Winnie: Task-Oriented Dialog System with Structure-Aware Contrastive Learning and Enhanced Policy Planning.
Kaizhi GaoTianyu WangZhongjing MaSuli ZouPublished in: AAAI (2024)
Keyphrases
- learning process
- learning systems
- learning algorithm
- learning tasks
- search control rules
- macro operators
- action selection
- decision theoretic
- learning problems
- data sets
- heuristic search
- policy gradient
- domain independent
- search space
- optimal policy
- knowledge acquisition
- graphical models
- supervised learning
- state space
- reinforcement learning