPPONENT IN THE PRECEDING ROUND (1 OR 2) C MY = MY CHOICE FOR THE CURRENT ROUND (1 OR 2) C IF (J-1) 20,20,30 20 MY = 1 IF(N.EQ.1) MY=2 RETURN 30 IF (J-N) 50,60,60 50 MY = IHE RETURN 60 MY = 2 RETURN END C C SUBROUTINE GRIM (N,J,IHE,MY) C C THIS IS THE GRIM STRATEGY: START WITH C AND SWITCH TO D C AS SOON AS THE OPPONENT DOES C IF (J-1) 10,10,20 10 IX = 1 20 IF (IHE.EQ.2) IX = 2 IF (J.EQ.N) IX = 2 MY = IX RETURN END ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Tue Feb 18 04:48:36 1986 Date: Tue, 18 Feb 86 04:48:32 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a023065; 17 Feb 86 0:13 EST Date: Sun 16 Feb 1986 20:57-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #28 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 18 Feb 86 04:39 EST AIList Digest Monday, 17 Feb 1986 Volume 4 : Issue 28 Today's Topics: Seminars - A Design System for Engineering (MIT) & Fuzzy Logic and Common Sense Knowledge (SD Sigart) & Knowledge Engineering, Ontology (Oregon State) & Explanation-Based Learning (MIT) & Reactive Systems (SRI) & Temporal Logic for Concurrent Programs (CMU), Course - Spring Quarter Seminar on Rule-Based Systems (SU) ---------------------------------------------------------------------- Date: 13 Feb 1986 10:39 EST (Thu) From: Claudia Smith Subject: Seminar - A Design System for Engineering (MIT) [Forwarded from the MIT bboard by SASW@MIT-MC.] AN INTEGRATED DESIGN SYSTEM FOR ENGINEERING Robin J. Popplestone Edinburgh University Scotland I discuss the representation of mechanical engineering designs in a Logic programming context, and the exploration of a space of different possible designs. Designs are represented in terms of modules, which are basic concrete engineering entities (eg. motor, keyway, shaft). Modules interact via ports, and have an internal structure expressed by the part predicate. A taxonomic organisation of modules is used as the basis for making design decisions. Subsystems employed by the design system include the spatial relational inference mechanism employed in the RAPT robot Language, the Noname geometric modeller developed at Leeds University and the Press symbolic equation solver. The system is being implemented in the POPLOG system. An assumption based truth maintenance system based on the work of de Kleer is being implemented to support the exploration of design space. Tuesday, Feb. 18, 1986 4pm NE43, 8th Floor Playroom Hosts: Professors Brooks and Lozano-Perez. ------------------------------ Date: 14 Feb 86 09:01 PST From: sigart@LOGICON.ARPA Subject: Seminar - Fuzzy Logic and Common Sense Knowledge (SD Sigart) The San Diego SIGART presents FUZZY LOGIC AND COMMON SENSE KNOWLEDGE Featured Speaker: Dr. Lotfi A. Zadeh Thursday, Feb 20, 1986 6:30-8:30pm at UCSD Humanities Library Rm. 1438 Dr.Zadeh will introduce the concept of a disposition and the principle that common sense knowledge is of a dispositional nature, i.e. we can infer dispositional rules which are true in most cases. The concept of dispositionality leads to the concept of usuality or the usual value of variables. We need to develop a system for computing with and inferring from dispositional knowledge. Dr. Zadeh will show how to use fuzzy logic to deal with the concepts of dispositionality and usuality in a way which cannot be done with classical logic. Fuzzy logic will therefore be shown to provide a framework for commonsense reasoning. ------------------------------ Date: Thu, 13 Feb 86 09:46:27 pst From: Tom Dietterich Subject: Seminar - Knowledge Engineering, Ontology (Oregon State) KNOWLEDGE ENGINEERING AS THE INVESTIGATION OF ONTOLOGICAL STRUCTURE Michael J. Freiling Tektronix Laboratories Beaverton, Oregon Wednesday, February 12, l986 Cordley Hall, Room 1109 Oregon State University Corvallis, Oregon Experience has shown that much of the difficulty of learning to build knowledge-based systems lies in learning to design representation structures that adequately capture the necessary forms of knowledge. Ontological analysis is a method we have found quite useful at Tektronix for analyzing and designing knowledge-based systems. The basic approach of ontological analysis is a step-by-step construction of knowledge structures beginning with basic objects and relationships in the task domain, and continuing through