Towards Automatic Evaluation of Dialog Systems: A Model-Free Off-Policy Evaluation Approach.
Haoming JiangBo DaiMengjiao YangWei WeiTuo ZhaoPublished in: CoRR (2021)
Keyphrases
- automatic evaluation
- policy evaluation
- model free
- dialog systems
- natural language generation
- reinforcement learning
- natural language
- temporal difference
- policy iteration
- reinforcement learning algorithms
- function approximation
- human judgments
- learning algorithm
- multi agent
- information extraction
- least squares
- knowledge base
- e learning