Corpus ID: 8969827

Linear Response Methods for Accurate Covariance Estimates from Mean Field Variational Bayes

@inproceedings{Giordano2015LinearRM,
  title={Linear Response Methods for Accurate Covariance Estimates from Mean Field Variational Bayes},
  author={R. Giordano and T. Broderick and Michael I. Jordan},
  booktitle={NIPS},
  year={2015}
}
  • R. Giordano, T. Broderick, Michael I. Jordan
  • Published in NIPS 2015
  • Mathematics, Computer Science
  • Mean field variational Bayes (MFVB) is a popular posterior approximation method due to its fast runtime on large-scale data sets. However, a well known major failing of MFVB is that it underestimates the uncertainty of model variables (sometimes severely) and provides no information about model variable covariance. We generalize linear response methods from statistical physics to deliver accurate uncertainty estimates for model variables—both for individual variables and coherently across… CONTINUE READING
    Beyond Tag-Along Systems
    Single microphone wind noise reduction using techniques of artificial bandwidth extension
    15

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 27 REFERENCES
    Dimerization of human growth hormone by zinc.
    157
    Organizational Demography and Inequality
    46
    Multiuser Detection
    4524
    J. Appi. phys
    • 1962
    Connecting mTORC1 signaling to SREBP-1 activation
    151
    Localized Bumps of Activity Sustained by Inhibition in a Two-Layer Thalamic Network
    26