Appel: LLL

Philippe Blache pb at
Tue Mar 14 12:19:25 UTC 2000

From: Claire Nedellec <Claire.Nedellec at>


                           CALL FOR PAPERS


                  13th - 14th September 2000, Lisbon - Portugal

                    Co-located with ICGI and CoNLL



Our purpose  is to provide  a forum for  discussion on all  aspects of
learning linguistic knowledge in logic.

This workshop  is a  follow-up of the  succesful LLL workshop  held in
1999    in    Bled,     (Slovenia)    and    co-located    with    ICML

It  will be co-located  with the  International Conference  on Grammar
Inference (ICGI) ( and the Conference
on Natural Language Learning (CoNLL)

 Aims and scope of the conference

The fact  that more  and more people  are interested in  the automatic
acquisition of lexicons  is due to the progress  in the development of
applications  in NLP,  terminology acquisition,  indexing, information
extraction,  retrieval, question-answering, etc.   Relational learning
seems  like  a valuable  alternative  to  data  analysis in  some  NLP
domains.   This is clearly  shown by  the recent  success of  both NLP
methods based on ILP or non-classic logics, and hybrid methods.

Interest in learning linguistic  knowledge has grown steadily over the
last  15  years.   As  compared  to  manual  acquisition,  specialized
resources can  be learned,  revised and extended  with respect  to the
task at hand for much less cost.

Despite the degree  of variation in the applications  and resources we
want  to  acquire, most  of  them  are learned  in  the  same way:  by
observing  regularities among  the  co-occurence of  phenomena in  the
corpus.  Therefore, a  large  amount  of work  is  naturally based  on
statistics, and attempts to develop robust and large-scale methods.

Moreover,  relational learning  and logic-based  learning  have proved
their capacity  to learn complex structured  knowledge from structured
data  and explicit  background knowledge.  Compared to  data analysis,
some of the  major advantages here are: a better  means to express the
representation;  a  method  that   is  easier  to  understand;  and  a
comprehensible learning result.

As a consequence,  interest is growing for a  corpus-based learning of
structures    that   represent    linguistic    resources   such    as
predicate-argument structures, grammars, ontologies, etc.

The goal of  this workshop is to bring  together researchers from many
subfields  of  AI  who  are  working  on  learning  from  text,  while
emphasizing the logic-based learning techniques and algorithms.  These
techniques include,

     * Instance-based and clustering approaches in relational learning
     * Scalability issues (applying Logic-based methods to large data sets)
     * Logical approaches to statistical NLP
     * Theory revision
     * Explanation-based learning
     * Higher-order logic for LLL
     * Handling very complex terms
     * Multi-predicate learning
     * Collaborative and interactive learning
     * Learning in description logics
     * Combinations of approaches and multi-strategy learning
     * Evaluation techniques

     * Information indexing, filtering, retrieval, extraction
     * Text classification methods
     * Question answering
     * Learning ontologies, thesauri and lexicon
     * Learning terminology
     * Learning predicate-argument structure
     * Shallow parsing
     * Learning grammar
     * Learning subcategorisation frames
     * Part-of-speech tagging
     * Morphosyntactic tagging
     * Morphological analysis

 In addition to these topics,  the workshop covers all theoretical and
methodological issues concerning learning from text using logic- based
techniques.  Submissions describing innovative applications are also

 Important dates

  Submission of papers by May, 15, 2000
  Acceptance notices mailed by June, 19, 2000
  Final, camera-ready papers due by July, 14, 2000


 The workshop  will be two  full days, including invited  talks, paper
presentations,  poster presentations,  and numerous  opportunities for
discussion.   There   will  be   joint  sessions  with   the  workshop
"International  Conference  on   Grammar  Inference"  (ICGI)  and  the
"Conference on Natural Language  Learning" (CoNLL) on topics of common
interest.   Joint  sessions  will  include  invited  talks  and  paper
presentations, depending on submissions.

 Submission procedure

  Full papers may be up to 10  pages, short papers up to 6 pages, both
in a  11 point  font and single-spaced.   We accept  either electronic
submission (preferred),  in Postscript, PDF  or Word format,  or paper
submissions (in 4 copies) to the following address:

  Claire Nedellec
  LLL workshop
  LRI, Bat 490                           e-mail: cn at
  Universite Paris-Sud                   Tel: +33 (0)1 69 15 66 26
  F-91405 Orsay                          Fax: +33 (0)1 69 15 65 86

 Program committee

  Program chair

  Claire Nedellec (LRI, University of Paris-Sud, France)


  Pieter Adriaans (Syllogic and University of Amsterdam, Netherlands)
  Roberto Basili (University of Roma, Italy)
  Gilles Bisson (INRIA, Grenoble, France)
  Henrik Boström (University of Stockholm, Sweden)
  Gosse Bouma (University of Groningen, Netherlands)
  James Cussens (University of York, UK)
  Tomaz Erjavec (Institute Jozef Stefan, Slovenia)
  Daniel Kayser (LIPN, University Paris-Nord, France)
  Suresh Manandhar (University of York, UK)
  Guenter Neumann (DFKI, Germany)
  Steve Pulman (University of Cambridge, UK)
  Christer Samuelsson (XRCE, Grenoble, France)
  Stefan Wrobel (University of Magdeburg, Germany)

  Arlindo Oliveira (INESC, Lisbon, Portugal)

 Claire Nedellec
 Inference and Machine Learning Group     e-mail: cn at
 LRI, Bat 490                             Tel: +33 (0)1 69 15 66 26
 Universite Paris-Sud                     Fax: +33 (0)1 69 15 65 86
 F-91405 Orsay
 Web :

Message diffusé par la liste Langage Naturel <LN at>
Informations, abonnement :
English version          :
Archives                 :

More information about the Ln mailing list