http://www.inria.fr/acacia/ECAI2002-OLT/

Welcome to the

Workshop on 
Machine Learning and Natural Language Processing for Ontology Engineering

 held in conjunction with the ECAI'02 conference

Lyon (France), July 22-23 2002

Scope

  
Call for papers (pdf file)  
Paper evaluation  
Participation conditions  
Workshop time-table
REGISTRATION OPENED
Important dates
Early Registration for to the workshop and the ECAIconference: May 30th 2002
History  
Scientific Support
Accepted papers available on-line
Workshop chairs
Workshop program
Workshop organizing committee  

Program committee

  

Supported by : TIA, Atala , OntoWeb

TIA

Scope:

Ontologies serve as a means to establish a conceptually concise basis for communicating knowledge for many purposes. Engineering them may be considered as a process that, starting from knowledge sources, produces a structured conceptual model. With this workshop, among all knowledge sources, we ill pay special interest to texts (technical documentation, interview transcripts, handbooks, documents gathered from the Web and so on), semi-structured data and existing knowledge bases. Among all possible techniques, reports about the use of NLP tools, linguistic approaches, machine learning algorithms and any combination of these will be encouraged. As a matter of fact, efforts in the machine learning community pursue the induction of more concise and more expressive knowledge structures (e.g. relational learning). Moreover, results in machine learning, NLP, linguistics are mature enough to be worth being integrated into knowledge engineering methods. It is time to evaluate how their combination could improve the efficiency of building ontologies as well as their quality and their relevance.
Building such knowledge structures raises some theoretical issues that are little studied. The originality of this workshop is to call for several disciplines such as linguistics, terminology, natural language processing, knowledge representation and machine learning to go deeply into these issues and related epistemological foundations. It will give a unique opportunity to these communities to confront their views and results. To this end, the workshop will focus not only on practical and technical problems but also on a theoretical reflection about building, maintaining and reusing terminological resources and ontologies. We would also like to debate the nature of ontologies, their genericity according to applications and sources. Cross-disciplinary contributions, in particular those involving linguistics, are strongly encouraged. 

Call for papers:

Ontologies serve as a means to establish a conceptually concise basis for communicating knowledge for many purposes. Recent years have seen a surge of interest that deals with the discovery, automatic or semi-automatic creation of complex, multirelational knowledge structures such as ontologies. For example, the natural language community tries to acquire word semantics from texts, database researchers tackle the problem of schema induction, and numerous intelligent information agents are built by learning complex structures from semi-structured input (HTML, XML files. This interest converges with the recent proposals from various communities to build a Semantic Web. One popular solution relies on ontologies and annotations of Web resources w.r.t. these ontologies. The size of the Web implies to be able to automate some parts of the process and to scale it up. Therefore NLP (Natural Language Processing) tools as well as learning techniques seem to be very promising. ).

The aims of this workshop are both to evaluate the maturity of these various results and to facilitate the scientific confrontation between the corresponding research communities.

Engineering ontologies may be considered as a process that, starting from (possibly evolutive) knowledge sources, produces a structured conceptual model. During this workshop, among all knowledge sources, special interest will be payed to texts (technical documentation, interview transcripts, handbooks, documents gathered from the Web and so on), semi-structured data and existing knowledge bases. Among all possible techniques, reports about the use of NLP tools, linguistic approaches, machine learning algorithms and any combination of these are encouraged. As a matter of fact, efforts in the machine learning community pursue the induction of more concise and more expressive knowledge structures (e.g. relational learning). Moreover, results (principles, methods and techniques) in machine learning, NLP, linguistics are mature enough to be worth being integrated into knowledge engineering methods. It is time to evaluate how their combination could improve the efficiency of building ontologies as well as their quality and their relevance.