representations of state, state transformations, and heuristics for selecting transformations. Formal tools that can be usefully employed in ontological analysis include domain equations, semantic grammars, and full-scale specification languages. The principles and tools of ontological analysis are illustrated with actual examples from knowledge-based systems we have built or analyzed with this method. ------------------------------ Date: Fri, 14 Feb 86 15:20 EST From: Brian C. Williams Subject: Seminar - Explanation-Based Learning (MIT) [Forwarded from the MIT bboard by SASW@MIT-MC.] Thursday , February 20 4:00pm Room: NE43- 8th floor Playroom The Artificial Intelligence Lab Revolving Seminar Series Explanation-Based Learning Tom Mitchell Rutgers University, New Brunswick, NJ The problem of formulating general concepts from specific training examples has long been a major focus of machine learning research. While most previous research has focused on empirical methods for generalizing from a large number of training examples using no domain-specific knowledge, in the past few years new methods have been developed for applying domain-specific knowledge to formulate valid generalizations from single training examples. The characteristic common to these methods is that their ability to generalize from a single example follows from their ability to explain why the training example is a member of the concept being learned. This talk proposes a general, domain-independent mechanism, call EBG, that unifies previous approaches to explanation-based generalization. The EBG method is illustrated in the context of several example problems, and used to contrast several existing systems for explanation-based generalization. The perspective on explanation-based generalization afforded by this general method is also used to identify open research problems in this area. ------------------------------ Date: Fri 14 Feb 86 18:27:38-PST From: LANSKY@SRI-AI.ARPA Subject: Seminar - Reactive Systems (SRI) AN ARCHITECTURE FOR INTELLIGENT REACTIVE SYSTEMS OR HOW NOT TO BE EATEN BY A TIGER Leslie Kaelbling SRI International AI Center and Stanford University 11:00 AM, WEDNESDAY, February 19 SRI International, Building E, Room EJ228 (new conference room) In this talk I will present an architecture for intelligent reactive systems. The ideas are fairly general, but are intended for use in programming Flakey to carry out complex tasks in a dynamic environment. Many previous robots simply 'closed their eyes' while a time-consuming system, such as a planner or vision system, was invoked, allowing perceptual inputs either to be lost or saved for later processing. In a truly dynamic world, things might change to such an extent that the results of the long calculation would no longer be useful. Worse yet, the robot might run into a wall or be eaten by a tiger. This architecture will allow the robot to remain aware during long computations, and to behave plausibly in novel situations. This talk represents work in progress, so much of the seminar will be devoted to general discussion. ------------------------------ Date: 14 February 1986 1045-EST From: Cathy Hill@A.CS.CMU.EDU Subject: Seminar - Temporal Logic for Concurrent Programs (CMU) Speaker: Aravinda Prasad Sistla Date: February 19, 1986 Time: 1:30 - 3:00 pm Place: WeH 4623 Title: ON EXPRESSING SAFETY AND LIVENESS PROPERTIES IN TEMPORAL LOGIC. Correctness properties of concurrent programs are usually classified as either safety properties or liveness properties. In general, proving a program correct involves in establishing that the program satisfies certain safety properties and certain liveness properties, and usually different techniques are applied in proving these properties. In this talk we consider many different definitions of these properties (e.g. safety,strong safety,liveness, absolute liveness etc.) and investigate what classes of these properties are expressible in temporal logic. We present syntactic characterization of formulae that express these properties. Finally, we give algorithms to recognize if a temporal specification is a safety property or liveness property. ------------------------------ Date: Wed 12 Feb 86 15:47:58-PST From: Ted Shortliffe Subject: Course - Spring Quarter