Copy Link
Add to Bookmark
Report
NL-KR Digest Volume 07 No. 03
NL-KR Digest (Mon Jan 29 09:18:30 1990) Volume 7 No. 3
Today's Topics:
NEW JOURNAL: PHILOSOPHY & AI
Corpus-based MT
NEW BOOK: METATAXIS IN PRACTICE
Meeting Announcement: 2nd Midwest AI + CogSci Conf.
ACL Conference Information
Submissions: nl-kr@cs.rpi.edu
Requests, policy: nl-kr-request@cs.rpi.edu
Back issues are available from host archive.cs.rpi.edu [128.213.7.6] 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: Thu, 7 Dec 89 10:41:32 -0500
From: rapaport@cs.Buffalo.EDU (William J. Rapaport)
Subject: NEW JOURNAL: PHILOSOPHY & AI
Kluwer Academic Publishers announces
PHILOSOPHY AND ARTIFICIAL INTELLIGENCE
An International Journal
Editorial Focus:
Machines and Mentality
Knowledge and its Representation
Epistemic Aspects of Computer Programming
Connectionist Conceptions
Artificial Intelligence and Epistemology
Computer Methodology
Computational Approaches to Philosophical Issues
Philosophy of Computer Science
Simulation and Modeling
Ethical Aspects of Artificial Intelligence
Editor: James H. Fetzer, University of Minnesota, Duluth, MN, USA
Book Review Editor: William J. Rapaport, SUNY Buffalo, Buffalo, NY, USA
PHILOSOPHY AND ARTIFICIAL INTELLIGENCE affords an international forum
for discussion and debate of important and controversial issues concern-
ing significant developments within its areas of editorial focus.
Well-reasoned contributions from diverse theoretical perspectives are
welcome, and every effort will be made to insure their prompt publica-
tion. Among the features that are intended to make this journal dis-
tinctive within the field are these:
o Strong stands on controversial issues are specifically encouraged;
o Important articles exceeding normal journal length may appear;
o Special issues devoted to specific topics will be a regular feature;
o Review essays discussing current problem situations will appear;
o Critical responses to previously published pieces are also invited.
This journal is intended to foster a tradition of criticism within the
AI and philosophical communities on problems and issues of common con-
cern. Its scope explicitly encompasses philosophical aspects of com-
puter science. All submissions will be subject to review. Publication
will begin with a single volume of four issues per year. The first
issue will appear in January 1991.
Contributors should send 4 copies of their manuscript to:
James H. Fetzer, Editor
PHILOSOPHY AND ARTIFICIAL INTELLIGENCE
Department of Philosophy
University of Minnesota
Duluth, MN 55812
USA
phil@ub.d.umn.edu
Correspondence concerning books for review should be sent to:
William J. Rapaport, Book Review Editor
PHILOSOPHY AND ARTIFICIAL INTELLIGENCE
Center for Cognitive Science
Department of Computer Science
SUNY Buffalo
Buffalo, NY 14260
USA
rapaport@cs.buffalo.edu
rapaport@sunybcs.bitnet
Subscription information and sample copies will be available from:
Kluwer Academic Publishers Group
P.O. Box 322
3300 AH Dordrecht
The Netherlands
or
Kluwer Academic Publishers
101 Philip Drive
Assinippi Park
Norwell, MA 02061-1667
USA
------------------------------
To: nl-kr@cs.rpi.edu
Date: Wed, 17 Jan 90 12:20:02 +0100
From: Klaus Schubert <dlt1!schubert@relay.EU.net>
Phone: +31 30 911911
Fax: +31 30 944048
Telex: 40342 bso nl
Subject: Corpus-based MT
******* NEW BOOK ******* NEW BOOK ******* NEW BOOK ******* NEW BOOK ******* NE
Just published:
Victor Sadler:
WORKING WITH ANALOGICAL SEMANTICS:
Disambiguation techniques in DLT.
(= Distributed Language Translation 5)
Dordrecht/Providence: Foris Publications 1989, 256 pp.
67,- guilders softbound
The book can be bought in bookshops or ordered directly from the publisher:
Foris Publications
Postbus 509
NL-3300 AM Dordrecht
Netherlands
Foris Publications USA
P. O. Box 5904
Providence RI 02903
USA
Distributor for Japan:
Toppan Company
Sufunotomo Bldg.
1-6 Kanda Surugadai
Chiyoda-ku
Tokyo 101
Japan
- ------------------------------------------------------------------------------
How can a computer system select the appropriate meaning for a given word,
phrase or sentence in context? More precisely, where machine translation is
concerned, how can the computer select the appropriate expression in the
target language to convey the meaning identified for the source language
expression? And, in the worst case, how can it choose between a number of
virtual synonyms in the target language: e.g. between English "fast, quick,
rapid" and "swift" for the French "rapide"?
