nl-kr-request@cs.rpi.edu (NL-KR Moderator Chris Welty) (07/10/90)
NL-KR Digest (Mon Jul 9 10:08:58 1990) Volume 7 No. 12 Today's Topics: UK redistribution of NL-KR Digest. Suggestions wanted for Computational Linguistics Course an announcement - NATURAL LANGUAGE SOFTWARE REGISTRY Educational Applications using On-Line Lexicons looking for a paper: DB & NL Pointers to Nlang systems summaries Re: Info on Text Compression sought. ADAPTATION ANOMALY IN DEVELOPING NERVOUS SYSTEMS (BBN Seminar) Re: a machine more intelligent than human beings? CFP-1990 Workshop on Current Trends in SNePS Submissions: nl-kr@cs.rpi.edu Requests, policy: nl-kr-request@cs.rpi.edu Back issues are available from host archive.cs.rpi.edu [128.213.5.17] in the files nl-kr/Vxx/Nyy (ie nl-kr/V01/N01 for V1#1), mail requests will not be promptly satisfied. If you can't reach `cs.rpi.edu' you may want to use `turing.cs.rpi.edu' instead. BITNET subscribers: we now have a LISTSERVer for nl-kr. You may send submissions to NL-KR@RPIECS and any listserv-style administrative requests to LISTSERV@RPIECS. ----------------------------------------------------------------- To: nl-kr@cs.rpi.edu Date: Wed, 27 Jun 90 16:50:14 BST >From: "G. Joly" (UCL-SM Psychiatry) <gjoly@NSFnet-Relay.AC.UK> Subject: UK redistribution of NL-KR Digest. Please note that UK subscribers (individuals or lists) are encouraged to receive their copies from nl-kr-list-request@uk.ac.nsfnet-relay. You should SIGNOFF NL-KR at LISTSERV@RPIECS.BITNET for yourself once you are on the list at the NSF. Gordon Joly aka nl-kr-list-request@nsf.ac.uk ------------------------------ To: nl-kr@cs.rpi.edu >From: ok@goanna.cs.rmit.OZ.AU (Richard A. O'Keefe) Newsgroups: comp.ai.nlang-know-rep Subject: Suggestions wanted for Computational Linguistics Course Keywords: education help Date: 27 Jun 90 12:44:23 GMT Currently RMIT Comp. Sci. offers 3-year BScs (plus MSc, PhD, various diplomas). Next year we plan to add an optional 4th "honours" year, with the idea that people who intend to advance to MSc would take that year first. I'll probably be teaching that course. I'll have no trouble at all filling the time, but I want the course to lay a solid foundation for future study. So I'm looking for suggestions, especially from people who have successfully taught a course at this level. As the students will not have had any previous exposure to linguistics, my current idea is that the first semester should be devoted to linguistic topics. Does that sound about right? Which are the most important topics to teach at this level _to_people_who_are_doing_computational_linguistics rather_than_linguistics_proper? Just how much of the _computational_ side is appropriate in a first course? What do you suggest as essential texts that students should have? What do they need access to that they needn't buy unless they go further? Which journals _must_ the library subscribe to if we want to students to do Computational Linguistics work at MSc level? (AJCL of course, what else?) And so on. Please send E-mail to ok@goanna.cs.rmit.OZ.AU and if there's interest I'll summarise. - - "private morality" is an oxymoron, like "peaceful war". ------------------------------ To: nl-kr@cs.rpi.edu Date: Wed, 27 Jun 90 13:07:49 CDT >From: Computational Linguists <registry%tira@gargoyle.uchicago.edu> Subject: an announcement - NATURAL LANGUAGE SOFTWARE REGISTRY NATURAL LANGUAGE SOFTWARE REGISTRY There are now many natural language software projects, both commercial and noncommercial. In order to facilitate the exchange and evaluation of software, the University of Chicago's Center for Information and Language Studies is undertaking to catalog projects for the community. If you have developed a piece of software for natural language processing that other researchers might find useful, you can help us by describing it below. An electronic version of the form is preferred; it is available by anonymous ftp to tira.uchicago.edu. Feel free to write up software that is useful to you, even if you are not the originator. Although the registry will be an ongoing project, we ask that you please return the form by July 30 for existing software. Results will be made available in this forum. Elizabeth Hinkelman, for CILS (registry@tira.uchicago.edu) Name: Person to contact for software (if different): Institution: Department: Street: City/State/Zip: Country: Phone (with country & area