NIPS Proceedingsβ

Cold-Start Reinforcement Learning with Softmax Policy Gradient

Part of: Advances in Neural Information Processing Systems 30 (NIPS 2017)


[PDF] [BibTeX] [Supplemental] [Reviews]


Conference Event Type: Poster


Policy-gradient approaches to reinforcement learning have two common and undesirable overhead procedures, namely warm-start training and sample variance reduction. In this paper, we describe a reinforcement learning method based on a softmax value function that requires neither of these procedures. Our method combines the advantages of policy-gradient methods with the efficiency and simplicity of maximum-likelihood approaches. We apply this new cold-start reinforcement learning method in training sequence generation models for structured output prediction problems. Empirical evidence validates this method on automatic summarization and image captioning tasks.