nl-kr-request@CS.ROCHESTER.EDU (NL-KR Moderator Brad Miller) (06/16/88)
NL-KR Digest (6/15/88 18:14:36) Volume 4 Number 59 Today's Topics: Feasible Learnability and Locality of Grammars (UNISYS seminar) Talk by Michael Lesk BBN AI Seminar -- Phylis Koton Lang. & Cognition Seminar MT Conference Budapest Submissions: NL-KR@CS.ROCHESTER.EDU Requests, policy: NL-KR-REQUEST@CS.ROCHESTER.EDU ---------------------------------------------------------------------- Date: Wed, 8 Jun 88 15:19 EDT From: finin@PRC.Unisys.COM Subject: Feasible Learnability and Locality of Grammars (UNISYS seminar) AI SEMINAR UNISYS PAOLI RESEARCH CENTER Feasible Learnability and Locality of Grammars Naoki Abe Computer and Information Science University of Pennsylvania Polynomial learnability is a generalization of a complexity theoretic notion of feasible learnability originally developed by Valiant in the context of learning boolean concepts from examples. In this talk I will present an intuitive exposition of this learning paradigm, and then apply this notion to the evaluation of grammatical formalisms for linguistic description from the point of view of feasible learnability. In particular, a novel, nontrivial constraint on the degree of ``locality'' of grammars will be defined which allows grammatical formalisms of much linguistic interest to be polynomially learnable. If time allows possible implications of this result to the theory of natural language acquisition will also be discussed. 2:00 pm Wednesday, June 1 Paoli Auditorium Unisys Paloi Research Center Route 252 and Central Ave. Paoli PA 19311 -- non-Unisys visitors who are interested in attending should -- -- send email to finin@prc.unisys.com or call 215-648-7446 -- ------------------------------ Date: Tue, 7 Jun 88 17:04 EDT From: Peter de Jong <dejong@wheaties.ai.mit.edu>, Laureen Fletcher <fletch%eyes@media-lab.media.mit.edu> Subject: Talk by Michael Lesk "Does Technology Affect How People Read?" Lessons from the 18th Century. This is about reprinting the first edition of "Tristram Shandy;" duplicating 18th century fonts, etc. with some discussion of the switch from reading aloud to reading silently. "How to Tell a Pine Cone from an Ice Cream Cone -- Sense Disambiguation Using Machine Readable Dictionaries" Does a "fireman" feed fires or put them out? It depends on whether or not he is on a steam locomotive. This talk explains a scheme for deciding which sense of an ambiguous word is meant by counting overlaps of words in definitions in a machine-readable dictionary. Michael Lesk Division Manager of Computer Sciences Research Bell Communications Research Morristown, New Jersey Friday, June 10, 1988 2:00 - 3:00 p.m. E15-401 Host: Peg Schafer ------------------------------ Date: Fri, 10 Jun 88 13:52 EDT From: Marc Vilain <MVILAIN@G.BBN.COM> Subject: BBN AI Seminar -- Phylis Koton BBN Science Development Program AI Seminar Series Lecture MODEL-BASED DIAGNOSTIC REASONING USING PAST EXPERIENCES Phylis Koton MIT Lab for Computer Science (ELAN@XX.LCS.MIT.EDU) BBN Labs 10 Moulton Street 2nd floor large conference room 10:30 am, Tuesday June 14 The problem-solving performance of most people improves with experience. The performance of most expert systems does not. People solve unfamiliar problems slowly, but recognize and quickly solve problems that are similar to those they have solved before. People also remember problems that they have solved, thereby improving their performance on similar problems in the future. This talk will describe a system, CASEY, that uses case-based reasoning to recall and remember problems it has seen before, and uses a causal model of its domain to justify re-using previous solutions and to solve unfamiliar problems. CASEY overcomes some of the major weaknesses of case-based reasoning through its use of a causal model of the domain. First, the model identifies the important features for matching, and this is done individually for each case. Second, CASEY can prove that a retrieved solution is applicable to the new case by analyzing its differences from the new case in the context of the model. CASEY overcomes the speed limitation of model-based reasoning by remembering a previous similar case and making small changes