Entropy-SGD optimizes the prior of a PAC-Bayes bound: Generalization properties of Entropy-SGD and data-dependent priors.
Gintare Karolina DziugaiteDaniel M. RoyPublished in: ICML (2018)
Keyphrases
- pac bayes
- data dependent
- risk bounds
- generalization bounds
- rademacher complexity
- information theory
- empirical risk minimization
- learning algorithm
- vc dimension
- hash functions
- bayesian framework
- reproducing kernel hilbert space
- stochastic processes
- statistical learning theory
- ranking algorithm
- theoretical framework
- upper bound