Bootstrapping a Neural Conversational Agent with Dialogue Self-Play, Crowdsourcing and On-Line Reinforcement Learning.
Pararth ShahDilek Hakkani-TürBing LiuGökhan TürPublished in: NAACL-HLT (3) (2018)
Keyphrases
- conversational agent
- reinforcement learning
- conversational agents
- fitted q iteration
- natural language interfaces
- network architecture
- human communication
- state space
- neural network
- function approximation
- model free
- reinforcement learning algorithms
- learning process
- machine learning
- optimal policy
- game playing
- temporal difference
- associative memory
- sensory inputs
- information extraction
- optimal control
- learning rules
- dynamic programming
- action selection
- multi agent
- image sequences