Deep Reinforcement Learning for Inquiry Dialog Policies with Logical Formula Embeddings.
Takuya HiraokaMasaaki TsuchidaYotaro WatanabePublished in: CoRR (2017)
Keyphrases
- reinforcement learning
- optimal policy
- logical formulas
- policy search
- markov decision process
- control policies
- markov decision processes
- function approximation
- reinforcement learning agents
- reward function
- control policy
- markov decision problems
- fitted q iteration
- partially observable markov decision processes
- natural language
- state space
- policy gradient methods
- learning environment
- reinforcement learning algorithms
- scientific inquiry
- vector space
- dimensionality reduction
- knowledge building
- infinite horizon
- manifold learning
- policy iteration
- action space
- model free
- long run
- partial order
- partially observable
- mixed initiative
- temporal difference
- science education
- data mining
- euclidean space
- conversational agent
- low dimensional
- machine learning
- user interface
- knowledge base
- spoken dialog
- e learning