[comp.ai.nlang-know-rep] NL-KR Digest Volume 4 No. 31

nl-kr-request@CS.ROCHESTER.EDU (NL-KR Moderator Brad Miller) (03/25/88)

NL-KR Digest             (3/24/88 15:35:30)            Volume 4 Number 31

Today's Topics:
        SUNY Buffalo Comp. Sci. Colloq:  David McDonald
        SUNY BUffalo Cognitive Science:  Ann Banfield
        SUNY Buffalo Cog/Ling Sci:  Ivan Sag
        Unisys AI Seminar, Remo Pareschi, 4/22
        seminar - The nature of children's early wh-questions
        Language & Cognition seminar

        Call for Papers - IJCAI 89 in Detroit

        New Chair in CL
        
Submissions: NL-KR@CS.ROCHESTER.EDU 
Requests, policy: NL-KR-REQUEST@CS.ROCHESTER.EDU
----------------------------------------------------------------------

Date: Thu, 10 Mar 88 10:12 EST
From: William J. Rapaport <sunybcs!rapaport@AMES.ARC.NASA.GOV>
Subject: SUNY Buffalo Comp. Sci. Colloq:  David McDonald


                   UNIVERSITY AT BUFFALO
                STATE UNIVERSITY OF NEW YORK

               DEPARTMENT OF COMPUTER SCIENCE

                         COLLOQUIUM

                   From Water to Wine:
Generating Natural Language Text from Today's Applications Programs

                   Dr. David D. McDonald
                Brattle Research Corporation

     Today's AI programs all too often cut  corners  in  the
conceptual  models  they  reason  with.  To generate natural
sounding texts from these models one needs to compensate for
these semantic deficits but without compromising the princi-
pled grammatical treatments in the generator.  I  will  talk
about  how  the  interface to our text generator, Mumble-86,
handles these problems, using an example from  a  knowledge-
based mission-planning system. Looking to the future, I will
discuss what we have learned about how to  correct  some  of
the  deficits, including locality of type information, anno-
tating units according to their use and context, and a  more
versatility  in  the representational formalism to allow for
alternative perspectives on the same facts.

             Date:   Thursday, 10th March, 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.

------------------------------

Date: Fri, 11 Mar 88 16:53 EST
From: William J. Rapaport <rapaport@cs.Buffalo.EDU>
Subject: SUNY BUffalo Cognitive Science:  Ann Banfield


                STATE UNIVERSITY OF NEW YORK AT BUFFALO

                  GRADUATE GROUP IN COGNITIVE SCIENCE

                                PRESENTS

                              ANN BANFIELD

                         Department of English
                  University of California at Berkeley

              THE LINGUISTICS OF SUBJECTIVITY IN NARRATIVE

Dr. Banfield will talk about her most recent work concerning the  issues
of non-narrated text, "point of view", and the philosophy of language as
it relates to subjectivity in narrative.

Using such disparate sources as Chomskyan linguistics, Russell's  theory
of  egocentric particulars, and Auerbach's notion of mimesis, and bring-
ing them to bear on the writings of Virginia  Woolf,  Gustave  Flaubert,
and other English and French novelists, Dr. Banfield produces strong and
interesting assertions about the nature of narrative.  Two of  her  most
controversial assertions are that the communication model of language is
inappropriate to a theory of subjectivity in narrative and that not  all
narratives  have  narrators.   Building  on  the tradition of generative
grammar rather than on structuralist principles of linguistics, Banfield
brings  a fresh perspective to current debates on the status of linguis-
tics for narrative theory.

Dr. Banfield's argument for a falsifiable  narrative  theory,  presented
most  fully  in her 1982 book _Unspeakable Sentences_ (Routledge & Kegan
Paul), has provoked considerable interest and controversy in the  fields
of literary linguistics, narrative theory, and the poetics of style.  In
addition, her theories have the potential to stimulate new discussion in
such  related  fields as linguistic pragmatics, artificial intelligence,
and the philosophy of the subject.

