A strongly quasiconvex PAC-Bayesian bound

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

We propose a new PAC-Bayesian bound and a way of constructing a hypothesis space, so that the bound is convex in the posterior distribution and also convex in a trade-off parameter between empirical performance of the posterior distribution and its complexity. The complexity is measured by the Kullback-Leibler divergence to a prior. We derive an alternating procedure for minimizing the bound. We show that the bound can be rewritten as a one-dimensional function of the trade-off parameter and provide sufficient conditions under which the function has a single global minimum. When the conditions are satisfied the alternating minimization is guaranteed to converge to the global minimum of the bound. We provide experimental results demonstrating that rigorous minimization of the bound is competitive with cross-validation in tuning the trade-off between complexity and empirical performance. In all our experiments the trade-off turned to be quasiconvex even when the sufficient conditions were violated.
Original languageEnglish
Title of host publicationProceedings of International Conference on Algorithmic Learning Theory, 15-17 October 2017, Kyoto University, Kyoto, Japan
EditorsSteve Hanneke, Lev Reyzin
PublisherProceedings of Machine Learning Research
Publication date2017
Pages466-492
Publication statusPublished - 2017
EventThe 28th International Conference on Algorithmic Learning Theory (ALT) - Kyoto, Japan
Duration: 15 Oct 201717 Oct 2017
http://www.comp.nus.edu.sg/~fstephan/alt/alt2017/

Conference

ConferenceThe 28th International Conference on Algorithmic Learning Theory (ALT)
LandJapan
ByKyoto
Periode15/10/201717/10/2017
Internetadresse
SeriesProceedings of Machine Learning Research
Volume76
ISSN1938-7228

ID: 197764786