Enhancing Reinforcement Learning with Label-Sensitive Reward for Natural Language Understanding.
Kuo LiaoShuang LiMeng ZhaoLiqun LiuMengge XueZhenyu HuHonglin HanChengguo YinPublished in: ACL (1) (2024)
Keyphrases
- natural language understanding
- reinforcement learning
- text understanding
- semantic analysis
- language understanding
- knowledge representation
- natural language
- function approximation
- semantic representations
- natural language processing
- temporal difference
- dialogue system
- state space
- eligibility traces
- spoken dialog systems
- reward function
- model free
- learning algorithm
- partially observable environments
- markov decision processes
- reinforcement learning algorithms
- multi agent
- machine learning
- database
- optimal policy
- abductive reasoning
- prior knowledge
- semantic representation
- keywords
- learning agent
- average reward
- web pages
- information retrieval