Generalization of state-observation-dependency in Partly Hidden Markov Models

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    3 Citations (Scopus)

    Abstract

    Generalized Partly Hidden Markov Model (GPHMM) is proposed by modifying Partly Hidden Markov Model (PHMM), and it is successfully applied to the speech recognition. PHMM, which was proposed in our previous paper, is the novel stochastic model, in which the pairs of the hidden states (H-state) and the observable states (O-state) determine the stochastic phenomena of the current observation and the next state transition. In the formulation of PHMM, we used common pair of H-state and O-state to determine both of these phenomena. In the formulation of GPHMM proposed here, we use common H-state but different O-states for the current observation and for the next state transition separately. This slight modification brought the big flexibility in the modeling of phenomena. Experimental results showed the effectiveness of GPHMM (without delta parameters): it reduced the word error by 17% compared to triphone HMM (with delta parameters), respectively.

    Original languageEnglish
    Title of host publication7th International Conference on Spoken Language Processing, ICSLP 2002
    PublisherInternational Speech Communication Association
    Pages2673-2676
    Number of pages4
    Publication statusPublished - 2002
    Event7th International Conference on Spoken Language Processing, ICSLP 2002 - Denver, United States
    Duration: 2002 Sep 162002 Sep 20

    Other

    Other7th International Conference on Spoken Language Processing, ICSLP 2002
    CountryUnited States
    CityDenver
    Period02/9/1602/9/20

    Fingerprint

    Hidden Markov Model
    flexibility
    Modeling
    Speech Recognition

    ASJC Scopus subject areas

    • Language and Linguistics
    • Linguistics and Language

    Cite this

    Ogawa, T., & Kobayashi, T. (2002). Generalization of state-observation-dependency in Partly Hidden Markov Models. In 7th International Conference on Spoken Language Processing, ICSLP 2002 (pp. 2673-2676). International Speech Communication Association.

    Generalization of state-observation-dependency in Partly Hidden Markov Models. / Ogawa, Tetsuji; Kobayashi, Tetsunori.

    7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association, 2002. p. 2673-2676.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Ogawa, T & Kobayashi, T 2002, Generalization of state-observation-dependency in Partly Hidden Markov Models. in 7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association, pp. 2673-2676, 7th International Conference on Spoken Language Processing, ICSLP 2002, Denver, United States, 02/9/16.
    Ogawa T, Kobayashi T. Generalization of state-observation-dependency in Partly Hidden Markov Models. In 7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association. 2002. p. 2673-2676
    Ogawa, Tetsuji ; Kobayashi, Tetsunori. / Generalization of state-observation-dependency in Partly Hidden Markov Models. 7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association, 2002. pp. 2673-2676
    @inproceedings{a1299c3adfd449d7b4788af4187864ee,
    title = "Generalization of state-observation-dependency in Partly Hidden Markov Models",
    abstract = "Generalized Partly Hidden Markov Model (GPHMM) is proposed by modifying Partly Hidden Markov Model (PHMM), and it is successfully applied to the speech recognition. PHMM, which was proposed in our previous paper, is the novel stochastic model, in which the pairs of the hidden states (H-state) and the observable states (O-state) determine the stochastic phenomena of the current observation and the next state transition. In the formulation of PHMM, we used common pair of H-state and O-state to determine both of these phenomena. In the formulation of GPHMM proposed here, we use common H-state but different O-states for the current observation and for the next state transition separately. This slight modification brought the big flexibility in the modeling of phenomena. Experimental results showed the effectiveness of GPHMM (without delta parameters): it reduced the word error by 17{\%} compared to triphone HMM (with delta parameters), respectively.",
    author = "Tetsuji Ogawa and Tetsunori Kobayashi",
    year = "2002",
    language = "English",
    pages = "2673--2676",
    booktitle = "7th International Conference on Spoken Language Processing, ICSLP 2002",
    publisher = "International Speech Communication Association",

    }

    TY - GEN

    T1 - Generalization of state-observation-dependency in Partly Hidden Markov Models

    AU - Ogawa, Tetsuji

    AU - Kobayashi, Tetsunori

    PY - 2002

    Y1 - 2002

    N2 - Generalized Partly Hidden Markov Model (GPHMM) is proposed by modifying Partly Hidden Markov Model (PHMM), and it is successfully applied to the speech recognition. PHMM, which was proposed in our previous paper, is the novel stochastic model, in which the pairs of the hidden states (H-state) and the observable states (O-state) determine the stochastic phenomena of the current observation and the next state transition. In the formulation of PHMM, we used common pair of H-state and O-state to determine both of these phenomena. In the formulation of GPHMM proposed here, we use common H-state but different O-states for the current observation and for the next state transition separately. This slight modification brought the big flexibility in the modeling of phenomena. Experimental results showed the effectiveness of GPHMM (without delta parameters): it reduced the word error by 17% compared to triphone HMM (with delta parameters), respectively.

    AB - Generalized Partly Hidden Markov Model (GPHMM) is proposed by modifying Partly Hidden Markov Model (PHMM), and it is successfully applied to the speech recognition. PHMM, which was proposed in our previous paper, is the novel stochastic model, in which the pairs of the hidden states (H-state) and the observable states (O-state) determine the stochastic phenomena of the current observation and the next state transition. In the formulation of PHMM, we used common pair of H-state and O-state to determine both of these phenomena. In the formulation of GPHMM proposed here, we use common H-state but different O-states for the current observation and for the next state transition separately. This slight modification brought the big flexibility in the modeling of phenomena. Experimental results showed the effectiveness of GPHMM (without delta parameters): it reduced the word error by 17% compared to triphone HMM (with delta parameters), respectively.

    UR - http://www.scopus.com/inward/record.url?scp=0141701004&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=0141701004&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:0141701004

    SP - 2673

    EP - 2676

    BT - 7th International Conference on Spoken Language Processing, ICSLP 2002

    PB - International Speech Communication Association

    ER -