Neural Information Processing Systems http://nips.cc/

Neural Information Processing Systems http://nips.cc/

But could more performance be squeezed out of networks by using different input representations?

Converting an n-dimensional vector to a probability distribution over n objects isacommonly used component inmanymachine learning tasks likemulticlass classification,multilabelclassification,attentionmechanismsetc.

Neural Information Processing Systems http://nips.cc/

Neural Information Processing Systems http://nips.cc/

Yetinreinforcementlearning,duetothenatureofthe Bellman equation, there isanopportunity toalsoexploit temporal regularization based on smoothness in value estimates over trajectories. This paper explores a class of methods for temporal regularization.

We consider deep policy learning with only batched historical trajectories. The main challenge of this problem is that the learner no longer has a simulator or "environment oracle" as in most reinforcement learning settings.

Adversarial training [10, 16, 18], which injects adversarially perturbed dataintotraining data,isapromising approach. Many other heuristics have been developed to make neural networks insensitive against small perturbations on inputs.

Neural Information Processing Systems http://nips.cc/