to its solution. It overcomes the inability of associational reasoning to deal with unanticipated problems by recognizing when it has not seen a similar problem before, and using model-based reasoning in those circumstances. The techniques developed for CASEY were implemented in the domain of medical diagnosis, and resulted in solutions identical to those derived by a model-based expert system for the same domain, but with an increase of several orders of magnitude in efficiency. Furthermore, the methods used by the system are domain-independent and should be applicable in other domains with models of a similar form. ------------------------------ Date: Tue, 14 Jun 88 08:51 EDT From: Dori Wells <DWELLS@G.BBN.COM> Subject: Lang. & Cognition Seminar BBN Science Development Program Language & Cognition Seminar Series CHILDREN'S REORGANIZATION OF KNOWLEDGE IN THE DOMAIN OF ASTRONOMY Stella Vosniaoov University of Illinois BBN Laboratories Inc. 10 Moulton Street Large Conference Room, 2nd Floor 10:30 a.m., Wednesday, June 15, 1988 Abstract: Some preliminary findings from an ongoing project on children's acquisition of knowledge in the domain of astronomy will be presented. The findings indicate that elementary school children's early beliefs are consistent with their phenomenal explanation of a stationary flat earth and an up and down movement of the sun and moon. These beliefs appear to be quite resistant to change and rise to a number of misconceptions which reveal children's difficulty to assimilate current scientific views. ------------------------------ Date: Fri, 10 Jun 88 08:28 EDT From: New Directions in Machine Translation - Budapest <mcvax!dlt1!mtconf@uunet.UU.NET> Subject: MT Conference Budapest As was announced earlier, BSO/Research (Utrecht) and the John von Neumann Society for Computing Sciences (Budapest) jointly organise the international conference NEW DIRECTIONS IN MACHINE TRANSLATION on 18 and 19 August 1988 (Thursday and Friday before Coling) in Budapest. We can now announce the final list of lectures. The conference can still accept some participants (but not papers, there are invited speakers only). If you want to register, please contact the John von Neumann Society for Computing Sciences Conference Secretariat Pf. 240 H-1368 Budapest 5 Hungary Telephone: .. 36 / 1 / 329390 Telex: 225792 mtesz h The registration fee is 170,- DM (programme, proceedings, welcome party, refreshments and lunches). Hotel accomodation can be ordered via the Conference Secretariat. Please do not try to register via this electronic address which is in the Netherlands, not in Hungary. The list of speakers is: W. JOHN HUTCHINS (Norwich): Recent developments in machine translation - a review of the last five years. TIBOR V'AMOS (Budapest): Language and computer society. IVAN I. OUBINE (Moscow): The state of the art in machine translation in the USSR. CHEN YUAN (Peking): Esperanto-based MT research in China. CHRISTIAN BOITET (Grenoble): Pros and cons of the pivot and transfer approaches in multilingual machine translation. MICHIKO KOSAKA (New York): A sublanguage approach to Japanese-English machine translation. IV'AN GUZM'AN DE ROJAS (La Paz): ATAMIRI - interlingual MT using the Aymara language. KLAUS SCHUBERT (Utrecht): The architecture of DLT - interlingual or double direct? CHRISTA HAUENSCHILD (Berlin): Discourse structure - some implications for machine translation. JUN-ICHI TSUJII (Kyoto): What is a cross-linguistically valid interpretation of discourse? CHRISTIAN GALINSKI (Vienna): Advanced terminology banks supporting knowledge-based MT Some reflections on the costs for setting up and operating a terminological data bank. WERA BLANKE (Berlin): Terminologia Esperanto-Centro Efforts for terminological standardization in the planned language. DIETRICH WEIDMANN (Schaffhausen): Universal applicability of dependency grammar. BENGT SIGURD (Lund): Translating to and from Swedish by SWETRA - multilanguage translation system. G'ABOR PR'OSZ'EKY (Budapest): Hungarian - a special challenge to machine translation? CLAUDE PIRON (Geneva): What we can learn about languages from mistakes made by professional translators. ------------------------------ End of NL-KR Digest *******************