Engineering such knowledge structures raises some theoretical issues that are little studied. The originality of this workshop is to call for several disciplines such as linguistics, terminology, natural language processing, knowledge representation and machine learning to go deeply into these issues and related epistemological foundations. It will give a unique opportunity to these communities to confront their views and results. To this end, the workshop will focus not only on practical and technical problems but also on a theoretical reflection about building, maintaining and reusing terminological resources and ontologies. We would also like to debate the nature of ontologies, their genericity according to applications and sources. Cross-disciplinary contributions, in particular those involving linguistics, are strongly encouraged.

Technical and theoretical issues to be discussed at the workshop include, but are not limited to:

  • Status of texts as knowledge sources, connections between ontologies and

    • texts
  • Linguistic and terminological ressources as knowledge sources
  • Learning from machine-readable dictionaries
  • Extending existing ontologies (Wordnet)
  • Text Mining for building ontologies
  • Linguistics (techniques and principles) to build ontologies
  • Natural Language Processing (NLP) tools for building and maintaining ontologies
  • Ontologies for Text and Document Processing
  • Learning selectional restrictions
  • Multi-relational learning, Inductive Logic Programming
  • Instance mining
  • Learning ontologies with inferences (e.g. using description logics)
  • Cooperative learning of ontologies
  • Ontologies and NLP tools for the semantic web
  • Learning ontologies from the Web (from DTDs, XML files, RDF files)

    • Paper evaluation

    Papers should be no longer than 5000 words. They can either report research work, practical experiments whether completed or in progress. Papers discussing more theoretical questions are also welcome.

    Each paper will be reviewed by two persons from the program committee having in mind the willingness to promote discussions and debates rather than selection. Papers will be published in paperback proceedings distributed to the workshop participants and available on-line after June 10th 2002. Please use the same format as the one suggested for the conference.

    Send papers by email (html, ps or pdf files) to OLT2002@sophia.inria.fr before March 15th.

    Participation conditions

    Beside the papers' authors, anyone wishing to take part in this workshop should send a one page abstract about his/her motivations to attend the workshop and/or his/her recent work related to the workshop topic. This page should also contain one question-issue to be debated during the workshop. Motivation abstracts will be reviewed.

    Send your text to OLT2002@sophia.inria.fr before May 24th.

    All workshop participants are required to register for the ECAI 2002 main conference.

    Workshop time-table

    In order to make exchanges easier during the workshop, each paper will be assigned a discutant selected among the authors of the other papers. Discutants will contribute to paper presentations and discussions.

    Paper presentations will be organized into thematic sessions. A large amount of time will be dedicated to debates at the end of each session or during specific sessions according to the questions suggested by the participants (see participation conditions).

    Important dates

    History

    This workshop is the first attempt to set up a dialog between two emerging communities that used to organise two workshop series : the "Ontology Learning" workshops and the "Ontologies and texts" workshops. Previous editions of "Ontology Learning" took place during ECAI2000 and IJCAI2001. The first meeting of "Ontologies and texts" was held at EKAW2000.

    Scientific Support

    This workshop is promoted by the following scientific associations:

    French working group on Terminology and Artificial Intelligence (TIA);

    ATALA, the French Association for Natural Language Processing;

    A3CTE, the French working group on Applications, Learning and Knowledge Acquisition from Electronic Documents

    ?? Ontoweb ???

    Workshop chairs

    Nathalie AUSSENAC-GILLES (IRIT, Toulouse, F)
    Alexander MAEDCHE (FZI, Univ. of Karlsruhe, G)

    Workshop organizing committee

    Brigitte BIEBOW (LIPN, Paris, Fr)
    Anne CONDAMINES (ERSS, Toulouse, Fr)
    Rose DIENG-KUNTZ (INRIA, Sophia Antipolis, Fr)
    Adeline NAZARENKO (LIPN, Paris, Fr)
    Claire NEDELLEC (LRI, Paris, Fr)
    Stephen STAAB (AIFB, Karslruhe, De)

    Program committee

    Page maintained by : Sophie Honnorat
    Last update: 07-Jui-2002