Unsupervised Word Segmentation and Lexicon Discovery Using Acoustic Word Embeddings

@article{Kamper2016UnsupervisedWS,
  title={Unsupervised Word Segmentation and Lexicon Discovery Using Acoustic Word Embeddings},
  author={Herman Kamper and Aren Jansen and Sharon Goldwater},
  journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing},
  year={2016},
  volume={24},
  pages={669-679}
}
  • Herman Kamper, Aren Jansen, Sharon Goldwater
  • Published 2016
  • Computer Science
  • IEEE/ACM Transactions on Audio, Speech, and Language Processing
  • In settings where only unlabeled speech data is available, speech technology needs to be developed without transcriptions, pronunciation dictionaries, or language modelling text. A similar problem is faced when modeling infant language acquisition. In these cases, categorical linguistic structure needs to be discovered directly from speech audio. We present a novel unsupervised Bayesian model that segments unlabeled speech and clusters the segments into hypothesized word groupings. The result… CONTINUE READING

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 51 CITATIONS

    A segmental framework for fully-unsupervised large-vocabulary speech recognition

    VIEW 10 EXCERPTS
    CITES BACKGROUND & METHODS

    Unsupervised neural and Bayesian models for zero-resource speech processing

    VIEW 4 EXCERPTS
    CITES BACKGROUND & METHODS

    Phoneme Based Embedded Segmental K-Means for Unsupervised Term Discovery

    VIEW 1 EXCERPT

    Audio Word2vec: Sequence-to-Sequence Autoencoding for Unsupervised Learning of Audio Segmentation and Representation

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Unsupervised learning of cross-modal mappings between speech and text

    VIEW 5 EXCERPTS
    CITES BACKGROUND
    HIGHLY INFLUENCED

    Semantic Speech Retrieval With a Visually Grounded Model of Untranscribed Speech

    VIEW 1 EXCERPT

    FILTER CITATIONS BY YEAR

    2016
    2020

    CITATION STATISTICS

    • 1 Highly Influenced Citations

    • Averaged 10 Citations per year from 2018 through 2020

    References

    Publications referenced by this paper.
    SHOWING 1-4 OF 4 REFERENCES

    A hierarchical system for word discovery exploiting DTW-based initialization

    VIEW 6 EXCERPTS
    HIGHLY INFLUENTIAL

    Unsupervised Pattern Discovery in Speech

    VIEW 15 EXCERPTS
    HIGHLY INFLUENTIAL

    Conjugate Bayesian analysis of the Gaussian distribution

    • K. P. Murphy
    • 2007, http://www.cs.ubc.ca/~murphyk/mypapers.html
    • 2007
    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL