Hopfield Models as Neural-Network Acceptors

Marc F.J. Drossaers

    Research output: Chapter in Book/Report/Conference proceedingChapterProfessional

    8 Downloads (Pure)

    Abstract

    The use of neural networks for integrated lin­guistic analysis may be profitable. This paper presents the first results of our research on that subject: a Hopfield model for syntactical analy­sis. We construct a neural network as an imple­mentation of a bounded push-down automaton, which can accept context-free languages with lim­ited center-embedding. The network's behavior can be predicted a priori, so the presented the­ory can be tested. The operation of the network as an implementation of the acceptor is provably correct. Furthermore we found a solution to the problem of spurious states in Hopfield models: we use them as dynamically constructed repre­sentations of sets of states of the implemented acceptor. The so-called neural-network acceptor we propose, is fast but large.
    Original languageEnglish
    Title of host publicationConnectionism and Natural Language Processing
    Subtitle of host publicationProceedings of the third Twente Workshop on Language Technology
    EditorsM.F.J. Drossaers, A. Nijholt
    Place of PublicationEnschede
    PublisherUniversity of Twente
    Pages57-66
    Publication statusPublished - 1992
    Event3rd Twente Workshop on Language Technology, TWLT 3 - University of Twente, Enschede, Netherlands
    Duration: 12 May 199213 May 1992
    Conference number: 3

    Publication series

    NameTwente Workshop on Language Technology
    PublisherUniversity of Twente
    Volume3
    ISSN (Print)0929-0672
    NameMemoranda Informatica
    PublisherUniversity of Twente
    Number92-64
    ISSN (Print)0924-3755

    Workshop

    Workshop3rd Twente Workshop on Language Technology, TWLT 3
    Abbreviated titleTWLT
    CountryNetherlands
    CityEnschede
    Period12/05/9213/05/92

    Fingerprint Dive into the research topics of 'Hopfield Models as Neural-Network Acceptors'. Together they form a unique fingerprint.

    Cite this