Seminar on Rule-Based Systems (SU) [Forwarded from the Stanford bboard by Laws@SRI-AI.] SEMINAR ON RULE-BASED EXPERT SYSTEMS Professors Buchanan and Shortliffe Comp. Sci. 524 Med. Inf. Sci. 229 Spring Quarter 1986 - 2 units Tuesday, 3:30-5:00PM TC-135 Conference Room, Medical Center [Class size limited to 16] This course is a graduate seminar for students wishing to gain a technical understanding of, as well as a historial perspective on, rule-based expert systems. The emphasis of the course will be on an analysis of the research lessons of MYCIN and related projects in the Knowledge Systems Laboratory, the strengths and limitations of the rule-based approach to knowledge representation, and the way in which AI research evolves as new ideas and concepts are discovered. The course will meet weekly for 90 minutes and will require substantial reading assignments for each session. The required text for the seminar is "Rule-Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project"; additional related papers will also be assigned. Working in pairs, all students will be responsible for leading the discussion once during the quarter. There will be a final exam. Prerequisites: at least one course in artificial intelligence and familiarity with LISP. Enrollment: limited to 16; signup in TC-135 or by contacting Ms. Alison Grant (GRANT@SUMEX or 7-6979). If the course is oversubscribed, preference will be given as follows: MS/AI and MIS grad students, other CSD grad students, non-CSD graduate students and medical students, CSD research staff, undergraduates, auditors. 2 units, Tu 3:30-5:00, Room TC-135 (Medical Center), Professors Buchanan and Shortliffe. The course will not be offered again until 1987-88. April 1: INTRODUCTION Readings: None April 8: KNOWLEDGE ENGINEERING Readings: Chapters 1,7,35,8,9 [Chapter 4 suggested before 7 for those unfamiliar with MYCIN] April 15: USING RULES Readings: Chapters 2,3,5,6 April 22: REASONING UNDER UNCERTAINTY Readings: Chapters 10,11,12,13 [updated version of Chapter 13 will be provided] April 29: GENERALIZED FRAMEWORKS Readings: Chapters 14,15,16,33 May 6: OTHER REPRESENTATIONS OF KNOWLEDGE Readings: Chapters 21,22,23,24 May 13: EXPLANATIONS/TUTORING Readings: Chapters 17,18,20,25,26 May 20: META-LEVEL KNOWLEDGE Readings: Chapters 27,28,29 May 22 (Thursday class, 3:30-5pm): EVALUATING PERFORMANCE Readings: Chapters 30,31 May 27: no class Readings: Chapters 32,34,36 June 3: SUMMARY AND CONCLUSIONS Readings: None ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Tue Feb 18 05:05:49 1986 Date: Tue, 18 Feb 86 05:05:44 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a023209; 17 Feb 86 0:45 EST Date: Sun 16 Feb 1986 21:16-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #29 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 18 Feb 86 04:42 EST AIList Digest Monday, 17 Feb 1986 Volume 4 : Issue 29 Today's Topics: Queries - Expert Systems Information & Rule Master Reviews & Games, Evolution and Learning Conference & Chess & Micro Prolog, Bindings - Prisoner's Dilemma Mailing List, Machine Learning - Hopfield Networks, Software Review - Personal Computer Scheme ---------------------------------------------------------------------- Date: Sun 16 Feb 86 22:36:39-EST From: "Randall Davis" Reply-to: davis@mit-mc Subject: access to info I'd like to assemble a list of resources of information about expert systems, organized along the lines indicated below. If you can think of a journal, trade magazine, newsletter, or regularly scheduled conference not listed below, and can supply the relevant details, please send them to me (not to the whole list, and please only respond if you have the details available and accurate). I'll filter the responses to eliminate duplicates and re-post to the list it for general consumption. ++++++++++++++++++++++++++++++++++++++++ FORMAT FOR PUBLICATIONS NAME PUBLISHER EDITOR(S) SUBSCRIPTION INFORMATION (frequency of publication, price) SUBSCRIPTION ADDRESS (where to write) CATEGORY: one of RESEARCH JOURNAL (eg, Artificial Intelligence) RESEARCH NEWSLETTER (eg, AAAI Magazine, SIGART) COMMERCIAL NEWSLETTER (eg, Expert Systems Strategies) FOCUS: eg, all areas of AI, expert systems technical issues, management issues, etc. FORMAT FOR CONFERENCES NAME SPONSORING