codes): Email network & address: Name of system: Type of system: commercial product research system other (specify) Application: machine translation database interface parsing generation understanding other (specify) Components: phonological analyzer/generator morphological analyzer/generator parser/generator semantic interpreter knowledge representation discourse structure pragmatic features other (specify) Are the components available as independent modules, or firmly embedded? Can they be extended easily, under certain conditions, or with difficulty? Size of system: Programming language: Operating system or hardware: Applicable natural language(s): Can other languages be substituted easily? Number of examples the system has been tested on: [Unit: sentences, messages, other (specify)] 1-10 - demonstration system 10-100 - small research system 100-1000 - larger research system 1000-10000 - robust or production quality system Is the project under development, completed, or ongoing? Summarize the main goals and ideas. Indicate what makes the project a useful and interesting tool for research applications. List documents in which the software is described: User documentation: System documentation: Available support: none upgrades source code consulting Format for software distribution: Cost: Restrictions on use: If you are willing to have the software reviewed, please send us a version along with this information. We are also interested in reports and documentation, even for software not reviewed. NL Software Registry Center for Information and Language Studies 1100 East 57th Street Chicago, IL 60637, USA registry@tira.uchicago.edu ------------------------------ To: nl-kr@cs.rpi.edu Date: Wed, 27 Jun 90 16:02:28 EDT >From: kegl@cs.swarthmore.edu (Judy Kegl) Subject: Educational Applications using On-Line Lexicons I am currently revising a survey article on the use of on line dictionaries in educational applications. In particular, preexisting comprehensive on-line dictionaries or on line printing tapes converted inoto a data base format that interface woth any and all educational applications are what I am looking for. Please send information and/or bibliographic references to Judy Kegl Depts. of Linguistics and Psychology Swarthmore College Swarthmore, PA 19081 kegl@campus.swarthmore.edu I would be happy to sent a copy of the draft as it stands to anyone intersted in commenting on it or sugesting additional information. Hard mail or electronic copies are availiable. Thanks. ------------------------------ To: nl-kr@cs.rpi.edu >From: unige@cernvax.cern.ch (nicolas mayencourt) Newsgroups: comp.databases,comp.ai.nlang-know-rep,comp.ai Subject: looking for a paper: DB & NL Keywords: data bases, natural language Date: 2 Jul 90 11:00:08 GMT Followup-To: comp.databases Hello, I am posting this message for a friend. She is working on data base interfaces and she has been looking for a reference for a long time. Here it is: W. Hoppner, K. Morik, H. Marbugger Taking it over: the natural language dialog systems HAMANS. Report ANS 26, May 1984. Does anybody know anything about it? Or how to get in contact with the authors? By the way, this paper should also be available in the following book: L. Bolc. (edt) Cooperative Interactive Systems Berlin: Springer, 1984. but any attempt to find it proved unsuccessful. You can either e-mail me or mail her. Her address is: Aicha Abdeljaoued Centre Universitaire d'Informatique Universite` de Geneve 12, rue du Lac CH - 1207 Geneve, Switzerland Thank you, -Ruggero Milanese (milanese@cuisun.unige.ch) ------------------------------ To: nl-kr@cs.rpi.edu >From: ccicpg!cci632!ritcsh!ultb.rit.edu!rtgrc@uunet.UU.NET (R.T. Gayvert) Newsgroups: comp.ai.nlang-know-rep Subject: Pointers to Nlang systems summaries Keywords: Natural Language summaries front-ends Date: 29 Jun 90 17:43:04 GMT I've been asked to put together a talk on the state-of-the-art in natural-language applications. In other words what kinds of commerical products are available for using natural language input to various computational systems. Does anyone have any good references to articles which sumarize current systems and perhaps make some predictions on the future of the field. Thanks in advance for any information. Please MAIL replys to (no matter what the headers say) Tom Ridley trr1442@ma.RIT.EDU ------------------------------ To: nl-kr@cs.rpi.edu Date: Wed, 