Advanced search
1 file | 5.79 MB
Author
Organization
Abstract
We present a novel model architecture which leverages deep learning tools to perform exact Bayesian inference on sets of high dimensional, complex observations. Our model is provably exchangeable, meaning that the joint distribution over observations is invariant under permutation: this property lies at the heart of Bayesian inference. The model does not require variational approximations to train, and new samples can be generated conditional on previous samples, with cost linear in the size of the conditioning set. The advantages of our architecture are demonstrated on learning tasks that require generalisation from short observed sequences while modelling sequence variability, such as conditional image generation, few-shot learning, and anomaly detection.

Downloads

  • 1802.07535.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 5.79 MB

Citation

Please use this url to cite or link to this publication:

Chicago
Korshunova, Iryna, Jonas Degrave, Ferenc Huszar, Yarin Gal, Arthur Gretton, and Joni Dambre. 2018. “BRUNO : a Deep Recurrent Model for Exchangeable Data.” In ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018) , 31:1–9.
APA
Korshunova, I., Degrave, J., Huszar, F., Gal, Y., Gretton, A., & Dambre, J. (2018). BRUNO : a deep recurrent model for exchangeable data. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018) (Vol. 31, pp. 1–9). Presented at the 32nd Conference on Neural Information Processing Systems (NIPS) .
Vancouver
1.
Korshunova I, Degrave J, Huszar F, Gal Y, Gretton A, Dambre J. BRUNO : a deep recurrent model for exchangeable data. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018) . 2018. p. 1–9.
MLA
Korshunova, Iryna et al. “BRUNO : a Deep Recurrent Model for Exchangeable Data.” ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018) . Vol. 31. 2018. 1–9. Print.
@inproceedings{8584686,
  abstract     = {We present a novel model architecture which leverages deep learning tools to perform exact Bayesian inference on sets of high dimensional, complex observations. Our model is provably exchangeable, meaning that the joint distribution over observations is invariant under permutation: this property lies at the heart of Bayesian inference. The model does not require variational approximations to train, and new samples can be generated conditional on previous samples, with cost linear in the size of the conditioning set. The advantages of our architecture are demonstrated on learning tasks that require generalisation from short observed sequences while modelling sequence variability, such as conditional image generation, few-shot learning, and anomaly detection.},
  author       = {Korshunova, Iryna and Degrave, Jonas and Huszar, Ferenc and Gal, Yarin and Gretton, Arthur and Dambre, Joni},
  booktitle    = {ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018) },
  issn         = {1049-5258 },
  language     = {eng},
  location     = {Montreal, CANADA },
  pages        = {1--9},
  title        = {BRUNO : a deep recurrent model for exchangeable data},
  volume       = {31},
  year         = {2018},
}