The paper proves a new fast-rate PAC-Bayes generalization bound controlled by the empirical flatness of the posterior, using shifted Rademacher processes.
PAC-Bayes Mini-tutorial: A Continuous Union Bound
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
When I first encountered PAC-Bayesian concentration inequalities they seemed to me to be rather disconnected from good old-fashioned results like Hoeffding's and Bernstein's inequalities. But, at least for one flavour of the PAC-Bayesian bounds, there is actually a very close relation, and the main innovation is a continuous version of the union bound, along with some ingenious applications. Here's the gist of what's going on, presented from a machine learning perspective.
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fast-rate PAC-Bayes Generalization Bounds via Shifted Rademacher Processes
The paper proves a new fast-rate PAC-Bayes generalization bound controlled by the empirical flatness of the posterior, using shifted Rademacher processes.