                        Thursday, March 24, 1988
                               4:00 P.M.
                        280 Park, Amherst Campus

There will be an evening discussion with Dr. Banfield  at  the  home  of
Erwin  Segal,  101  Carriage  Circle, Amherst.  For further information,
call Bill Rapaport (Dept. of Computer Science, 636-3193 or 3180) or Gail
Bruder (Dept of Psychology, 636-3676).

------------------------------

Date: Wed, 16 Mar 88 18:17 EST
From: William J. Rapaport <rapaport@cs.Buffalo.EDU>
Subject: SUNY Buffalo Cog/Ling Sci:  Ivan Sag


                STATE UNIVERSITY OF NEW YORK AT BUFFALO

                     The Steering Committee of the
              GRADUATE STUDIES AND RESEARCH INITIATIVE IN

                   COGNITIVE AND LINGUISTIC SCIENCES

                                PRESENTS

                              IVAN A. SAG

                       Department of Linguistics
             Stanford University and University of Chicago

                LINGUISTIC THEORY AND COGNITIVE SCIENCE

In this talk, I outline one view of how the concerns of  Cognitive  Sci-
ence  impose  constraints  on  the  design  of  linguistic  theories.  I
emphasize the importance of  such  design  properties  as  monotonicity,
simultaneous constraint satisfaction, declarativeness, and reversibility
of grammars.  Despite many appearances to the contrary, much  of  modern
linguistic  theory  can  be formulated in such terms.  Pollard and Sag's
"Information-Based Syntax and Semantics" (1987, to appear), is  in  fact
an  attempt  to  weave results from a number of diverse traditions (LFG,
GPSG, GB, Categorial Grammar, and Unification-Based Grammar  Formalisms)
into a sound theoretical framework that has just such design properties.
I will survey several results of this research program and offer sugges-
tions about directions for future research that integrates comprehensive
linguistic descriptions so designed with models of language processing.

                         Monday, March 21, 1988
                               4:15 P.M.
                        280 Park, Amherst Campus

There will also be an  informal  evening  discussion  at  Judy  Duchan's
house,  130  Jewett Parkway, at a time to be announced.  Call Bill Rapa-
port (Dept. of Computer Science, 636-3193 or 3180) for further  informa-
tion.

------------------------------

Date: Fri, 18 Mar 88 16:30 EST
From: finin@PRC.Unisys.COM
Subject: Unisys AI Seminar, Remo Pareschi, 4/22


			      AI SEMINAR
		     UNISYS PAOLI RESEARCH CENTER
				   
				   
	   A Definite Clause Version of Categorial  Grammar
				   
			    Remo Pareschi
		      University of Pennsylvania


Categorial Grammar is a framework for natural language analysis where
grammatical constituents are viewed either as functions or arguments,
and are associated with corresponding syntactic types.  We introduce a
first-order version of such a framework, based on the idea of encoding
syntactic types as definite clauses.  Thus, we drop all explicit
requirements of adjacency between combinable constituents, and we
capture word-order constraints simply by allowing subformulae of
complex types to share variables ranging over string positions. We are
in this way able to account for constructions involving discontinuous
constituents.  Such constructions are difficult to handle in the
more traditional version of Categorial Grammar, which is based on
the idea of strict string adjacency between combinable constituents.

The second half of this enterprise consists of showing how, for this
formalism, parsing can be implemented as theorem proving.  Our
approach to encoding types as definite clauses presupposes a
modification of standard Horn logic syntax to allow internal
implications in definite clauses. This modification is needed to
account for the types of higher-order functions and, as a consequence,
standard Prolog-like Horn logic theorem proving is not powerful
enough.  We tackle this problem by treating implication as in
intuitionistic logic. Such a treatment has been proposed elsewhere as
a useful extension of Prolog for implementing hypothetical reasoning
and modular logic programming, and finds here an application of
general interest for computational linguistics.  In fact, this same
approach to the types of higher-order functions in Categorial Grammar
can be used for the mechanism of unbounded dependencies in Generalized
Phrase Structure Grammar. In both cases, the semantic interpretation
of sen- tences can be made to follow directly from the process of
proving their syntactic well-formedness.

