Skip to yearly menu bar Skip to main content


Poster

PAC-Bayesian Theory Meets Bayesian Inference

Pascal Germain · Francis Bach · Alexandre Lacoste · Simon Lacoste-Julien

Area 5+6+7+8 #29

Keywords: [ Learning Theory ] [ (Other) Bayesian Inference ]


Abstract:

We exhibit a strong link between frequentist PAC-Bayesian bounds and the Bayesian marginal likelihood. That is, for the negative log-likelihood loss function, we show that the minimization of PAC-Bayesian generalization bounds maximizes the Bayesian marginal likelihood. This provides an alternative explanation to the Bayesian Occam's razor criteria, under the assumption that the data is generated by an i.i.d. distribution. Moreover, as the negative log-likelihood is an unbounded loss function, we motivate and propose a PAC-Bayesian theorem tailored for the sub-gamma loss family, and we show that our approach is sound on classical Bayesian linear regression tasks.

Live content is unavailable. Log in and register to view live content