Re: [Corpora-List] Seeking new results on WSD in applications

From: Philip Resnik (resnik@umiacs.umd.edu)
Date: Fri Oct 21 2005 - 18:28:10 MET DST

  • Next message: Steven Johnson: "[Corpora-List] Question re access to corpora of transcribed conversational Spanish"

    I recently posted a query to this list asking for pointers to "any
    recent (2004 or later) results bearing on the usefulness of word sense
    disambiguation (WSD) for end-user applications (such as machine
    translation or question answering, and excluding WSD in support of
    NLP-internal tasks such as parsing, semantic role labeling, learning
    selectional preferences, etc.)." I'd like to express my thanks to
    Chris Brew, Chris Callison-Burch, Katrin Erk, Jesus Jimenez, Rada
    Mihalcea, Paul Rayson, and Luis Villarejo for their replies. Here's
    my attempt at organizing/summarizing.

    One category of responses concerned recent work on using WSD
    techniques in connection with lexical selection in MT (with or without
    an explicit sense inventory). Relevant references include:

      Clara Cabezas and Philip Resnik, ``Using WSD Techniques for Lexical
      Selection in Statistical Machine Translation'', Technical report
      CS-TR-4736/LAMP-TR-124/UMIACS-TR-2005-42, July 2005.
      http://lampsrv01.umiacs.umd.edu/pubs/TechReports/LAMP_124/LAMP_124.pdf

      Marine Carpuat and Dekai Wu. "Word Sense Disambiguation
      vs. Statistical Machine Translation". 43rd Annual Meeting of the
      Association for Computational Linguistics (ACL-2005). Ann Arbor, MI:
      Jun 2005.

      Marine Carpuat and Dekai Wu. "Evaluating the Word Sense
      Disambiguation Performance of Statistical Machine
      Translation". Second International Joint Conference on Natural
      Language Processing (IJCNLP-2005). Jeju, Korea: Oct 2005.

      Je'su's Gime'nez, Llui's Ma'rquez, and German Rigau, Automatic
      Translation of WordNet Glosses. (EUROLAN'05 Cross-Language Knowledge
      Induction Workshop), July 2005.
      http://www.lsi.upc.edu/~jgimenez/PUBS/eurolan2005-gmr.pdf

      D. Vickrey, L. Biewald, M. Teyssier, and D. Koller. "Word-Sense
      Disambiguation for Machine Translation." Conference on Empirical
      Methods in Natural Language Processing (EMNLP), October 2005.
      http://ai.stanford.edu/~koller/papers.cgi?topic=9

    Another set of references concerns use of WSD-like tagging of semantic
    categories in applications involving "interpretation of results" by
    the user, such as requirements engineering, content analysis,
    dictionary search tools, and providing assistance for human
    translators.

      Lo"fberg L, Juntunen J-P, Nykanen A, Varantola K, Rayson P, Archer
      D. (2004). Using a semantic tagger as dictionary search tool. In
      Williams G. and Vessier S. (eds.) Proceedings of the 11th EURALEX
      (European Association for Lexicography) International Congress
      (Euralex 2004), Lorient, France, 6-10 July 2004. Universite' de
      Bretagne Sud. Volume I, pp. 127-134.

      S. Sharoff, P. Rayson, O. Mudraya, A. Wilson and T. McEnery
      (2004). A tool for assisting translators using automatic semantic
      annotation. Presented at Corpus Use and Learning to Translate
      (CULT-BCN) Barcelona, January 22nd-24th 2004.

      Sawyer, P., Rayson, P. and Cosh, K. (forthcoming) Shallow Knowledge
      as an Aid to Deep Understanding in Early-Phase Requirements
      Engineering. IEEE Transactions on Software Engineering.

    Two replies related to enabling tasks, not end-user applications per se:

      Colin Bannard and Chris Callison-Burch. "Paraphrasing with bilingual
      parallel corpora." ACL-2005.
      http://www-csli.stanford.edu/~ccb/publications/ACL_2005a.pdf. They
      "found that WSD played an important role, though only did simulated
      WSD rather than doing it properly" (Chris C-B, p.c.)

      Katrin Erk is "working on the automatic assignment of FrameNet
      frames, which can be seen as a WSD task", and she suggested looking
      at Tatu and Moldovan, "A Semantic Approach to Recognizing Textual
      Entailment" (EMNLP-2005). She also mentioned related work aimed at
      using FrameNet for Question Answering:
      http://www.springerlink.com/(k23eicebjg3n5w45qp0frc45)/app/home/contribution.asp?referrer=parent&backto=issue,6,42;journal,579,2263;linkingpublicationresults,1:105633,1

    Finally, Chris Brew kindly suggested four other places/directions in
    which to look for applications: ICALL (Intelligent Computer-Aided
    Language Learning); support for lexicographers; support for librarians
    assigning Library of Congress labels (or similar) to books (or
    similar); disambiguation of user queries to search engines in order to
    place ads within the search results. (If anyone has specific recent
    references related to these suggestions I'd love to know what they
    are.)

    Thanks again to everyone for their responses!

      Philip

      ----------------------------------------------------------------
      Philip Resnik, Associate Professor
      Department of Linguistics and Institute for Advanced Computer Studies

      1401 Marie Mount Hall UMIACS phone: (301) 405-6760
      University of Maryland Linguistics phone: (301) 405-8903
      College Park, MD 20742 USA Fax: (301) 314-2644 / (301) 405-7104
      http://umiacs.umd.edu/~resnik E-mail: resnik@umiacs.umd.edu



    This archive was generated by hypermail 2b29 : Fri Oct 21 2005 - 18:44:52 MET DST