				   
  		      3:30 pm Tuesday, March 22
		     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: Mon, 21 Mar 88 14:57 EST
From: Anurag.Acharya@CENTRO.SOAR.CS.CMU.EDU
Subject: seminar - The nature of children's early wh-questions


		 Computation Linguistics Research Seminar

		The nature of children's early wh-questions	
		
		  	Karin Stromswold
		M.I.T. and Harvard Medical School

		    3:00-4:00pm, Thursday, 
			March 24, 1988

		       Scaife Hall 220
		  Carnegie Mellon University
	
	Wh-questions have been the subject of a great deal of recent 
linguistic research.  Which, if any, of the recently proposed analyses 
of wh-questions are supported by developmental data?  Using a computer
search, I culled all of the wh-questions asked in the CHILDes transcripts
of 12 children between the ages 0;11 and 6;6.  This yielded a corpus of 
over 16,000 wh-questions.  I used this corpus to answer the following
two questions:
	
	1. Are  subject questions gapless?
	2. Do argument and adjunct questions differ structurally?

In my talk, I will argue that the developmental data suggest that subject 
questions do have gaps and that argument and adjunct questions have 
different structural representations.

------------------------------

Date: Wed, 23 Mar 88 11:31 EST
From: Dori Wells <DWELLS@G.BBN.COM>
Subject: Language & Cognition seminar


                       BBN Science Development Program
                    Language & Cognition Seminar Series

                   PLANNING COHERENT MULTISENTENTIAL TEXT

                              Eduard Hovy
                  Information Sciences Institute of USC
                          4676 Admiralty Way
                    Marina del Rey, CA 90282-6695

                         BBN Laboratories Inc.
                          10 Moulton Street
                    Large Conference Room, 2nd Floor

                 10:30 a.m., Thursday, March 31, 1988



Abstract:  Generating multisentential text is hard.  Though most text 
generators are capable of simply stringing together more than one sentence, 
they cannot determine coherent order.  Very few programs attempt to plan 
out the structure of multisentential paragraphs. 

Clearly, the key notion is coherence.  The reason some paragraphs are 
coherent is that the information in successive sentences follows some 
pattern of inference or of knowledge with which the hearer is familiar, 
so that the hearer is able to relate each part to the whole.  To signal 
such inferences, people usually link successive blocks of text in one 
of a fixed set of ways.  The inferential nature of such linkage was 
noted by Hobbs in 1978.  In 1982, McKeown built schemas (scripts) for 
constructing some paragraphs with stereotypical structure.  Around 
the same time, after a wide-ranging linguistic study, Mann proposed 
a relatively small number of intersentential relations that suffice to 
bind together coherently most of the things people tend to speak about. 

The talk will describe a prototype text structurer that is based on the 
inferential ideas of Hobbs, uses Mann's relations, and is more general 
than the schema applier built by McKeown. The structurer takes the form 
of a standard hierarchical expansion planner, in which the relations 
act as plans and their constraints on relation fillers (represented 
in a formalism similar to Cohen and Levesque's work) as subgoals in the 
expansion. The structurer is conceived as part of a general text planner, 
but currently functions on its own. It is being tested in two domains: 
database output and expert system explanantion. 

------------------------------

Date: Wed, 16 Mar 88 21:31 EST
From: Sridharan <sridhara@cel.fmc.com>
Subject: Call for Papers - IJCAI 89 in Detroit

Eleventh International Joint Conference on 
Artificial Intelligence

Detroit, Michigan U.S.A

August 20 thru 26 1989

The International Joint Conferences on Artificial Intelligence (IJCAI) 
continue to be the premier forum for international scientific 
exchange and presentation of AI research.  The next conference will 
be held in Detroit, Michigan USA from August 20 through August 26, 
1989.  The conference is sponsored by the International Joint 
Conferences on Artificial Intelligence Inc. (IJCAII) and is co-
sponsored and hosted by the American Association for Artificial 
Intelligence (AAAI) and a broadly-based consortium of academic, 
industrial and governmental institutions in the Southeastern 
Michigan region.

