Date: Sat 13 Feb 1988 22:50-PST From: AIList Moderator Kenneth Laws Reply-To: AIList@KL.SRI.COM Us-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList V6 #29 - Seminars, Conferences To: AIList@KL.SRI.COM Status: R AIList Digest Sunday, 14 Feb 1988 Volume 6 : Issue 29 Today's Topics: Policy - Seminars, Seminar - New Logics for Linguistic Descriptions (BBN) & The Power of Vacillation (SUNY) & Qualitative ODEs using Linear Approximations (UToronto), Conference - Computational Learning Theory & 3rd European Working Session on Learning 1988 & Micros, Expert Systems in Planning, Transport, Building ---------------------------------------------------------------------- Date: Fri, 5 Feb 88 09:07:24 PST From: rifrig@Sun.COM (Christopher Rigatuso) Subject: Policy - Seminars How come we get seminar announcements for Feb 4th, on Feb 5th? It seems that this type of thing happens occasionally on this alias. --Chris. [I used to send out seminar notices as they came in, trying to maximize their usefulness. Readers found this annoying, however, because lengthy seminar and conference notices were mixed into nearly every digest. The list members seem to prefer having conference notices segregated so that this substream may be easily skipped or archived. Such collection introduces a delay of up to a week. Note that the purpose of seminar abstracts (N.B.: abstracts are required) on AIList is to inform those who can't attend seminars rather than to alert those who can. -- KIL] ------------------------------ Date: Tue 9 Feb 88 09:03:58-EST From: Dori Wells Subject: Seminar - New Logics for Linguistic Descriptions (BBN) BBN Science Development Program AI/Education Seminar Series SOME NEW LOGICS FOR LINGUISTIC DESCRIPTIONS William Rounds CSLI, Stanford University Xerox PARC (ROUNDS@Russell.Stanford.EDU) BBN Laboratories Inc. 10 Moulton Street Large Conference Room, 2nd Floor 10:30 a.m., Tuesday, February 23, 1988 Abstract: Unification-based grammar formalisms typically use attribute- value matrices as repositories of information derived from utterances. In previous work we have shown how to represent grammatical specifications as logical formulas which speak directly about these matrices. This involved the use of a particularly simple form of deterministic propositional dynamic logic. In this talk, we will review this logic, and then discuss how to extend the logic to speak about set-valued matrices, which involves a notion of nondeterminism. Examples will be given involving modeling common knowledge as a certain non-wellfounded set (its elements include the set itself), and some coordination phenomena in lexical-functional grammar. Each example illustrates a particular kind of logical expression. ------------------------------ Date: 9 Feb 88 15:29:07 GMT From: decvax!sunybcs!rapaport@ucbvax.Berkeley.EDU (William J. Rapaport) Subject: Seminar - The Power of Vacillation (SUNY) STATE UNIVERSITY OF NEW YORK AT BUFFALO DEPARTMENT OF COMPUTER SCIENCE COLLOQUIUM THE POWER OF VACILLATION John Case Department of Computer Science State University of New York at Buffalo Recursion theory provides a relatively abstract, elegant account of the absolute boundaries of computability by discrete machines. The insights it can provide are best described as philosophical. In this talk I examine a sub- part of this theory pertaining to machine learning, specifi- cally, in this case, language learning. I will describe Gold's influential, recursion theoretic, language-learning paradigm (and variations on the theme), point out its easily seen, considerable weaknesses, but then argue, by means of example theorems, that it is possible, nonetheless, to obtain some insights into language learning within the general context of this paradigm. For example, I will squeeze some insight out of a theorem to the effect that allowing a kind of vacillation in the convergent behavior of algorithmic, language-learning devices leads, perhaps unexpectedly, to greater learning power. I'll sketch the proofs of a couple of the theorems, in part to convince you they are true, but mostly because the proofs are beautiful and illustrative of techniques in the area. Date: Thursday, 11th February, 1988 Time: 3:30 pm to 4:30 pm Place: Bell 337, Amherst Campus Wine and Cheese will be served at 4:30 pm at Bell 224. For further information, call (716) 636-3199. ------------------------------ Date: 11 Feb 88 22:14:07 GMT From: Armin Haken Subject: Seminar - Qualitative ODEs using Linear Approximations (UToronto) There will be an AI seminar on Tuesday 23 February at 2PM in room SF 1105, given by Dr. Elisha Sacks of MIT. [...] Hosting is Hector Levesque. Automatic Qualitative Analysis of Ordinary Differential Equations Using Piecewise Linear Approximations by Elisha Sacks