IRList Digest Wednesday, 25 November 1987 Volume 3 : Issue 42 Today's Topics: Query - Matching knowledge representations - Help needed in searching for AIDS research - Hypertext thesis info. Discussion - Hyperdocument discussion Announcement - HUMANIST group COGSCI - Using machine readable Longman's Dictionary for NLP - Dynamic connectionism, Threshold of knowledge, Ideonomy News addresses are Internet or CSNET: fox@vtcs1.cs.vt.edu BITNET: fox@vtcs1.bitnet ---------------------------------------------------------------------- Date: Wed, 18 Nov 87 14:15 EDT From: LEWIS%cs.umass.edu@RELAY.CS.NET Subject: Refs on deciding relevance by matching knowledge reps Dear IRLIST: I would be interested in hearing of references on determining the relevance of documents to a query based on matching some sort of non-surface text representations of the contents of document and query. Best example is a recent dissertation by DeeAnn Emmel Lewis (no relation) which looks at the correlation between relevance judgments and similarity of case frame representations of query and document. Other examples which hover around the edge of this topic would be Farradane's relational indexing, use of thesauri, work on indexing on multi-word phrases (especially ala Sparck-Jones and Tait, where phrases are generated by NLP techniques), etc. Many thanks, David D. Lewis CSNET: lewis@cs.umass.edu COINS Dept. BITNET: lewis@umass University of Massachusetts, Amherst Amherst, MA 01003 [Note: Some points that come to mind -- dissertation by S. Weiss, work by Oddy and Belkin on ASKs, work on IR for chemistry. Please let us know what others you collect. - Ed] ------------------------------ Date: Mon, 23 Nov 87 17:24:15 est From: Dr Anna Olesinska Serafin, Kummelbyvagen 1K, 191 40 Sollentuna, Sweden. Subject: AIDS Research for Poland [Forwarded from CRTNET 111 - November 23, 1987 - Ed] I am a doctor from the Department of Education at the Warsaw Medical School, Poland, and currently here in Sweden as a guest scientist doing research on AIDS. I have come to know about computer communication during my desperate search for information on AIDS and I hope that this electronic message and request will help provide me with some answers that I could not find in databases and libraries. The information obtained will be tremendously useful to the Ministry of Health in Poland for the formulation of the National AIDS program. I would also appreciate it if you could also copy this message either electronically or in print to others whom you think might help. Your answers could be mailed to P2269@QZCOM or to me by post. Thank you in anticipation. I would like to seek CONTACT SOURCES and REFERENCES TO PUBLICATIONS on the following : 1) guidelines for AIDS campaigns (local, national, international) 2) programs for public AIDS education and prevention and for the "risk groups", and their evaluations. 3) AIDS testing programs and regulations [Note: I included this since I figure there are lots of sources that could be consulted that Poland could be made aware of. - Ed] ------------------------------ Date: Tue, 17 Nov 87 22:35:20 EST From: grads014@grumpy.cis.guelph.netnorth To: hypertext thesis I am currently researching browsing and navigation problems in large loosely structured information bases (hypertext). I intend to write a thesis on this topic towards a Master of Science in Computing & Information Science at the University of Guelph, Ontario, Canada. Could you ple [Note: message was cut off - I expect David is looking for references. There was a very nice bibliography handed out at Hypertext '87 from folks at Brown - I suggest that as a good starting point. - Ed] grads014%uogvax2.BITNET@wiscvm.wisc.edu David Hendry ------------------------------ Date: Mon, 23 Nov 87 16:10:40 EST From: rada@mcs.nlm.nih.gov (Roy Rada CSB) Subject: hyperdocument discussion Ed, In the latest IRList Digest there was a call for interest in a hypertext discussion group (through a note from Scott Peterson). I'm interested in participating. Yours, Roy ------------------------------ Date: 17 November 1987, 16:04:16 EST From: MCCARTY at UTOREPAS Subject: HUMANIST group Yaacov Choueka (Bar-Ilan Univ.) sent me a reference concerning your electronic discussion group. I would be very interested in hearing more about it. Would you kindly send me whatever information you have? Thanks very much. I run a ListServ discussion group, HUMANIST, that may interest you. I append information about it below. Yours, Willard McCarty What is HUMANIST? HUMANIST is a Bitnet/NetNorth/EARN discussion group for people who support computing in the humanities. Those who teach, review software, answer questions, give advice, program, write