/PAC-Bayesian-Theory-Meets-Bayesian-Inference

Code to related to my NIPS 2016 paper

Primary LanguagePythonBSD 2-Clause "Simplified" LicenseBSD-2-Clause

PAC Bayesian Theory Meets Bayesian Inference

This python code has been used to conduct the experiments presented in Section 6 of the following NIPS paper.

Pascal Germain, Francis Bach, Alexandre Lacoste, Simon Lacoste Julien. PAC Bayesian Theory Meets Bayesian Inference. Neural Information Processing Systems (NIPS), 2016.
http://arxiv.org/abs/1605.08636

Content

  • bayesian_regression.py contains the Bayesian linear regression learning algorithm used to produce the experiments. The two next files import this.

  • model_selection_experiment.py contains the code to reproduce Figures 1a and 1b of the paper.

  • bound_values_experiment.py contains the code used to produce Figure 1c of the paper.

Disclaimer

The current version is poorly documented. Feel free to contact me if you have any questions or comments!