This talk explores automating the qualitative analysis of physical systems. Scientists and engineers model many physical systems with ordinary differential equations. They deduce the behavior of the systems by analyzing the equations. Most realistic models are nonlinear, hence difficult or impossible to solve explicitly. Analysts must resort to approximations or to sophisticated mathematical techniques. I describe a program, called PLR (for Piecewise Linear Reasoner), that formalizes an analysis strategy employed by experts. PLR takes parameterized ordinary differential equations as input and produces a qualitative description of the solutions for all initial values. It approximates intractable nonlinear systems with piecewise linear ones, analyzes the approximations, and draws conclusions about the original systems. It chooses approximations that are accurate enough to reproduce the essential properties of their nonlinear prototypes, yet simple enough to be analyzed completely and efficiently. PLR uses the standard phase space representation. It builds a composite phase diagram for a piecewise linear system by pasting together the local phase diagrams of its linear regions. It employs a combination of geometric and algebraic reasoning to determine whether the trajectories in each linear region cross into adjoining regions and summarizes the results in a transition graph. Transition graphs explicitly express many qualitative properties of systems. PLR derives additional properties, such as boundedness or periodicity, by theoretical methods. PLR's analysis depends on abstract properties of systems rather than on specific numeric values. This makes its conclusions more robust and enables it to handle parameterized equations transparently. I demonstrate PLR on several common nonlinear systems and on published examples from mechanical engineering. || Armin Haken armin@ai.toronto.edu || || (416)978-6277 ...!utcsri!utai!armin || || UofT DCS, Toronto M5S 1A4 CDN armin%ai.toronto@csnet-relay || ------------------------------ Date: Wed, 10 Feb 88 15:29:57 CST From: pitt@p.cs.uiuc.edu (Lenny Pitt) Subject: Conference - Computational Learning Theory CALL FOR PAPERS Workshop on Computational Learning Theory Cambridge, Massachusetts August 3-5, 1988 The first workshop on Computational Learning Theory will be held at MIT August 3-5, 1988. It is expected that most papers will consist of rigorous and formal analyses of theoretical issues in Machine Learning. Empirical work will be considered only if it is testing some hypothesis that has a quantitative theoretical basis. Possible topics include, but are not limited to: 1. resource, convergence-rate and robustness analysis (time, space, number of examples, noise sensitivity, etc.) of specific learn- ing algorithms, 2. general learnability and non-learnability results in existing computational learning models and general upper and lower bounds on resources required for learning, and 3. new computational learning models, extensions of existing learn- ing models, and theoretical comparisons among learning models. Papers that make formal connections with work in Robotics, Neural Nets, Pattern Recognition, Adaptive Signal Processing and Cryptography are also welcome. TO REGISTER FOR THE WORKSHOP Due to space limitations, registration for the workshop will be limited to 60. If you would like to participate, send a brief (one page max.) description of your current research, by April 15, to the address below. Participants will be notified, and sent registration information, by June 1. It is possible that some financial support will be available for graduate student partici- pants. TO SUBMIT A PAPER Authors should submit extended abstracts that consist of: (1) A cover page with title, author's names and addresses (e-mail also if possible), and a 200 word summary. (2) A body not exceeding 5 pages in twelve-point font. A brief statement of the definitions and model used followed by a list of theorems with proof sketches is suggested. A succinct statement on the significance of the results should also be included. Authors should send 8 copies of their submissions to John Cherniavsky Workshop on Computational Learning Theory Department of Computer Science Georgetown University Washington, D.C. 20057 The deadline for receiving submissions is April 15, 1988. This deadline is FIRM. Authors will be notified by June 1, final camera-ready papers will be due July 1. Organizing/program committee: David Haussler, UC Santa Cruz, (workshop co-chair); Leonard Pitt, U. Illinois, (workshop co-chair); John Cherniavsky, Georgetown University, (program committee chair); Ronald Rivest, MIT, (local arrangements); Dana Angluin, Yale University; Carl Smith, NSF; Leslie Valiant, Harvard University; Manfred