Extracting Information from Spoken User Input. A Machine Learning Approach

P.K. Lendvai

    Research output: ThesisDoctoral ThesisScientific

    23 Downloads (Pure)

    Abstract

    We propose a module that performs automatic analysis of user input in spoken dialogue systems using machine learning algorithms. The input to the module is material received from the speech recogniser and the dialogue manager of the spoken dialogue system, the output is a four-level pragmatic-semantic representation of the user utterance. Our investigation shows that when the four interpretation levels are combined in a complex machine learning task, the performance of the module is significantly better than the score of an informed baseline strategy. However, via a systematic, automatised search for the optimal subtask combinations we can gain substantial improvement produced by both classifiers for all four interpretation subtasks. A case study is conducted on dialogues between an automatised, experimental system that gives information on the phone about train connections in the Netherlands, and its users who speak in Dutch. We find that drawing on unsophisticated, potentially noisy features that characterise the dialogue situation, and by performing automatic optimisation of the formulated machine learning task it is possible to extract sophisticated information of practical pragmatic-semantic value from spoken user input with robust performance. This means that our module can with a good score interpret whether the user of the system is giving slot-filling information, and for which query slots (e.g., departure station, departure time, etc.), whether the user gave a positive or a negative answer to the system, or whether the user signals that there are problems in the interaction.
    Original languageEnglish
    QualificationDoctor of Philosophy
    Supervisors/Advisors
    • Daelemans, W.M.P., Promotor
    • Bunt, Harry, Promotor
    • van den Bosch, Antal, Promotor
    • Krahmer, Emiel, Promotor
    Award date20 Dec 2004
    Place of PublicationEnschede
    Publisher
    Print ISBNs9090188746
    Publication statusPublished - 2004

    Fingerprint

    Learning systems
    Semantics
    Learning algorithms
    Managers
    Classifiers

    Cite this

    Lendvai, P.K.. / Extracting Information from Spoken User Input. A Machine Learning Approach. Enschede : [n.n.], 2004. 162 p.
    @phdthesis{e6cdf799a1094cb7aeac5ee9e456d74e,
    title = "Extracting Information from Spoken User Input. A Machine Learning Approach",
    abstract = "We propose a module that performs automatic analysis of user input in spoken dialogue systems using machine learning algorithms. The input to the module is material received from the speech recogniser and the dialogue manager of the spoken dialogue system, the output is a four-level pragmatic-semantic representation of the user utterance. Our investigation shows that when the four interpretation levels are combined in a complex machine learning task, the performance of the module is significantly better than the score of an informed baseline strategy. However, via a systematic, automatised search for the optimal subtask combinations we can gain substantial improvement produced by both classifiers for all four interpretation subtasks. A case study is conducted on dialogues between an automatised, experimental system that gives information on the phone about train connections in the Netherlands, and its users who speak in Dutch. We find that drawing on unsophisticated, potentially noisy features that characterise the dialogue situation, and by performing automatic optimisation of the formulated machine learning task it is possible to extract sophisticated information of practical pragmatic-semantic value from spoken user input with robust performance. This means that our module can with a good score interpret whether the user of the system is giving slot-filling information, and for which query slots (e.g., departure station, departure time, etc.), whether the user gave a positive or a negative answer to the system, or whether the user signals that there are problems in the interaction.",
    author = "P.K. Lendvai",
    note = "Pagination: 162",
    year = "2004",
    language = "English",
    isbn = "9090188746",
    publisher = "[n.n.]",

    }

    Lendvai, PK 2004, 'Extracting Information from Spoken User Input. A Machine Learning Approach', Doctor of Philosophy, Enschede.

    Extracting Information from Spoken User Input. A Machine Learning Approach. / Lendvai, P.K.

    Enschede : [n.n.], 2004. 162 p.

    Research output: ThesisDoctoral ThesisScientific

    TY - THES

    T1 - Extracting Information from Spoken User Input. A Machine Learning Approach

    AU - Lendvai, P.K.

    N1 - Pagination: 162

    PY - 2004

    Y1 - 2004

    N2 - We propose a module that performs automatic analysis of user input in spoken dialogue systems using machine learning algorithms. The input to the module is material received from the speech recogniser and the dialogue manager of the spoken dialogue system, the output is a four-level pragmatic-semantic representation of the user utterance. Our investigation shows that when the four interpretation levels are combined in a complex machine learning task, the performance of the module is significantly better than the score of an informed baseline strategy. However, via a systematic, automatised search for the optimal subtask combinations we can gain substantial improvement produced by both classifiers for all four interpretation subtasks. A case study is conducted on dialogues between an automatised, experimental system that gives information on the phone about train connections in the Netherlands, and its users who speak in Dutch. We find that drawing on unsophisticated, potentially noisy features that characterise the dialogue situation, and by performing automatic optimisation of the formulated machine learning task it is possible to extract sophisticated information of practical pragmatic-semantic value from spoken user input with robust performance. This means that our module can with a good score interpret whether the user of the system is giving slot-filling information, and for which query slots (e.g., departure station, departure time, etc.), whether the user gave a positive or a negative answer to the system, or whether the user signals that there are problems in the interaction.

    AB - We propose a module that performs automatic analysis of user input in spoken dialogue systems using machine learning algorithms. The input to the module is material received from the speech recogniser and the dialogue manager of the spoken dialogue system, the output is a four-level pragmatic-semantic representation of the user utterance. Our investigation shows that when the four interpretation levels are combined in a complex machine learning task, the performance of the module is significantly better than the score of an informed baseline strategy. However, via a systematic, automatised search for the optimal subtask combinations we can gain substantial improvement produced by both classifiers for all four interpretation subtasks. A case study is conducted on dialogues between an automatised, experimental system that gives information on the phone about train connections in the Netherlands, and its users who speak in Dutch. We find that drawing on unsophisticated, potentially noisy features that characterise the dialogue situation, and by performing automatic optimisation of the formulated machine learning task it is possible to extract sophisticated information of practical pragmatic-semantic value from spoken user input with robust performance. This means that our module can with a good score interpret whether the user of the system is giving slot-filling information, and for which query slots (e.g., departure station, departure time, etc.), whether the user gave a positive or a negative answer to the system, or whether the user signals that there are problems in the interaction.

    M3 - Doctoral Thesis

    SN - 9090188746

    PB - [n.n.]

    CY - Enschede

    ER -