Conditionally Gaussian PAC-Bayes

Abstract

Recent studies have empirically investigated different methods to train stochastic neural networks on a classification task by optimising a PAC-Bayesian bound via stochastic gradient descent. Most of these procedures need to replace the misclassification error with a surrogate loss, leading to a mismatch between the optimisation objective and the actual generalisation bound. The present paper proposes a novel training algorithm that optimises the PAC-Bayesian bound, without relying on any surrogate loss. Empirical results show that this approach outperforms currently available PAC-Bayesian training methods.

Cite

Text

Clerico et al. "Conditionally Gaussian PAC-Bayes." Artificial Intelligence and Statistics, 2022.

Markdown

[Clerico et al. "Conditionally Gaussian PAC-Bayes." Artificial Intelligence and Statistics, 2022.](https://mlanthology.org/aistats/2022/clerico2022aistats-conditionally/)

BibTeX

@inproceedings{clerico2022aistats-conditionally,
  title     = {{Conditionally Gaussian PAC-Bayes}},
  author    = {Clerico, Eugenio and Deligiannidis, George and Doucet, Arnaud},
  booktitle = {Artificial Intelligence and Statistics},
  year      = {2022},
  pages     = {2311-2329},
  volume    = {151},
  url       = {https://mlanthology.org/aistats/2022/clerico2022aistats-conditionally/}
}