The knowledge required for solving such problems ranges all the way from
an acquaintance with language-specific collocations (a "troop" of monkeys,
but a "pride" of lions), to specialized knowledge of the subject matter,
independent of the language in which it is expressed.
Spurred by the ambitious goal of achieving high-quality machine translation
without post-editing, the author adopts an innovative approach based on the
principle of analogy: the expression in focus is matched, in its context,
against candidate examples, also in their context, in a textual knowledge
bank, and the best available match determines the translation.
This book, the second on semantics in the Distributed Language Translation
series, draws heavily on the practical experience of building the prototype
DLT system, completed in 1988. After an exhaustive survey of the various
types of ambiguity, it explains and evaluates the techniques applied in the
prototype system and goes on to offer a detailed design for the production
stage. This design introduces a powerful concept new to machine translation:
the corpus-based Bilingual Knowledge Bank, which integrates in a single
dynamic data structure (a source text linked with its translation) both the
machine dictionary and the representation of the linguistic and
extra-linguistic knowledge necessary for disambiguation.
- ------------------------------------------------------------------------------
Table of Contents
Acknowledgements
Introduction
Part I: Prototype R&D
Chapter 1.
The field: Types of ambiguity relevant to machine translation
1.1 Types and terms
1.1.1 Lexical ambiguity
1.1.2 Relational ambiguity
1.2 The scale of difficulty
Chapter 2.
The background: Development of the prototype architecture
2.1 Dumb syntax, smart semantics
2.2 The knowledge sources
2.2.1 Potential sources
2.2.2 Knowledge sources in the prototype
Chapter 3.
The theory: Assessing plausibility
3.1 What constitutes plausibility?
3.2 Methods of assessing plausibility
3.2.1 The conventional approach: features and primitives
3.2.2 The DLT prototype: collecting the evidence
3.2.3 The first DLT prototype: making meaning explicit
3.2.4 The second DLT prototype: leaving meaning implicit
Chapter 4.
The techniques: Reasoning by analogy
4.1 The word match
4.2 Disambiguating the source language
4.2.1 Lexical choice: the word pair match
4.2.2 The Y match
4.2.3 The X match
4.2.4 Combining the X and Y matches
4.2.5 Combining word pair matches
4.2.6 Resolving relational ambiguities
4.2.7 The disambiguation dialogue
4.3 Disambiguating the intermediate language
4.3.1 Contextual cues in the bilingual dictionary
4.3.2 Word match and expectancy match
4.3.3 Method of the expectancy match
4.3.4 Combining the match scores
4.3.5 Functional disambiguation
Chapter 5.
The evaluation: Tests and limitations
5.1 Some tests of \s-2SWESIL\s+2
5.1.1 Testing the word match function
5.1.2 Testing the word pair match function
5.1.3 The Meijby Test
5.1.4 Applications for referential disambiguation
5.2 The limitations of \s-2SWESIL\s+2
5.2.1 Strategic problems
5.2.2 Limitations on existing functions
5.2.3 Adding new functions
Part II: Design for a production system
Chapter 6.
The Bilingual Knowledge Bank: An integrated knowledge source for
machine translation
6.1 Rationale for a BKB
6.1.1 Objectivity: the corpus as primary knowledge resource
6.1.2 Dictionary building: the need for a bilingual corpus
6.1.3 Reversibility: the use of bilingual context
6.1.4 Consistency: discarding the LKB
6.1.5 Quantity: structuring the corpus
6.1.6 Scope: increasing the breadth of knowledge
6.1.7 Specificity: processing encyclopaedic knowledge
6.1.8 Sensitivity: distinguishing word senses
6.1.9 Productivity: towards full alignment
6.1.10 Dynamicity: updating the knowledge sources
6.1.11 Probability: relativising the frequencies
6.1.12 The conclusion: the need for an on-line corpus
6.1.13 Comparison with other research
6.2 Constructing the BKB
6.2.1 Structural disambiguation
6.2.2 Identifying translation units
6.2.3 Identifying referents
6.3 Advantages and spin-offs of the BKB approach
6.3.1 Basic advantages
6.3.2 The spin-offs
6.3.3 Possible objections
Chapter 7.
Disambiguation with a BKB: Something old and something new, something
borrowed...
7.1 Towards a new process architecture
7.1.1 Selecting the translation units
7.1.2 Challenging the selections
7.1.3 Backtracking
7.2 Examples of (simulated) BKB-based disambiguation
7.2.1 An example of lexical disambiguation
7.2.2 Choosing between TL synonyms
7.2.3 Structural disambiguation
7.2.4 Functional disambiguation
7.2.5 Referential disambiguation
7.3 Principles and techniques
7.3.1 TU selection by the Metataxor
7.3.2 Semantic coherence and the Examiner
7.3.3 Evaluating referential relations
7.3.4 Evaluating functional relations
7.3.5 Backtracking and semantic feedback
7.4 A new look at the disambiguation dialogue
7.4.1 Structural ambiguity
7.4.2 Referential ambiguity
7.4.3 Functional ambiguity
7.4.4 Lexical ambiguity
Chapter 8.