documentation, or otherwise support research and teaching in this area are included. Although HUMANIST is primarily intended to help these people exchange all kinds of information, it is primarily meant for interaction rather than publication or advertisement. In general, members of the network are encouraged to ask questions and offer answers, to begin and contribute to discussions, to suggest problems for research, and so forth. One of the specific motivations for establishing HUMANIST was to allow people involved in this area to form a common idea of the nature of their work, its requirements, and its standards. Institutional recognition is not infrequently inadequate, at least partly because computing in the humanities is an emerging and highly cross-disciplinary field. Its support is significantly different from the support of other kinds of computing, with which it may be confused. HUMANIST is a project of the Special Interest Group for Humanities Computing Resources, which is affiliated with the Association for Computing in the Humanities (ACH) and the Association for Literary and Linguistic Computing (ALLC). Although participants in HUMANIST are not required to be members of either organization, membership is highly recommended. Information about the SIG may be obtained from George Brett (ecsghb@tucc.bitnet), Michael Sperberg-McQueen (u18189@uicvm.bitnet), or the undersigned. New members of HUMANIST are welcome, provided that they fit the broad guidelines described above. To subscribe send a message to me at the network address below, giving your name, address, telephone number, and a short description of what you do to support computing in the humanities. (This description will eventually be distributed to all HUMANISTs as a means of introducing you to the community and helping us define ourselves professionally. It may subsequently be printed in the Newsletter of the ACH unless you declare otherwise.) Please describe your academic background and research interests, both in computing and otherwise; the nature of your job; and, if relevant, its place in your university. Willard McCarty Editor of HUMANIST Centre for Computing in the Humanities University of Toronto (MCCARTY@UTOREPAS.BITNET) ------------------------------ Date: Mon, 9 Nov 1987 10:28 EST From: Peter de Jong Subject: Cognitive Science Calendar [Extract - Ed] Date: Sunday, 8 November 1987 17:42-EST From: Marc Vilain Re: BBN AI Seminar -- Bran Boguraev BBN Science Development Program AI Seminar Series Lecture THE USE OF AN ON-LINE DICTIONARY FOR NATURAL LANGUAGE PROCESSING Bran Boguraev Computer Laboratory, University of Cambridge (UK) (bkb%computer-lab.cambridge.ac.uk@NSS.Cs.Ucl.AC.UK) BBN Labs 10 Moulton Street 2nd floor large conference room 10:30 am, Friday November 13 This talk is an attempt at a retrospective analysis of the collective experience stemming from the use of the machine-readable version of the Longman Dictionary of Contemporary English for natural language processing. It traces the relationships between specific requirements for lexical data and issues of making such data available for diverse research purposes. A particular model of on-line dictionary use is presented, which promotes a strong separation between the processes of extracting information from machine-readable dictionaries and using that information within the pragmatic context of computational linguistics. The talk further analyses some characteristics of the raw lexical data in electronic sources and outlines a methodology for making maximal use of such potentially rich, but inherently unreliable, resources. ------------------------------ Date: Tue, 10 Nov 1987 10:12 EST From: Peter de Jong Subject: Cognitive Science Calendar [Extract - Ed] Date: Monday, 9 November 1987 12:20-EST From: Elizabeth Willey Re: CBIP CONNECTIONIST SEMINAR DYNAMICAL CONNECTIONISM Elie Bienenstock Universite de Paris-Sud Wednesday, 11 November E25-406, 12:00 In connectionist models, computation is usually carried out in a space of activity levels, the connectivity state being frozen. in contrast, dynamical connectionist models manipulate connectivity states. For instance, they can solve various graph matching problems. They also have the typical associative memory and error-correcting properties of usual connectionist models. Applications include invariant pattern recognition; dynamical connectionist models are able to generalize over transformation groups rather than just Hamming distance. It is proposed that these principles underlie much of brain function; fast- modifying synapses and high-resolution temporal correlations may embody the dynamical links used in this new connectionist approach. ============== NE43, 