Warmuth, UC Santa Cruz. ------------------------------ Date: Tue, 9 Feb 88 10:32:34 PST From: Haym Hirsh Subject: Conference - 3rd European Working Session on Learning 1988 [forwarding on request of Sleeman] From: Derek Sleeman EWSL 88 3rd European Working Session on Learning Turing Institute, Glasgow, Scotland - October 3-5, 1988 CALL FOR PAPERS EWSL is an annual meeting on Machine Learning enabling European researchers to present recent research results. However, participation is NOT limited to Europeans. The first EWSL meeting was held at Orsay, France, in February 1986; the second was held in Bled, Yugoslavia. TOPICS The emphasis will be on Machine Learning, but relevant Cognitive Science studies pertinent to the theme will be most welcomed. SUBMISSION OF PAPERS Authors should submit five copies of papers (in English) of no more than 5000 words by the 1 May 1988 to the Program Chairman: Derek Sleeman (EWSL-88) Department of Computing Science University of Aberdeen ABERDEEN AB9 2UB Scotland UK Tel No. Aberdeen (+44 224) 272288; Telex 73458 The title page should contain the following information: Authors' names & addresses; and for the contact person telephone and telex numbers; together with an abstract of 100-200 words, and upto 10 descriptive key- words. TIMETABLE - submission deadline: 1 May 1988 - notification of acceptance or rejection: 1 July 1988 - camera ready copy: 15 August 1988 All delegates will receive a copy of the proceedings on registration; it is intended to publish a selection of the papers in a revised form after the meeting. PROGRAMME COMMITTEE MEMBERS Ivan BRATKO, Ljubljana, Yugoslavia Bernd NORDHAUSEN, Muenchen. Pavel BRAZDIL, Porto, Portugal Claus ROLLINGER, Stuttgart. Francoise FOGELMAN, Paris. Derek SLEEMAN, Aberdeen. Yves KODRATOFF, Orsay. Martin STACEY, Aberdeen. Donald MICHIE, Glasgow. Luc STEELS, Brussels. Steve MUGGLETON, Glasgow. Bob WIELINGA, Amsterdam. FORMAT OF THE CONFERENCE The conference will consist of invited talks, submitted technical papers, short project progress reports, and in-depth discussions on spe- cial topics. Suggestions for panel topics are invited and should be sent to the Program Chairman. LOCAL ORGANISATION Tim Niblett/Jim Richmond Turing Institute George House 36 North Hanover Street GLASGOW G1 2AD Phone: +44 41 552 6400. ------------------------------ Date: 9 Feb 88 22:08:35 GMT From: munnari!dbrmelb.oz.au!ron%dbrmelb.dbrhi.OZ@uunet.UU.NET (Ron Sharpe) Subject: Conference - Micros, Expert Systems in Planning, Transport, Building INTERNATIONAL WORKSHOP: MICROCOMPUTERS AND EXPERT SYSTEMS FOR PLANNING, TRANSPORT AND BUILDING When : Tues. April 26th 1988 (Services planning) Wed. April 27th 1988 (Building planning) Thur. April 28th 1988 (Expert systems) Where : AUSTRADE (Australian Trade Commission) 24th Floor, Harbour Centre, 25 Harbour Road Wan Chai HONG KONG Theme : Government and private enterprise currently face increasing responsibilities that must be met with diminishing resources in a period of rapid technological and social change. Microcomputer technology offers opportunities for improved capabilities for evaluating the services required, and in providing information for decision making. These seminars and workshops will describe the development of over 15 microcomputer-based packages expressly designed for planners, engineers and management. Expert systems applications under development will be presented. Eleven key planners and engineers from the People's Republic of China & Phillipines will be participating in the workshop (under sponsorship from AIDAB - Australian International Development Assistance Bureau). This will provide a valuable opportunity for other participants wishing to establish linkages in those countries. Cost : $HK300 per day, or $HK550 for 2 days, or $HK800 for 3 days (includes luncheon, refreshments and software technical handouts). Discount: 10% for early registration before March 7th. (Australian costs: $A60, $A110, $A160 respectively). As the number of places is limited, acceptance will be in order in receipt of registration. Further details & Enquires: Dr Ron Sharpe, Commonwealth Scientific & Industrial Research Organisation (CSIRO), Div. of Construction & Engineering, PO Box 56, Highett, Vic 3190, Australia. Phone (+61 3) 556 2211, Fax (+61 3) 553 2819, Telex AA33766 e-mail: ron@dbrmelb.dbrhi.oz Dr T Y Chen, Centre of Computer Studies & Applications, University of Hong Kong, Pokfulam Rd, Hong Kong. Phone 5-859 2491, Telex 71919 e-mail: tychen@hkucs.uucp Dr Anthony Gar-On Yeh, Centre of Urban Studies & Urban Planning, University of Hong Kong, Pokfulam Road, Hong Kong Phone 5-8592721-7, Telex 71919 e-mail: hkucs!hkucc!hdxugoy.uucp ------------------------------ End of AIList Digest ********************