ORGANIZATION FREQUENCY OF OCCURENCE ADDRESS FOR INFORMATION I have details for Journals AI Journal Journal of Automated Reasoning Newsletters AAAI Magazine Expert Systems Strategies Conferences IJCAI, AAAI and would welcome all other info, especially non-US listings. ------------------------------ Date: Fri 14 Feb 86 12:56:42-PST From: Bill Park Subject: Rule Master Reviews? To whom it may concern: If you have any experience with Rule Master, would you please tell me what you think of it? We are considering using it in a project related to NASA's space station. Thanks, Bill Park (Park@SRI-AI) (415)859-2233 SRI International Menlo Park, CA ------------------------------ Date: 14 Feb 86 02:43:55 GMT From: nike!im4u!ut-sally!ut-ngp!gknight@ucbvax.berkeley.edu (gknight) Subject: Games, Evolution & Learning Conference query. Can anyone give me information on a conference entitled "Games, Evolution & Learning," held in New Mexico in 1984 or 1985? The organizer (or any other contact person)? Proceedings, if available? A list of speakers and paper titles? Etc., etc. Please send by mail directly to me and I'll post a summary of info received for the information of others on the nets. Many thanks, Gary Knight, 3604 Pinnacle Road, Austin, TX 78746 (512/328-2480). Biopsychology Program, Univ. of Texas at Austin. "There is nothing better in life than to have a goal and be working toward it." -- Goethe. ------------------------------ Date: 14 Feb 86 09:42 EST From: Vu.wbst@Xerox.COM Subject: Chess game informations needed. I'm reading about expert system, and would like to try to build an expert system. I would appreciate any helpfull hints, pointers to any existing Chess game expert system, in Interlisp-D would a plus. I would like to thank you in advance for any help. Dinh Regular mail: Dinh Vu Xerox Corporation 800 Philips Rd, Bld 129-38B Webster, Ny 14580. ------------------------------ Date: Fri 14 Feb 86 16:56:18-EST From: FWHITE@G.BBN.COM Subject: Prolog on VMS and/or MAC Does anybody know of a public domain version of Prolog for VAX/VMS or the Macintosh? Or how about a commercial version? Jeff Berliner (BERLINER@G.BBN.COM) ------------------------------ Date: 15 Feb 86 18:43:00 PST From: MEGIDDO@IBM-SJ.ARPA Subject: Prisoner's Dilemma Prisoner's dilemma tournament mailing list; Please send back a note if you wish to receive future announcements. ------------------------------ Date: 7 Feb 86 20:13:13 GMT From: decwrl!pyramid!ut-sally!mordor!ehj@ucbvax.berkeley.edu (Eric H Jensen) Subject: Re: Hopfield Networks? In article <1960@peora.UUCP> jer@peora.UUCP (J. Eric Roskos) writes: >In a recent issue (Issue 367) of EE Times, there is an article titled >"Neural Research Yields Computer that can Learn". This describes a >simulation of a machine that uses a "Hopfield Network"; from the ... I got the impression that this work is just perceptrons revisited. All this business about threshold logic with weighting functions on the inputs adjusted by feedback (i.e. the child reading) ... Anybody in the know have a comment? eric h. jensen (S1 Project @ Lawrence Livermore National Laboratory) Phone: (415) 423-0229 USMail: LLNL, P.O. Box 5503, L-276, Livermore, Ca., 94550 ARPA: ehj@angband UUCP: ...!decvax!decwrl!mordor!angband!ehj [What is new is that there are now training algorithms for multilayer networks -- something that Minsky and Papert declared unlikely in their famous book on perceptrons. Also new is the development of special hardware, both chips and full [Boltzmann] processors for the implementation of such networks. Hopfield networks require symmetric connections and a form of "relaxation" processing or simulated annealing; Hopfield characterizes this as constraint satisfaction or discrete optimization by moving through the center of a data space (in somewhat the same manner as the Karmarkar algorithm) instead of touring the vertices in the manner of the simplex algorithm. Other multilayer connectionist networks have recently been developed that do not require symmetric or even feedback connections, except for training feedback. The breakthrough in these latter networks seems to be the notion of adjusting each coefficient in proportion to its "responsibility" in making a good or bad decision. Determination of proportionate responsibility can be made using partial derivatives. Another possibility that I find intriguing is the use of a domain-knowledgeable expert system for identifying "guilty" coefficients, as in the system for predicting horse races reported in Heuristics for Inductive Learning by Steven Salzberg of Applied Expert Systems, IJCAI 85, pp. 603-609. -- KIL] ------------------------------ Date: Tue 11 Feb 86 22:38:38-CST From: Rob Pettengill Subject: Personal Computer Scheme I recently purchased an implementation of the Scheme dialect of lisp for my PC. I am familiar with GC Lisp, IQ Lisp, and Mu Lisp for the PC. I use Lambdas and 3600s with ZetaLisp at work. TI PC Scheme is a very complete implementation of scheme for the IBM and TI personal computers and compatibles. It combines high speed code execution, a good debugging and editing environment, and very low cost. The Language: * Adheres faithfully to the Scheme standard. * Has true lexical scoping. * Prodedures and environments are first class data objects. * Is properly tail recursive - there is no penalty compared to iteration. * Includes window and graphics extensions. The Environment: * An incremental optimizing compiler (not native 8086 code) * Top level read-compile-print loop. * Interactive debugger allows run time error recovery. * A minimal Emacs-like full screen editor with a scheme mode featuring parethesis matching and auto indenting of lisp code. * An execute DOS command or "push" to DOS capability - this is only practical with a hard disk because of the swap file PCS writes. * A DOS based Fast Load file format object file conversion utility. * A fast 2 stage garbage collector. First Impressions: Scheme seems to be much better sized to a PC class machine than the other standard dialects of lisp because of its simplicity. The TI implementation appears to be very solid and complete. The compiled code that it produces (with debugging switches off) is 2 to 5 times faster than the other PC lisps that I have used. With the full screen editor loaded (there is also a structure editor) there seems to be plenty of room for my code in a 640k PC. TI recommends 320k or 512k with the editor loaded. The documentation is of professional quality (about 390 pages), but not tutorial. Abelson and Sussman^2's "Structure and Interpretation of Computer Programs" is a very good companion for learning scheme as well as the art and science of programming in general. My favorite quick benchmark - (define (test n) (do ((i 0 (1+ i)) (r () (cons i r))) ((>= i n) r))) runs (test 10000) in less than 10 seconds with the editor loaded - of course it takes a couple of minutes to print out the ten thousand element list that results. The main lack I find is that the source code for the system is not included - one gets used to that in good lisp environments. I have hit only a couple of minor glitches, that are probably pilot error, so far. Since the system is compiled with debugging switches off it is hard to get much useful information about the system from the dubugger. Based on my brief, but very positive experience with TI PC scheme and its very low price of $95 - I recommend it to anyone interested in a PC based lisp. You can order from Texas Instruments at 1-800-TI-PARTS. (Standard disclaimers about personal opinions and having no commercial interest in the product ...) Rob Pettengill ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Mon Feb 17 04:19:05 1986 Date: Mon, 17 Feb 86 04:19:02 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a023543; 17 Feb 86 1:55 EST Date: Sun 16 Feb 1986 22:35-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V4 #30 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Mon, 17 Feb 86 04:04 EST AIList Digest Monday, 17 Feb 1986 Volume 4 : Issue 30 Today's Topics: Query - Ambiguous Sentences, Cognitive Psychology - Definition & Novice-Expert Differences, Theory - Dreyfus' Technology Review Article ---------------------------------------------------------------------- Date: Fri 14 Feb 86 14:29:21-PST From: FIRSCHEIN@SRI-AI.ARPA Subject: Ambiguous sentences. I wonder whether AILIST readers have a favorite short sentence for illustrating multiple ambiguity, say greater than 5 meanings? ------------------------------ Date: Sat, 15 Feb 86 21:38:39 EST From: