Hierarchical sparse modeling using Spike and Slab priors

Yuanming Suo, Minh Dao, Trac Tran, Umamahesh Srinivas, Vishal Monga

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

    9 Citations (Scopus)

    Abstract

    Sparse modeling has demonstrated its superior performances in many applications. Compared to optimization based approaches, Bayesian sparse modeling generally provides a more sparse result with a knowledge of confidence. Using the Spike and Slab priors, we propose the hierarchical sparse models for the scenario of single task and multitask - Hi-BCS and CHi-BCS. We draw the connections of these two methods to their optimization based counterparts and use expectation propagation for inference. The experiment results using synthetic and real data demonstrate that the performance of Hi-BCS and Chi-BCS are comparable or better than their optimization based counterparts.

    Original languageEnglish (US)
    Title of host publication2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings
    Pages3103-3107
    Number of pages5
    DOIs
    StatePublished - Oct 18 2013
    Event2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Vancouver, BC, Canada
    Duration: May 26 2013May 31 2013

    Publication series

    NameICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
    ISSN (Print)1520-6149

    Other

    Other2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013
    CountryCanada
    CityVancouver, BC
    Period5/26/135/31/13

    Fingerprint

    Experiments

    All Science Journal Classification (ASJC) codes

    • Software
    • Signal Processing
    • Electrical and Electronic Engineering

    Cite this

    Suo, Y., Dao, M., Tran, T., Srinivas, U., & Monga, V. (2013). Hierarchical sparse modeling using Spike and Slab priors. In 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings (pp. 3103-3107). [6638229] (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). https://doi.org/10.1109/ICASSP.2013.6638229
    Suo, Yuanming ; Dao, Minh ; Tran, Trac ; Srinivas, Umamahesh ; Monga, Vishal. / Hierarchical sparse modeling using Spike and Slab priors. 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. pp. 3103-3107 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).
    @inproceedings{2df75f55722846649fb14d02aea0d003,
    title = "Hierarchical sparse modeling using Spike and Slab priors",
    abstract = "Sparse modeling has demonstrated its superior performances in many applications. Compared to optimization based approaches, Bayesian sparse modeling generally provides a more sparse result with a knowledge of confidence. Using the Spike and Slab priors, we propose the hierarchical sparse models for the scenario of single task and multitask - Hi-BCS and CHi-BCS. We draw the connections of these two methods to their optimization based counterparts and use expectation propagation for inference. The experiment results using synthetic and real data demonstrate that the performance of Hi-BCS and Chi-BCS are comparable or better than their optimization based counterparts.",
    author = "Yuanming Suo and Minh Dao and Trac Tran and Umamahesh Srinivas and Vishal Monga",
    year = "2013",
    month = "10",
    day = "18",
    doi = "10.1109/ICASSP.2013.6638229",
    language = "English (US)",
    isbn = "9781479903566",
    series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",
    pages = "3103--3107",
    booktitle = "2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings",

    }

    Suo, Y, Dao, M, Tran, T, Srinivas, U & Monga, V 2013, Hierarchical sparse modeling using Spike and Slab priors. in 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings., 6638229, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, pp. 3103-3107, 2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013, Vancouver, BC, Canada, 5/26/13. https://doi.org/10.1109/ICASSP.2013.6638229

    Hierarchical sparse modeling using Spike and Slab priors. / Suo, Yuanming; Dao, Minh; Tran, Trac; Srinivas, Umamahesh; Monga, Vishal.

    2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. p. 3103-3107 6638229 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

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

    TY - GEN

    T1 - Hierarchical sparse modeling using Spike and Slab priors

    AU - Suo, Yuanming

    AU - Dao, Minh

    AU - Tran, Trac

    AU - Srinivas, Umamahesh

    AU - Monga, Vishal

    PY - 2013/10/18

    Y1 - 2013/10/18

    N2 - Sparse modeling has demonstrated its superior performances in many applications. Compared to optimization based approaches, Bayesian sparse modeling generally provides a more sparse result with a knowledge of confidence. Using the Spike and Slab priors, we propose the hierarchical sparse models for the scenario of single task and multitask - Hi-BCS and CHi-BCS. We draw the connections of these two methods to their optimization based counterparts and use expectation propagation for inference. The experiment results using synthetic and real data demonstrate that the performance of Hi-BCS and Chi-BCS are comparable or better than their optimization based counterparts.

    AB - Sparse modeling has demonstrated its superior performances in many applications. Compared to optimization based approaches, Bayesian sparse modeling generally provides a more sparse result with a knowledge of confidence. Using the Spike and Slab priors, we propose the hierarchical sparse models for the scenario of single task and multitask - Hi-BCS and CHi-BCS. We draw the connections of these two methods to their optimization based counterparts and use expectation propagation for inference. The experiment results using synthetic and real data demonstrate that the performance of Hi-BCS and Chi-BCS are comparable or better than their optimization based counterparts.

    UR - http://www.scopus.com/inward/record.url?scp=84890502852&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84890502852&partnerID=8YFLogxK

    U2 - 10.1109/ICASSP.2013.6638229

    DO - 10.1109/ICASSP.2013.6638229

    M3 - Conference contribution

    AN - SCOPUS:84890502852

    SN - 9781479903566

    T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

    SP - 3103

    EP - 3107

    BT - 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings

    ER -

    Suo Y, Dao M, Tran T, Srinivas U, Monga V. Hierarchical sparse modeling using Spike and Slab priors. In 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. p. 3103-3107. 6638229. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). https://doi.org/10.1109/ICASSP.2013.6638229