Scaling up deep reinforcement learning for multi-domain dialogue systems.
Heriberto CuayáhuitlSeunghak YuAshley WilliamsonJacob CarsePublished in: IJCNN (2017)
Keyphrases
- multi domain
- spoken dialogue systems
- reinforcement learning
- dialogue system
- dialogue management
- cross domain
- domain specific
- search computing
- natural language
- state space
- mixed initiative
- learning algorithm
- machine learning
- transfer learning
- heterogeneous networks
- learning process
- multi agent
- tutorial dialogue
- data mining
- human users
- domain independent
- optimal policy
- unlabeled data
- semi supervised learning
- knowledge base