4 Jul 90 17:18:24 EST >From: drh@ultima.cs.uts.oz.au (David (Ray) Haynes) Subject: Re: Info on Text Compression sought. Newsgroups: comp.ai.nlang-know-rep In comp.ai.nlang-know-rep you write: >I would be grateful for pointers to literature in this area. 1) Could you provide more information, please? The project I was working on last year looked at state of the art compression for both text and graphics. - are you interested in text only? - have you considered graphics compression for text? - have you considered storing images as well as text? - why 16 byte records? - other questions, I can't think of just now; perhaps you could provide more technical information, because 2) The projects I am currently working on involve natural language interfaces to database (in general, and office information systems in particular). We are using Nexpert, CLOS and Oracle on Sparc workstations, and hope to involve both graphic and text display of office objects. Could you reciprocate by providing an introductory reading list on natural language interfaces to databases (as I am only new to this area)? Thanks, David Raymond Haynes Senior Research Assistant School Of Computing University of Technology, Sydney phone (02) 281 2552 e-mail drh@ultima.cs.uts.oz.au ------------------------------ To: nl-kr@cs.rpi.edu >From: "Damaris M. Ayuso" <dayuso@BBN.COM> Reply-To: dayuso@BBN.COM Date: Mon, 2 Jul 90 13:20:42 EDT Subject: ADAPTATION ANOMALY IN DEVELOPING NERVOUS SYSTEMS (BBN Seminar) BBN STC Science Development Program AI Seminar Series Lecture ADAPTATION ANOMALY IN DEVELOPING NERVOUS SYSTEMS: A COMPUTATION MODEL (Genetic Algorithm) YUVAL DAVIDOR Department for Applied Mathematics Weizmann Institute yuval%wisdom.weizmann.ac.il@cunyvm.cuny.edu BBN STC, 2nd floor large conference room 10 Moulton St, Cambridge MA, 02138 Wednesday July 11th, 1990, 10:30 AM A development-theoretic model of the nervous system is postulated through the use of a genetic algorithm. This model suggests an explanation, possibly universal, to the dynamics of natural selection as a result of interaction between environmental conditions and intrinsic features of a developing (nervous) system. The algorithm is a generalized procedure simulating evolution in which natural selection operates concurrently on all levels. The model predicts that systems, which initiate development in an environment of a finite degree of uncertainty, have a long term fitness increase if the system starts from a state of moderate redundancy. Due to selection, the redundancy is reduced thus allowing an efficient subsequent specialization of the more adapted elements of the system. The initial simplification of the system, which is reversed at a later stage, is called here adaptation anomaly, and results from selection sifting out the less adapted elements of the system. Development of the brain exhibits such behaviour and fits the theoretic model. Development and maturation of the nervous system involve cell migration and aggregation, elaboration of processes, selection of connections, and adaptation of particular model of transmission. Studies in neurobiology have, for many years, been struggling conceptually with the idea that the nervous system develops in an apparently wasteful fashion. More elements than are actually needed are involved and/or being formed at each stage of development. Therefore, initial stages of development are followed by cell death, elimination of extra processes, and withdrawal of extra connections which consequently allow further growth and specialization. Genetic algorithms, being computer simulations of simplified interactions of population genetics with the environment, offer a unique tool to analyze complex aspects and behaviour of natural systems. The most intriguing aspect of genetic algorithms is that predictions can be derived even in cases where little is known on transient phases or explicit relations between elements of the environment and intrinsic features of a system. ******************************************************* Suggestions for AI Seminar speakers are always welcome. Please e-mail suggestions to Damaris Ayuso (dayuso@bbn.com) or Marie Meteer (mmeteer@bbn.com). ******************************************************* ------------------------------ To: nl-kr@cs.rpi.edu >From: Gordon Joly <gj@psychiatry.sm.ucl.ac.uk> Date: Tue, 26 Jun 90 19:24:31 BST Subject: Re: a machine more