IRList Digest Tuesday, 22 Apr 1986 Volume 2 : Issue 20 Today's Topics: Query - Retrieve similar pictures from DB of binary images? SIGIR - Forum schedule Announcement - 2nd Conf. on Computer Interfaces & Intermediaries Announcement - ACL Nat'l Conf. information file COGSCI - Robust NL Interface, Inexact Reasoning CSLI - Foundations of Representation, Understanding Computers & Cognition CSLI - Lexical Rules and Lexical Representation ---------------------------------------------------------------------- Date: Tue, 8 Apr 86 06:37:18 est Subject: Query on Picture Retrieval/Similarity From: cross%wsu@csnet-relay.ARPA We are starting to build a system that will retrieve trademarks from a DB which resemble a given trademark in order to assess possible trademark infringement. I would appreciate pointers to systems which retrieve pictures from a DB which are similar to a given picture. Picture means binary images in this context. Please reply to me and I will summarize to IRLIST. ---- 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 Phone: 509-335-6319 or 509-335-6636 ----------------------------- Date: Wed, 9 Apr 1986 14:30 CST From: Vijay V. Raghavan Subject: Next Forum Dear Ed, There will probably be some delay in my sending the material for the next forum. This is because I am hoping to get the preliminary program for the Pisa conference included. I hope to get this sometime next week from Rabbitti. With regards, Vijay. ----------------------------- Date: Sat, 19 Apr 86 19:23 EST From: "Richard (Dick) Marcus" Subject: Second Conference on Computer Interfaces and Intermediaries for Information Retrieval The conference agenda is now in near-final form. Sessions will run from early morning Thursday, 28 May, to Saturday PM, May 31, at Boston Park Plaza Hotel. Sessions scheduled: 1- AI in Retrieval 2- Gateways 3- Interfaces 4- Common Command Language 5- Natural Language The following speakers will present papers on those topics: Martha Williams (U Illinois), Carol Fenichel (Hahnemann U), Gio Wiederhold (Stanford), Nick Belkin (Rutgers), Linda Smith (U Illinois), Michael Pincus (Thunderstone Corp), Bruce Croft (U Mass), Ed Fox (VPI), Viktor Hampel (Lawrence Livermore Natl Lab), Gladys Cotter (DTIC), Judy Hushon (BBN), Richard Marcus (MIT), Charles Meadow (U Toronto), David Toliver (ISI), Bill Mischo (U Illinois), Michael Monahan (GAEC Computers Intl), Donald Hawkins (Bell Labs), Marcia Bates (UCLA), Hilary Burton (Lawrence Livermore Natl Lab), Dineh Moghdam Davis (Bentley C), Emily Fayen (U Penn), Rita Bergman (CCA), Lionel Bernstein (U Illinois-Chicago), Gabriel Jacobson (GTE Labs), Ralph Weischedel (BBN), Sharon Salveter (Boston U), Craig Thompson (Texas Instruments), and Gerard Salton (Cornell U). In addition, there will be system presentations and demonstrations by Artificial Intelligence Corp (INTELLECT), Dialog (DialogLink), Disclosure (Micro Disclosure), H.W. Wilson (WILSEARCH), ISI (Sci-Mate), KNM Inc. (SIRE), Telebase (systems (EASYNET), and MIT (CONIT). Early registration is $225 ($175 for government attendees, $20 extra for pay-at-door). For further info, including brochure with full agenda details and registration form plus info on special rates for hotel and transportation, call or write the American Defense Preparedness Association (ADPA), Rosslyn Center, Suite 900, 1700 N Moore St., Arlington, VA 22209, 703/522-1820 or Ms. Carol Jacobson at the Defense Technical Information Center (DTIC) 202/274-5367. In my objective opinion, it looks like a good, one-track conference! ----------------------------- Date: Wed, 16 Apr 86 00:49:50 est From: walker@mouton.ARPA Subject: ACL Annual Meeting, 10-13 June, Columbia University, New York City The Program and Registration Information brochure is just being mailed to all ACL members and to selected members of AAAI and LSA. If you are not sure you will be receiving it and would like a net copy, send a message to one of the addresses below; and I will (try to--even electronic mail is not always reliable) send you one. Please include the phrase "ACL net info" in the subject line. And include the full net address in the body of the message; the complexity of network connections coupled with the poverty of our mail system sometimes makes "replies" unsendable. The file has about 20,000 characters; it contains the full program (33 papers; an invited presentation by Gary Hendrix; two forums, one on Connectionism with Terry Sejnowski and Dave Waltz, the other on Machine Translation with Martin Kay and Maghi King); descriptions of the 6 tutorials (Intro to Computational Linguistics, Natural Language Generation, Structuring the Lexicon, Recent Developments in Syntactic Theory and Their Computational Import, Current Approaches to Natural Language Semantics, and Machine Translation--all held on 10 June); registration information and directions; and an Application Form that can be printed out, filled in (or filled in, printed out), and mailed in. Inexpensive air-conditioned dormitory accommodations are available, and some good rates for hotels have been secured. We are still encouraging people who would like to exhibit or demonstrate programs to contact Ralph Grishman (Computer Science, New York University, 251 Mercer Street, New York, NY 10012; 212:460-7492; grishman@nyu.arpa). Don Walker walker@mouton.arpa walker%mouton@csnet-relay {ucbvax, ihnp4, ...