TY  - RPRT
TI  - Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Data
AU  - Gintare Karolina Dziugaite
AU  - Daniel M. Roy
PY  - 2017
UR  - https://arxiv.org/abs/1703.11008
ID  - 1703.11008
ER  -