8TH FLOOR THUR, 11/12, 4:00PM ON THE THRESHOLD OF KNOWLEDGE The Case for Inelegance Dr. Douglas B. Lenat Principal Scientist, MCC In this talk, I would like to present a surprisingly compact, powerful, elegant set of reasoning methods that form a set of first principles which explain creativity, humor, and common sense reasoning -- a sort of "Maxwell's Equations" of Thought. I'd like very much to present them, but, sadly, I don't believe they exist. So, instead, I'll tell you what I've been working on down in Texas for the last three years. Intelligent behavior, especially in unexpected situations, requires being able to fall back on general knowledge, and being able to analogize to specific but far-flung knowledge. As Marvin Minsky said, "the more we know, the more we can learn". Unfortunately, the flip side of that comes into play every time we build and run a program that doesn't know too much to begin with, especially for tasks like semantic disambiguation of sentences, or open-ended learning by analogy. So-called expert systems finesse this by restricting their tasks so much that they can perform relatively narrow symbol manipulations which nevertheless are interpreted meaningfully (and, I admit, usefully) by human users. But such systems are hopelessly brittle: they do not cope well with novelty, nor do they communicate well with each other. OK, so the mattress in the road to AI is Lack of Knowledge, and the anti-mattress is Knowledge. But how much does a program need to know, to begin with? The annoying, inelegant, but apparently true answer is: a non-trivial fraction of consensus reality -- the few million things that we all know, and that we assume everyone else knows. If I liken the Stock Market to a roller-coaster, and you don't know what I mean, I might liken it to a seesaw, or to a steel spring. If you still don't know what I mean, I probably won't want to deal with you anymore. It will take about two person-centuries to build up that KB, assuming that we don't get stuck too badly on representation thorns along the way. CYC -- my 1984-1994 project at MCC -- is an attempt to build that KB. We've gotten pretty far along already, and I figured it's time I shared our progress, and our problems, with "the lab." Some of the interesting issues are: how we decide what knowledge to encode, and how we encode it; how we represent substances, parts, time, space, belief, and counterfactuals; how CYC can access, compute, inherit, deduce, or guess answers; how it computes and maintains plausibility (a sibling of truth maintenance); and how we're going to squeeze two person-centuries into the coming seven years, without having the knowledge enterers' semantics "diverge". ============== Friday, 13 November 12:00pm E25-401 Ideonomy: Founding a 'Science of ideas' In a book published in 1601, Francis Bacon urged that modern science should have the equivalent of an 'ideonomic' character, as well as being based on experimentation and induction. My talk concerns a five-year effort to lay foundations for a science of ideas which I call Ideonomy. Whereas the field of Artificial intelligence is primarily aimed at the automation of mind, cognitive science at the modeling of human intelligence and thought, and logic at the formalization of reasoning, ideonomy is preoccupied with the discovery, classification, and systematization of universal ideas, with aiding and abetting man's use of ideas, and with automating the generation of ideas. The ideonomist holds that inattention to the latter things has hobbled the development, and limited the success of the other fields; and that properly all four subjects should be developed simultaneously and in close coordination, being mutually necessary and synergistic. At present ideonomy is divided into some 320 subdivisions, a few of which are: the study of ignorance, the study of analogies, the study of form, the study of causes, the study of questions, the study of answers, the study of processes, and the study of cognitive and heuristice principles. In each of these cases it seeks to identify: the types (of these things), higher and lower taxa, examples, interrelationships, causes, effects, reasons for studying, needed materials and methods, fundamental concepts, abstract and practical relations to other ideonomic divisions, and the like. We can also characterize ideonomy in another way, such as: the study of how elementary ideas can be combined, permuted, and trnsformed as exhaustive groups of ideas; A new language designed to facilitate thought and creativity; An attempt to exploit the qualitiative laws of the universe. ------------------------------ END OF IRList Digest ********************