Towards intelligent disambiguation: Inferencing over the BKB
8.1 Inferring indirect referential relations
8.1.1 Inference based on explicit references
8.1.2 Inference in the absence of explicit references
8.2 Recognizing contradictions and inconsistencies
8.3 Exploring deeper implications
8.4 The BKB as a basis for inference procedures
8.4.1 Discovering inference rules
8.4.2 Applying inference rules
8.4.3 The knowledge representation
Chapter 9.
Spin-off: BKBs, MKBs and other animals...
9.1 Other translation environments
9.2 Monolingual knowledge banks
9.3 Database applications
In conclusion
References
Index
*****************************************************************************
------------------------------
To: nl-kr@cs.rpi.edu
Date: Mon, 8 Jan 90 13:00:11 +0100
From: Klaus Schubert <dlt1!schubert@relay.EU.net>
Phone: +31 30 911911
Fax: +31 30 944048
Telex: 40342 bso nl
Subject: NEW BOOK: METATAXIS IN PRACTICE
******* NEW BOOK ******* NEW BOOK ******* NEW BOOK ******* NEW BOOK ******* NE
Just published:
METATAXIS IN PRACTICE
Dependency syntax for multilingual machine translation
(= Distributed Language Translation 6)
Ed. Dan Maxwell / Klaus Schubert.
Dordrecht/Providence: Foris Publications 1989, 323 pp.
52,- guilders softbound
The book can be bought in bookshops or ordered directly from the publisher:
Foris Publictions
Postbus 509
NL-3300 AM Dordrecht
Netherlands
Foris Publications USA
P. O. Box 5904
Providence RI 02903
USA
Distributor for Japan:
Toppan Company
Sufunotomo Bldg.
1-6 Kanda Surugadai
Chiyoda-ku
Tokyo 101
Japan
- ------------------------------------------------------------------------------
Dependency theory focuses primarily on those grammatical functions that are
directly relevant to translation rather than deriving functions indirectly
from form descriptions. Metataxis is a model for dependency-based transfer
from a source to a target language, which has been implemented in current
prototype versions of the multilingual DLT machine translation system.
Theoretical linguists are sometimes criticized for showing too little
interest in frequent and normal properties of everyday written language and
for being too exclusively preoccupied with rare and marginal phenomena of the
language system. This book is an attempt to take a different perspective.
In the first part of the book, a number of authors from research centres in
Europe and Asia apply the DLT model of dependency syntax to a variety of
structurally diverse languages, thus providing evidence for the
cross-linguistic feasibility of DLT's approach. Indo-European as well as
non-Indo-European languages are described. The emphasis is on a thorough
coverage of the constructions encountered in present-day non-literary texts.
In the second part, two metataxes (transfer syntaxes) that are currently
being used in the DLT machine translation system are described. This part
includes also a contribution dealing with aspects of computational
formalization and implementation of a metataxis system.
- ------------------------------------------------------------------------------
CONTENTS
PART I
Klaus Schubert: Dependency syntax for parsing and generation
Henning Lobin: A dependency syntax of German
Ingrid Schubert: A dependency syntax of Danish
Marek 'Swidzi'nski: A dependency syntax of Polish
Probal Dasgupta: A dependency syntax of Bangla
Kalevi Tarvainen: A dependency syntax of Finnish
G'abor Pr'osz'eky / Ilona Koutny / Bal'azs Wacha: A dependency syntax of
Hungarian
Shigeru Sato: A dependency syntax of Japanese
Klaus Schubert: A dependency syntax of Esperanto
PART II
Klaus Schubert: The theory and practice of metataxis
Dorine Tamis: Esperanto-French metataxis
Dan Maxwell: English-Esperanto metataxis
Job M. van Zuijlen: Aspects of metataxis formalization
Index
*******************************************************************************
------------------------------
To: nl-kr@cs.rpi.edu
Date: Thu, 25 Jan 1990 14:05:43 PST
From: Tim Koschmann <tdk@sumex-aim.stanford.edu>
Subject: Meeting Announcement: 2nd Midwest AI + CogSci Conf.