The conference is designed to give representation to all subfields of 
AI including research of all kinds.  The conference will also highlight 
the relationship of AI to other related disciplines.  The technical 
program will comprise a Paper Track focusing on empirical, 
analytical, theoretical, conceptual, foundational aspects and applied 
research; and a Videotape Track focusing on applications in all 
subfields best suited for this form of presentation.

The Eleventh IJCAI will feature:

 o  an outstanding technical program;
 o  state-of-the-art exhibit of AI related hardware and software;
 o  stimulating and informative tutorials;
 o  special events that include prizes, awards, panels and workshops;
 o  visits to academic and industrial research centers and 
 o  automobile manufacturing plants;  and
 o  an interesting variety of extra-conference activities.


The official language of the conference is English, both for papers 
and videotapes.  The major areas and subareas are indicated below.

A. AI Tools and Technologies
A1. Machine Architectures, Languages, Shells
A2. Parallel and Distributed Processing
A3. Real-Time Performance

B. Fundamental Problems, Methods, Approaches
B1. Search Methods
B2. Knowledge Acquisition, Learning, Analogy
B3. Cognitive Modeling
B4. Planning, Scheduling, Reasoning about Actions
B5. Automated Deduction
B6. Patterns of Commonsense Reasoning
B7. Other issues in Knowledge Representation

C. Fundamental Applications
C1. Natural Language, Speech Understanding and Generation
C2. Perception, Vision, Robotics
C3 Intelligent Tutoring Systems
C4. Design, Manufacturing, Control

D. Perspectives and Attitudes
D1. Philosophical Foundations
D2. Social Implications


Submission Requirements and Guidelines

Important Dates

Submissions must be postmarked by:		December 12, 1988
Notification of acceptance or rejection:	March 27, 1989
Edited version to be produced by: 		April 27, 1989
Conference:					August 20-26, 1989


Program Chair: Paper submissions, reviewing, invited talks, panels, 
awards and all matters related to the technical program.
			Dr. N.S. Sridharan
			FMC Corporation, Central Engineering Labs.
			1205 Coleman Avenue, Box 580
			Santa Clara, CA 95052 USA
			(408) 289-0315		sridharan@cel.fmc.com

Videotape Track Chair: For videotape submissions, editing and 
scheduling of video presentations.  
			Dr. John Birk
			Hewlett-Packard Labs. 
			3500 Deer Creek Road, P.O. Box 10350
			Palo Alto, CA 94304-1317 USA
			(415) 857-2568 	birk@hplabs.hp.com
Other Contacts:
Local Arrangements Chair: Enquiries about local arrangements.
			Dr. Ramasamy Uthurusamy
			General Motors Research Laboratories 
			Computer Science Department
			Warren, Michigan, 48090-9055 USA 
			(313) 986-1989		samy%gmr@relay.cs.net

Tutorials, Exhibits and Registration:
			Ms. Claudia Mazzetti
			AAAI Office, 445 Burgess Drive, Suite 100
			Menlo Park, CA 94025 USA
			(415) 328-3123

General Chair:  For all general conference related matters.
			Professor Wolfgang Bibel
			Computer Science, University of British Columbia
			6356 Agricultural Road, Vancouver, B.C.
			V6T 1W5  Canada
			(604) 228-3061		bibel@ubc.csnet

IJCAII Secretary-Treasurer:  
	  		Dr. Donald Walker
			AI and Information Science Research
			Bell Communications Research
			445 South Street, MRE 2A379
			Morristown, NJ 07960-1961 USA
			(201) 829-4312		walker@flash.bellcore.com
			
Paper Track Submission:

Authors should submit five (5) copies of their papers in hard copy 
form.  Papers should be a minimum of 2000 words (about four pages 
single spaced) to a maximum of 5000 words (about 10 pages single 
spaced).  Papers should be printed on 8.5"x11.0" or European A4 
sized paper, with 1.5" margins, using 12 pt type and be of letter-
quality print (no dot matrix printouts).  Each full page figure counts 
for 500 words.

