资源分类
》 资源 》论文》PAC CONFIDENCE SETS FOR DEEP NEURAL NET-WORKS VIA CALIBRATED PREDICTION

PAC CONFIDENCE SETS FOR DEEP NEURAL NET-WORKS VIA CALIBRATED PREDICTION

2019-12-30 | |  86 |   56 |   0

Abstract

We propose an algorithm combining calibrated prediction and generalization bounds from learning theory to construct confidence sets for deep neural networks with PAC guarantees—i.e., the confidence set for a given input contains the true label with high probability. We demonstrate how our approach can be used to construct PAC confidence sets on ResNet for ImageNet, and on a dynamics model the half-cheetah reinforcement learning problem.

上一篇：DEEP HOYER :L EARNING SPARSER NEURALN ETWORK WITH DIFFERENTIABLE SCALE -I NVARIANTS PARSITY MEASURES

下一篇：EDITABLE NEURAL NETWORKS

用户评价
全部评价

热门资源

  • Learning to Predi...

    Much of model-based reinforcement learning invo...

  • Stratified Strate...

    In this paper we introduce Stratified Strategy ...

  • The Variational S...

    Unlike traditional images which do not offer in...

  • A Mathematical Mo...

    Direct democracy, where each voter casts one vo...

  • Rating-Boosted La...

    The performance of a recommendation system reli...