intelligent than human beings? [ I include this with some apprehension, as I don't want this digest to start looking like comp.ai. This has been the only reply so far and thus seems harmless, direct further replies to the original poster (deng@shire.cs.psu.edu). - CW ] > While they are still working on the project, scientists, > laywers, human and machine rights groups over the world are > enganged in heated discussion as whether MisChess can be > considered as intelligent as we are. > > So, what is your opinion? > I don't play chess, but I consider my self "intelligent"; this whole thing smacks of the Turing Test and even more of Searle's Chinese Room, a red herring in itself. The subject field "a machine more intelligent than human beings?" is a little off colour. Why "more" intelligent? Does we have a good idea of a scale (an IQ test) for a machine? Or some other metric? When something converses, hold data and organises that data in way that could be called "learning (from mistakes)", then I guess that could be called "intelligent". Perhaps such a system should do a Ph.D., just to show off a little. Gordon Joly. JANET:gj@uk.ac.ucl.sm.psych ARPA:gjoly@nsf.ac.uk UUCP:gj%uk.ac.ucl.sm.psychiatry@UUNET.UU.NET Dept. of Psychiatry, University College & Middlesex School of Medicine, Middlesex Hospital, Mortimer Street, LONDON W1N 8AA. Tel +44 71 380 9472 FAX 323 1459 ------------------------------ To: nl-kr@cs.rpi.edu Date: Tue, 26 Jun 90 15:14:24 EDT >From: kumard@cs.Buffalo.EDU (Deepak Kumar) Subject: CFP-1990 Workshop on Current Trends in SNePS CALL FOR PAPERS CURRENT TRENDS IN SNePS --- THE 1990 WORKSHOP October 5 & 6, 1990 State University of New York at Buffalo Buffalo, NY. DESCRIPTION SNePS is a state-of-the-art knowledge representation and reasoning system used for Artificial Intelligence and Cognitive Science research. It is a propositional semantic network based system designed by members of The SNePS Research Group in conjunction and under the supervision of Dr. Stuart Shapiro and Dr. William Rapaport. SNePS 2.1, an implementation of SNePS in CommonLisp runs on several computers and is distributed under license from The Research Foundation of State University of New York. SNePS, in its various incarnations, is actively being used and developed at various AI research labs around the world. The aim of this workshop is to assemble researchers who are (have been, or are considering) using SNePS as a research tool for AI modeling, those who are (have been, or are considering) evaluating SNePS as an AI research environment, and those who are interested in commenting on or discussing SNePS and/or the philosophy of knowledge representation it embodies. Attendance at the workshop will be kept small (by invitation only) to allow for maximum possible interaction among participants. The workshop will be held on the campus of SUNY at Buffalo. All papers presented will be edited and compiled into published Proceedings. A tutorial session on SNePS will be given on the first morning, and we will schedule a time slot for live demonstrations of SNePS-based or SNePS-related systems. To be invited to the workshop, you must submit an Invitation Request consisting of at least one of the following (preferably by e-mail): 1. Submit a one page abstract of a paper to be considered for presentation. 2. Submit a short write-up of your current research and how it does/might relate to SNePS and/or its philosophy of knowledge representation. 3. Indicate your intention to attend the tutorial. 4. Submit a short description outlining the nature of the demonstration you wish to give, its hardware/software/time requirements, and how the system to be demonstrates relates to SNePS. DEADLINES Invitation Requests due on Wenesday, August 15, 1990. Invitations issued by September 3, 1990. SUBMISSIONS TO: Syed Ali 226 Bell Hall SUNY at Buffalo Buffalo NY, 14260 syali@cs.buffalo.edu Send any queries to the program committee: Sy Ali (syali@cs.buffalo.edu) Hans Chalupsky (hans@cs.buffalo.edu) Deepak Kumar (kumard@cs.buffalo.edu) William Rapaport (rapaport@cs.buffalo.edu) Stuart C. Shapiro (shapiro@cs.buffalo.edu) Everyone will be expected to make their own living arrangements. We will send a list of nearby hotels with the invitations. There may be a possibility of making some arrangements with local SNeRG members. ------------------------------ End of NL-KR Digest *******************