bzs%bostonu.csnet@CSNET-RELAY.ARPA Subject: Re: Sparklers from the Tech Review >From: larry@Jpl-VLSI.ARPA >COGNITIVE PSYCHOLOGY (a more restricted area than Cognitive Science) attempts >to understand biologically based thinking using behavioral and psychiatric >concepts and methods. This includes the effects emotional and social forces >exert on cognition. This group is increasingly borrowing from the following >groups. Just curious, but as an undergraduate studying such things at Cornell in the early 70's I remember being lectured over and over again about 'Cognitive Psychology' which at that point in time seemed to be a school derived largely from Festinger's works in Cognitive Dissonance et al (and Brehm and others.) It was generally posed as being orthogonal to behaviorism (Skinnerianism.) Is this the same 'cognitive psychology' I suffered through? Or has the term changed? What do they call the old stuff, or are we allowed to speak of it anymore (oops)? I suppose this definition *might* be referring to the same thing, but I don't see how. -Barry Shein, Boston University ------------------------------ Date: 14 Feb 86 14:38:36 EST (Fri) From: Robert Rist Subject: novice-expert differences You can trace back the articles you need if you look at Snow, R. E., Federico, P. & Montague, W. E. (Eds.). (1980) Aptitude, learning and instruction, Volume 2. This has articles by VanLehn and Brown, Stevens and Collins, Anderson and Norman. Lesgold, A. M. (1984). Acquiring expertise. In Anderson and Kosslyn (Eds.), Tutorials in learning and memory. Pointers to lots of different research domains. Chi, M. T. H., Glaser, R. & Rees, E. (1982). Expertise in problem solving. In Sternberg, R. J. (Ed.), Advances in the psychology of human intelligence. This is one of the best summary articles I have seen. Anderson, J. R. (Ed.) (1981). Cognitive skills and their acquisition. A mixed bag, but it contains some real classics. Gentner, D. & Stevens, A. L. (1983). Mental models. The stuff on multiple models and debugging is very interesting. If you're interested in learning, you could also look at Anzai, Y. (1984). Cognitive control of real-time event-driven systems. Cognitive Science, 8, 221-254. Anzai, Y. & Simon, H. A. (1979). The theory of learning by doing. Psych. Review, Vol 86, 124-140. Anderson, R. J. (1985). Cognitive psychology and its implications. This has a chapter on expertise development that gives an overview plus list of references. Have fun, Rob Rist ------------------------------ Date: Thu, 13 Feb 86 09:44:18 est From: rjk@mitre-bedford.ARPA (Ruben) Subject: Response to "Thompson@umass-cs.csnet" re: "Expertize" In lieu of replying to the apparently incorrect address "Thompson@umass-cs.csnet", I send my tidbit to the AILIST. >From an abstract but empirically motivated view of the relationship between expertise and category formation (a criterion useful for discriminating relatively novice from expert behavior), I suggest Eleanor Rosch's (U. of C. at Berkeley) work on prototypes. A particularly good SUMMA is her article "Human Categorization," of which I read in draft form but which SHOULD (?) have been published in ADVANCES IN CROSS-CULTURAL PSYCHOLOGY (Vol. 1), M. Warren (ed.), Academy Press, London, circa 1976. I think that her approach to categorization raises some intelligent and persuable questions about the role of expertize in categorization: this article is worth reading, even if it only touches on this question. Rosch planned to do further research to follow up her questions viz. expertize, but I have not yet seen it. (Let me know if you follow this up.) Ruben J. Kleiman rjk@MITRE-BEDFORD [The address Thompson%UMASS-CS.CSNet@CSNet-Relay should work (regardless of capitalization). The gateway requires that all CSNet mail from the Arpanet be addressed to @CSNet-Relay, and that all other @-signs be changed to %-signs. The .CSNet prior to the @CSNet-Relay is sometimes optional. -- KIL] ------------------------------ Date: 8 Feb 86 00:35:57 GMT From: decwrl!glacier!kestrel!ladkin@ucbvax.berkeley.edu Subject: Re: Technology Review article In article <15030@rochester.UUCP>, lab@rochester.UUCP (Lab Manager) writes: > "In 3000 years, Philosophy has still not lived up to its promises and > there's no reason to think it ever will." An