• DocumentCode
    442047
  • Title

    Fusion approach of formal semantics for the parsing of UDC sentences

  • Author

    Xu, Er-Qing

  • Author_Institution
    Dept. of Linguistics, Zhejiang Univ., Hangzhou, China
  • Volume
    6
  • fYear
    2005
  • fDate
    18-21 Aug. 2005
  • Firstpage
    3734
  • Abstract
    Aiming at the problem that the currently available parsing methods have difficulty in obtaining the D-structure/meaning of UDC sentences, this paper puts forward the fusion approach of formal semantics (FAFS), which combines λ-abstraction, λ-calculus, HPSG (head-driven phrase structure grammar), and knowledge in the realm of functional grammar. Firstly, employing the context-sensitive attribute grammar of HPSG, FAFS establishes the S-structure of UDCs, distinguishes between strong and weak UDCs, and extracts the fillers for the traces. Then FAFS obtains the WFF of UDCs according to the HPSG parse tree. In view of functional grammar, strong and weak UDCs are in fact ellipsis and substitution. With the result of HPSG parsing and filler extraction, as well as the WFF of UDCs, FAFS employs λ-abstraction and λ-calculus to derive the D-structure/meaning of a UDC and complete the parsing task. The overall computational efficiency of FAFS is in the cubic time, and thus FAFS is efficient. Finally, application examples were examined.
  • Keywords
    grammars; lambda calculus; natural languages; programming language semantics; context-sensitive attribute grammar; filler extraction; formal semantics; functional grammar; fusion approach; head-driven phrase structure grammar; lambda-abstraction; lambda-calculus; natural language understanding; parsing method; unbounded dependency construction sentence; Calculus; Computational efficiency; Computer science; Cybernetics; Fuses; Machine learning; Natural language processing; Natural languages; Turing machines; FAFS; HPSG; Natural language understanding; lambda—abstraction; lambda—calculus;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
  • Conference_Location
    Guangzhou, China
  • Print_ISBN
    0-7803-9091-1
  • Type

    conf

  • DOI
    10.1109/ICMLC.2005.1527590
  • Filename
    1527590