Sample Efficient Deep Reinforcement Learning for Dialogue Systems With Large Action Spaces.
Gellért WeiszPawel BudzianowskiPei-Hao SuMilica GasicPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2018)
Keyphrases
- reinforcement learning
- action space
- dialogue system
- continuous state
- markov decision processes
- state space
- state and action spaces
- optimal policy
- spoken dialogue systems
- machine learning
- transfer learning
- dynamic programming
- real valued
- single agent
- domain independent
- function approximation
- action selection
- domain knowledge
- hidden markov models
- state action
- reinforcement learning methods
- natural language
- continuous state spaces