interesting comment. Whenever a problem is solved in Philosophy, it spawns a whole new field of specialists, and is no longer called Philosophy. Witness Physics, which used to be called Natural Philosophy. When Newton took over, it gradually became a new subject. Witness our own subject, which arose out of the attempts of Frege to provide a formal foundation for mathematical reasoning, via Russell, Church, Curry, Kleene, Turing and von Neumann. Much work in natural language understanding arises from the work of Montague, and more recently speech act theory is being used, from Grice, Searle and Vanderveken. The list goes on, and so do I. Would that AI bear such glorious fruit. I think it might. Peter Ladkin ------------------------------ Date: 9 Feb 86 16:05:00 GMT From: pur-ee!uiucdcs!uiucuxc!bantz@ucbvax.berkeley.edu Subject: Re: Technology Review article Dreyfus's book "What Computers Can't Do" was a pretty sorry affair, insofar as it purported to have a positive argument about intrinsic limits of computers. However uncomfortable it makes the AI community feel, though, the journalistic baiting with extensive quotations from the AI community itself, ought to have demonstrated the virtues of a bit more humility than is often shown. [I'm refering to his gleeful quotation of predictions that, by 1970 or so a computer would be world chess champion, that fully literate translations of natural languages would be routine...] The responses here, so far, seem to be guilty of what Dreyfus is accused of: failing to engage the opponent seriously, and relying on personal expressions of distaste or ridicule. Specifically, Dreyfus does reject the typology of learning in AI, on the not implausible grounds that it is self-serving, and not obviously correct (or uniquely correct). [Please! I am *not* a fan of Dreyfus, and do not endorse most of his claims.] ------------------------------ Date: Sun 16 Feb 86 22:33:41-PST From: Ken Laws Reply-to: AIList-Request@SRI-AI.ARPA Subject: In Support of the Dreyfi I have now had a chance to read the Technology Review article (thanks to a copy from Oscar Firschein). If it is a fair sample of Hubert and Stuart Dreyfus' forthcoming book, Mind Over Machines, the book should be required reading. Not that I necessarily agree with their positions -- I see their points as problems to be solved rather than proofs of futility -- but they have now solidified their stronger arguments and (I presume) shed many of their weaker ones. I recently read the introduction to the second edition of Hubert's What Computers Can't Do and found myself disagreeing with about one item per page. (To be fair, they or anyone else would find similar disagreement with my [fuzzy] ideas if I had the ability and temerity to expose them in writing.) I did not experience anywhere near the same density of objections to this new article, Why Computers May Never Think Like People. I am optimistic that we will be able to build "digital" intelligences (with perhaps a few analog circuits thrown in as necessary), but I cannot support my optimisim as well as they support their pessimism. They are right that the AI "paradigms" of the past have proven weak and inextensible, and that those of the present are also likely to fail. (Five years hence, will not each researcher's proposals start with "Previous work in this field has had limited success due to ..., but our new approach will ...?) They are wrong to assume that the logic-based symbol-processing paradigm is the only card AI holds. (Sorry, guys, but I'm not a logic lover. Explicit definitions and rules for commonsense reasoning are a useful exercise, but flexible -- and sometimes errorful -- intelligence will ultimately depend on a patchwork of heuristics and analogies.) Many of the "feature vs aspect" problems raised by the brothers are being faced by those of us researching perception. Our results are sparse to date, but that is no proof that pattern recognition and concept formation are inherently human capabilities. Hubert and Stuart, as the loyal opposition to past naivete, may help us to face and overcome the true difficulties in real-world intelligence -- if they don't get our funding killed first. -- Ken Laws ------------------------------ End of AIList Digest ********************