Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model

Kouki Miyazawa, Hideaki Kikuchi, Reiko Mazuka

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

    4 Citations (Scopus)

    Abstract

    All normal humans can acquire their native phoneme systems simply by living in their native language environment. However, it is unclear as to how infants learn the acoustic expression of each phoneme of their native languages. In recent studies, researchers have inspected phoneme acquisition by using a computational model. However, these studies have used read speech that has a limited vocabulary as input and do not handle a continuous speech that is almost comparable to a natural environment. Therefore, in this study, we use natural continuous speech and build a self-organization model that simulates the cognitive ability of the humans, and we analyze the quality and quantity of the speech information that is necessary for the acquisition of the native vowel system. Our model is designed to learn values of the acoustic characteristic of a natural continuous speech and to estimate the number and boundaries of the vowel categories without using explicit instructions. In the simulation trial, we investigate the relationship between the quantity of learning and the accuracy for the vowels in a single Japanese speaker's natural speech. As a result, it is found that the vowel recognition accuracy of our model is comparable to that of an adult.

    Original languageEnglish
    Title of host publicationProceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010
    Pages2914-2917
    Number of pages4
    Publication statusPublished - 2010
    Event11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010 - Makuhari, Chiba
    Duration: 2010 Sep 262010 Sep 30

    Other

    Other11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010
    CityMakuhari, Chiba
    Period10/9/2610/9/30

    Fingerprint

    Learning
    Acoustics
    Language
    Aptitude
    Vocabulary
    Phoneme
    Unsupervised Learning
    Continuous Speech
    Research Personnel
    Native Language
    Explicit Instruction
    Self-organization
    Computational Model
    Cognitive Ability
    Acoustic Characteristics
    Natural Speech
    Simulation
    Vowel Systems

    Keywords

    • Language acquisition
    • Neural network
    • Vowels

    ASJC Scopus subject areas

    • Language and Linguistics
    • Speech and Hearing

    Cite this

    Miyazawa, K., Kikuchi, H., & Mazuka, R. (2010). Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010 (pp. 2914-2917)

    Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model. / Miyazawa, Kouki; Kikuchi, Hideaki; Mazuka, Reiko.

    Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. 2010. p. 2914-2917.

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

    Miyazawa, K, Kikuchi, H & Mazuka, R 2010, Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model. in Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. pp. 2914-2917, 11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010, Makuhari, Chiba, 10/9/26.
    Miyazawa K, Kikuchi H, Mazuka R. Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. 2010. p. 2914-2917
    Miyazawa, Kouki ; Kikuchi, Hideaki ; Mazuka, Reiko. / Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model. Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. 2010. pp. 2914-2917
    @inproceedings{854bcc38dc554b6382cd88bfc8a3fd59,
    title = "Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model",
    abstract = "All normal humans can acquire their native phoneme systems simply by living in their native language environment. However, it is unclear as to how infants learn the acoustic expression of each phoneme of their native languages. In recent studies, researchers have inspected phoneme acquisition by using a computational model. However, these studies have used read speech that has a limited vocabulary as input and do not handle a continuous speech that is almost comparable to a natural environment. Therefore, in this study, we use natural continuous speech and build a self-organization model that simulates the cognitive ability of the humans, and we analyze the quality and quantity of the speech information that is necessary for the acquisition of the native vowel system. Our model is designed to learn values of the acoustic characteristic of a natural continuous speech and to estimate the number and boundaries of the vowel categories without using explicit instructions. In the simulation trial, we investigate the relationship between the quantity of learning and the accuracy for the vowels in a single Japanese speaker's natural speech. As a result, it is found that the vowel recognition accuracy of our model is comparable to that of an adult.",
    keywords = "Language acquisition, Neural network, Vowels",
    author = "Kouki Miyazawa and Hideaki Kikuchi and Reiko Mazuka",
    year = "2010",
    language = "English",
    pages = "2914--2917",
    booktitle = "Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010",

    }

    TY - GEN

    T1 - Unsupervised learning of vowels from continuous speech based on self-organized phoneme acquisition model

    AU - Miyazawa, Kouki

    AU - Kikuchi, Hideaki

    AU - Mazuka, Reiko

    PY - 2010

    Y1 - 2010

    N2 - All normal humans can acquire their native phoneme systems simply by living in their native language environment. However, it is unclear as to how infants learn the acoustic expression of each phoneme of their native languages. In recent studies, researchers have inspected phoneme acquisition by using a computational model. However, these studies have used read speech that has a limited vocabulary as input and do not handle a continuous speech that is almost comparable to a natural environment. Therefore, in this study, we use natural continuous speech and build a self-organization model that simulates the cognitive ability of the humans, and we analyze the quality and quantity of the speech information that is necessary for the acquisition of the native vowel system. Our model is designed to learn values of the acoustic characteristic of a natural continuous speech and to estimate the number and boundaries of the vowel categories without using explicit instructions. In the simulation trial, we investigate the relationship between the quantity of learning and the accuracy for the vowels in a single Japanese speaker's natural speech. As a result, it is found that the vowel recognition accuracy of our model is comparable to that of an adult.

    AB - All normal humans can acquire their native phoneme systems simply by living in their native language environment. However, it is unclear as to how infants learn the acoustic expression of each phoneme of their native languages. In recent studies, researchers have inspected phoneme acquisition by using a computational model. However, these studies have used read speech that has a limited vocabulary as input and do not handle a continuous speech that is almost comparable to a natural environment. Therefore, in this study, we use natural continuous speech and build a self-organization model that simulates the cognitive ability of the humans, and we analyze the quality and quantity of the speech information that is necessary for the acquisition of the native vowel system. Our model is designed to learn values of the acoustic characteristic of a natural continuous speech and to estimate the number and boundaries of the vowel categories without using explicit instructions. In the simulation trial, we investigate the relationship between the quantity of learning and the accuracy for the vowels in a single Japanese speaker's natural speech. As a result, it is found that the vowel recognition accuracy of our model is comparable to that of an adult.

    KW - Language acquisition

    KW - Neural network

    KW - Vowels

    UR - http://www.scopus.com/inward/record.url?scp=79959816529&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=79959816529&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:79959816529

    SP - 2914

    EP - 2917

    BT - Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010

    ER -