}!bellcore!walker address mail to: Donald E. Walker (ACL) Bell Communications Research 445 South Street, MRE 2A379 Morristown, NJ 07960, USA 201:829-4312 ACL Annual Meeting, 10-13 June 1986, Columbia University, New York City ----------------------------- Date: Fri, 11 Apr 86 19:02:36 est From: DEJONG%OZ.AI.MIT.EDU@mc.lcs.mit.edu Subject: Cognitive Science Calendar [Extract - Ed] Tuesday, 15 April 10:30am Room: 2nd floor large conference room BBN Labs, 10 Moulton Street, Cambridge BBN SCIENCE DEVELOPMENT PROGRAM - AI/EDUCATION SEMINAR "Pragmatic Modeling: Toward a Robust Natural Language Interface" Professor Sandra Carberry University of Delaware Naturally occurring dialogue is both imperfect and incomplete. Not only does the information-seeker fail to communicate all aspects of his underlying task and partially constructed plan for accomplishing it, but also his utterances are often imperfectly or imcompletely formulated. It appears that human information-seekers expect an information-provider to facilitate a productive exchange by assimilating the dialogue and using this knowledge to remedy many of the information-seeker's faulty utterances. This talk will describe an on-going research effort aimed both at developing techniques for inferring and constructing a user model from an information-seeking dialogue and at identifying strategies for using this model to develop more robust natural language interfaces. Emphasis will be on the dynamic construction of the task-related plan motivating the information-seeker's queries, and its application in handling pragmatically ill-formed and incomplete utterances. Wednesday, 16, April 4:00pm (refreshments 3:45) Room: NE43-512A LCS THEORY OF COMPUTATION SEMINAR "Inexact Reasoning Using Graphs" JUDEA PEARL Computer Science Department UCLA Probability theory is shunned by most researchers in Artifical Intelligence. New calculi, claimed to better represent human reasoning under uncertainty, are being invented and reinvented at an ever-increasing rate. A major reason for the emergence of this curious episode has been the objective of making reasoning systems TRANSPARENT i.e., capable of producing PSYCHOLOGICALLY MEANINGFUL explanations for the intermediate steps used in deriving the conclusions. While traditional probability theory, admittedly, has erected cultural barriers against meeting this requirement, we shall show that these barriers are superficial, and can be eliminated with the use of DEPENDENCY GRAPHS. The nodes in these graphs represent propositions (or variables), and the arcs represent causal dependencies among conceptually-related propositions. We further argue that the basic steps invoked while people query and update their knowledge correspond to mental tracings of preestablished links in such graphs, and it is the degree to which an explanation mirrors these tracings that determines whether it is considered "psychologically meaningful". The first part of the talk will examine what properties of probabilistic models can be captured by graphical representations, and will compare the properties of two such representations: Markov Networks and Bayes Networks. The second part will introduce a calculus for performing inferences in Bayes Networks. The impact of each new evidence is viewed as a perturbation that propagates through the network via local communication among neighboring concepts. We show that such autonomous propagation mechanism leads to flexible control strategies and sound explanations, that it supports both predictive and diagnostic inferences, that it is guaranteed to converge in time proportional to the network's diameter, and that every proposition is eventually accorded a measure of belief consistent with the axioms of probability theory. In conclusion, we will show that the current trend of abandoning probability theory is grossly premature--taking graph propagation as the basis for probabilistic reasoning satisfies most computational requirements for managing uncertainties in reasoning systems and, simultaneously, it exhibits epistemological features unavailable in any competing formalism. HOST: Ronald Rivest ----------------------------- Date: Thu, 10 Apr 86 18:55:11 est From: EMMA@su-csli.ARPA Subject: Calendar, April 10, No. 11 [Extract - Ed] CSLI ACTIVITIES FOR THURSDAY, April 10, 1986 Representation: Foundations of Representation Ken Olson (Olson.pa@xerox) What is it for a thing to represent another? Answers that rely in any simple way on resemblance and causality are easily dismissed. Peirce thought that representation was an irreducibly three-place relation between a sign, an object, and what he called an interpretant. But while Peirce's view has much to recommend it, the notion of an interpretant seems to introduce an unwelcome mentalistic element. At least it is unwelcome if we wish to account for mental representation as one species of the more general notion instead of giving it privileged status. I claim, however, that the notion of interpretant does not presuppose a full-fledged mind. Other ideas of Peirce's also deserve attention. Situation theory may finally be the proper medium in which to realize his goal of a general theory of signs. CSLI ACTIVITIES FOR THURSDAY, April 17, 1986 Understanding Computers and Cognition by Terry Winograd Discussion led by Brian Smith (Briansmith.pa@xerox) For some time, Terry Winograd has believed that the general semantical and theoretical approaches embodied in current AI systems are inadequate for dealing with human language and thought. What distinguishes his views from those of various other AI critics is the scope of what he takes to be the problem. In particular, as he argues in his new book, he is convinced that that nothing within what he calls the ``rationalistic tradition''---in which he would presumably include most CSLI research---will overcome these inherent limitations. In this TINLunch we will discuss the argument presented in the book, try to separate the various threads that lead to Terry's conclusion, and assess its relevance to the CSLI research program. (The book, which is not difficult to read, should be available at local bookstores; some selected portions will be made available in the usual places.) ----------------------------- Date: Thu, 17 Apr 86 00:59:40 est From: EMMA@su-csli.ARPA Subject: Calendar, April 17, No. 12 [Extract - Ed] CSLI SEMINAR Lexical Rules and Lexical Representations Mark Gawron, Paul Kiparsky, Annie Zaenen 4:15 p.m., April 24, CSLI Trailer Classroom This is the third of a series of talks reflecting the ongoing elaboration of a model of lexical representation. In the first, Mark Gawron discussed a frame-based lexical semantics and its relationship to a theory of lexical rules. In the second, Paul Kiparsky proposed a theory of the linking of thematic roles to their syntactic realizations, emphasizing its interactions with a theory of morphology; and in this one, a sub-workgroup of the lexical project will sketch a unification based representation for the interaction of the different components of the lexical representation and both syntax and sentence semantics.