Each paper should contain the following information:  
 o  Title of paper
 o  Full names of all authors and complete addresses
 o  Abstract of 100-200 words
 o  Length of the paper in words
 o  The area/subarea in which the paper should be reviewed
 o  Declaration of multiple submissions.

If the paper submitted to IJCAI-89 is similar in substance or form to 
another paper submitted to other major conferences in 1989, this 
must be declared by the author.

Papers will be uniformly subject to peer review.  Selection criteria 
include accuracy and originality of ideas, clarity and significance of 
results and the quality of the presentation.  Late submissions will be 
automatically rejected without review.  The decision of the Program 
Committee will be final and cannot be appealed.  Papers selected will 
be scheduled for presentation and will be printed in the Proceedings.

Videotape Track Submission:

Authors should submit one (1) copy of a videotape of 15 minutes 
maximum duration, of applied research, accompanied by a 
submission letter  that includes 
 o  Title
 o  Full names of authors and complete addresses 
 o  Tape format (indicate one of NTSC, PAL or SECAM; and one of VHS   
    or .75" U-matic)
 o  Duration of tape in minutes
 o  An abstract not to exceed 100 words.   

Late submissions will be automatically rejected without review.  
Tapes will not be returned; authors must retain extra copies for 
making revisions.  All submissions will be converted to NTSC format 
before review.  Permission to copy for review purposes is 
required  and authors should indicate this in the submission letter.

This track is reserved for displaying interesting applications to real-
world problems arising in industrial, commercial, space, defense and 
educational arenas.  This track is designed to demonstrate the 
current level of usefulness of AI tools, techniques and methods. 

Tapes will be reviewed and selected for presentation during the 
conference.  The following criteria will guide the selection:
 o  Level of interest to the conference audience
 o  Clarity of goals, methods and results 
 o  Presentation quality  (including audio, video and pace)

Preference will be given to applications that show a good level of 
maturity.  Tapes which are deemed to be advertising commercial 
products, propaganda, purely expository materials, merely taped 
lectures or other material not of scientific or technical value will be 
rejected.  

------------------------------

Date: Sat, 19 Mar 88 09:08 EST
From: Jock McNaught <jock@CCL.UMIST.AC.UK>
Subject:    New Chair in CL


  UNIVERSITY OF MANCHESTER INSTITUTE OF SCIENCE AND TECHNOLOGY
                             (UMIST)

             Department of Language and Linguistics
                               and
              Centre for Computational Linguistics

               CHAIR IN COMPUTATIONAL LINGUISTICS

Applications  are  invited  for  a  new  chair  in  Computational
Linguistics  whose  purpose is to provide new leadership in  this
subject  field  and  generally  to  strengthen  the  department's
research  base.   Candidates  should  have  a  strong  record  of
academic and/or professional experience within the broad field of
computation   applied  to  natural   language   processing.   The
department's well-established and funded Centre for Computational
Linguistics  concentrates  on  postgraduate studies  and  largely
applied research.  The successful applicant is expected to play a
leading  part  in  the  stimulation  of  research   as  well   as
developing  teaching programmes and generally contributing to the
administration of the department.

Informal  enquiries  may be made to Professor J C  Sager  or  the
Registrar.

Salary will be in the professorial range with a minimum of 23,380
pounds sterling per annum (max.  permitted average 28,820  pounds
sterling). Requests for application forms and further particulars
should  be sent to the Registrar,  Room B9,  UMIST,  P O Box  88,
Manchester M60 1QD, United Kingdom, to whom completed application
forms should be returned as soon as possible.

UMIST is an equal opportunities employer.


E-mail contact addresses:

    jcs%ccl.umist.ac.uk@ean-relay.ac.uk (ean)
    jcs%ccl.umist.ac.uk@cunyvm.cuny.edu (arpa)
         jcs%ccl.umist.ac.uk@ac.uk (earn)
    $...!ukc!ccl.umist.ac.uk!jock (uucp)

Telephone: +44.61.236.3311 extension 2333

------------------------------

End of NL-KR Digest
*******************