March 30-April 1
2nd Midwest Artificial Intelligence and
Cognitive Science Society Conference
Contact:
John Dinsmore
Dept. of Computer Science
Southern Illinois University at Carbondale
Carbondale, IL 62901
614/536-2327
e-mail: GA3182@SIUCVMB.BITNET
------------------------------
To: nl-kr@cs.rpi.edu
Date: Fri, 26 Jan 90 14:22:45 EST
From: walker@flash.bellcore.com (Don Walker)
Subject: ACL Conference Information
Association for Computational Linguistics
ACL CONFERENCE INFORMATION
January 1990
PREPARATIONS FOR ACL-90 WELL UNDER WAY; 145 PAPERS SUBMITTED
The 28th Annual Meeting of the ACL will be held 6-9 June 1990 at
the University of Pittsburgh. The physical accommodations are
particularly attractive and convenient, with talks, exhibits, and
registration all in close proximity. Program and registration
information will be mailed in late February. The Program Committee
is being chaired by Bob Berwick [MIT AI Laboratory, Room 838, 545
Technology Square, Cambridge, MA 02139 USA; (+1 617) 253-8918;
berwick@wheaties.ai.mit.edu]. Local Arrangements, including exhibits
and demonstrations, are being handled by Rich Thomason [Intelligent
Systems Program, Cathedral of Learning 1004, University of Pittsburgh,
Pittsburgh, PA 15260, USA; (+1 412) 624-5791; thomason@cad.cs.cmu.edu].
Tutorials are being organized by Dan Flickinger [Hewlett-Packard
Research Laboratories, 1501 Page Mill Road, Palo Alto, CA 94304,
USA; (+1 415) 857-8789; flickinger@hp.com].
COLING-90 IN HELSINKI THIS AUGUST; NEW PROGRAM STRUCTURE
COLING-90, the 13th International Conference on Computational
Linguistics will take place in Helsinki, Finland, 20-24 August
1990. The conference will be divided into topical papers on crucial
issues in computational linguistics and brief reports with software
demonstrations. A strong emphasis is placed on controversial
proposals and their resolution. Equal time will be given to
presenting papers and discussing them. A description of the
philosophy of the Program Committee is featured in the forthcoming
issue of The FINITE STRING (Volume 15, Number 4) by the Program
Chair, Hans Karlgren [KVAL, Skeppsbron 26, S-111 30 Stockholm,
SWEDEN; (+46 8) 7896683; coling@ qzcom.bitnet or coling@com.qz.se].
Fred Karlsson is responsible for Local Arrangements [Dept of General
Linguistics, University of Helsinki, Hallituskatu 11, SF-00100
Helsinki, FINLAND; (+358 0) 1911; (+358 0) 656591; coling@finuh.bitnet].
Conference management will be handled through Riitta Ojanen [Kaleva
Travel Agency Ltd, Congress Service, Box 312, SF-00121 Helsinki,
FINLAND; (+358 0) 602711; (+358 0) 629019 fax]. Registration
information should be requested directly from Ojanen or Karlsson.
Proceedings will be available through the ACL Office after the
conference.
5TH EUROPEAN CHAPTER IN EAST GERMANY IN APRIL 1991
The Fifth Conference of the European Chapter of the ACL will be
held 9-11 April 1991 in East Berlin. The Program and Local
Arrangements Committees will be chaired by Juergen Kunze and Dorothee
Reimann [Akademie der Wissenschaften der DDR, Zentralinstitut fuer
Sprachwissenschaft, Prenzlauer Promenade 149-152, DDR-1100 Berlin,
GERMAN DEMOCRATIC REPUBLIC; (+37 2) 47 97 153 or 47 97 168; telex:
114713].
BERKELEY SITE SELECTED FOR ACL-91
The 29th Annual Meeting of the ACL will be held 18-21 June 1991 at
the University of California in Berkeley. The Program Committee
will be chaired by Doug Appelt [Artificial Intelligence Center,
SRI International, 333 Ravenswood Road, Menlo Park, CA 94025, USA;
(+1 415) 859-6150; appelt@ai.sri.com]. Responsibility for Local
Arrangements will be shared by Peter Norvig [Division of Computer
Science, University of California, 573 Evans Hall, Berkeley, CA
94720, USA; (+1 415) 642-9533; norvig@teak.berkeley.edu] and Bob
Wilensky [Division of Computer Science, University of California,
571 Evans Hall, Berkeley, CA 94720, USA; (+1 415) 642-7034;
wilensky@teak. berkeley.edu].
3RD APPLIED CONFERENCE SET FOR APRIL 1992 IN ITALY
The 3rd Conference on Applied Natural Language Processing of the
ACL will be held 1-3 April 1992 in Trento, Italy. Oliviero Stock
is Program CoChair and will be responsible for Local Arrangements
[IRST: Istituto per la Ricerca Scientifica e Tecnologica, I-38100
Trento, Loc. Pante di Povo, ITALY; (+39 461) 810105;
stock%irst@uunet.uu.net]. Lyn Bates will be Program CoChair [BBN
Systems & Technologies Corporation, 10 Moulton Street, Cambridge,
MA 02238, USA; (+1 617) 873-3634; bates@bbn.com].
------------------------------
End of NL-KR Digest
*******************
