Deep Reinforcement Learning with a Natural Language Action Space.
Ji HeJianshu ChenXiaodong HeJianfeng GaoLihong LiLi DengMari OstendorfPublished in: ACL (1) (2016)
Keyphrases
- action space
- reinforcement learning
- natural language
- state space
- markov decision processes
- state and action spaces
- real valued
- continuous state
- stochastic processes
- action selection
- control policies
- machine learning
- reinforcement learning methods
- continuous state spaces
- function approximation
- state action
- function approximators
- learning algorithm
- supervised learning
- information extraction
- single agent
- markov decision process
- partially observable
- reinforcement learning algorithms
- finite state
- optimal policy
- dynamic programming
- heuristic search
- control policy
- cooperative
- multi agent