Timezone: »

PAC-Bayes under potentially heavy tails
Matthew Holland

Tue Dec 10 10:45 AM -- 12:45 PM (PST) @ East Exhibition Hall B + C #227

We derive PAC-Bayesian learning guarantees for heavy-tailed losses, and obtain a novel optimal Gibbs posterior which enjoys finite-sample excess risk bounds at logarithmic confidence. Our core technique itself makes use of PAC-Bayesian inequalities in order to derive a robust risk estimator, which by design is easy to compute. In particular, only assuming that the first three moments of the loss distribution are bounded, the learning algorithm derived from this estimator achieves nearly sub-Gaussian statistical error, up to the quality of the prior.

Author Information

Matthew Holland (Osaka University)