Interactive Reinforcement Learning with Dynamic Reuse of Prior Knowledge from Human and Agent Demonstrations.
Zhaodong WangMatthew E. TaylorPublished in: IJCAI (2019)
Keyphrases
- reinforcement learning
- prior knowledge
- dynamic environments
- multi agent
- action selection
- multi agent systems
- state abstraction
- artificial agents
- learning capabilities
- agent model
- autonomous agents
- learning algorithm
- reinforcement learning agents
- multiagent systems
- partially observable
- machine learning
- training data
- autonomous learning
- cooperative
- learning agent
- partial information
- optimal policy
- intelligent agents
- changing environment
- agent technology
- human behavior
- human subjects
- user interaction
- markov decision process
- virtual agents
- conversational agents
- open environments
- software agents