Interactive Explanations: Diagnosis and Repair of Reinforcement Learning Based Agent Behaviors.
Christian Arzate CruzTakeo IgarashiPublished in: CoG (2021)
Keyphrases
- reinforcement learning
- multi agent
- action selection
- learning agent
- conversational agents
- multi agent systems
- intelligent agents
- function approximation
- learning agents
- autonomous learning
- decision making
- agent behavior
- software agents
- autonomous agents
- selective perception
- multi agent environments
- user interaction
- reinforcement learning agents
- minority game
- medical diagnosis
- real robot
- multiple agents
- multiagent systems
- state abstraction
- embodied conversational agents
- model based diagnosis
- agent interactions
- partially observable
- plan repair
- learning capabilities
- reward function
- state space
- mobile agents
- fault diagnosis
- agent architecture
- machine learning
- repair actions
- dynamic programming
- reinforcement learning algorithms
- sensory inputs
- markov decision processes
- single agent
- agent learns
- exploration strategy
- model free
- cooperative
- reward shaping
- dynamic environments
- optimal policy
- fault detection
- conversational agent
- agent technology
- state transitions
- fault isolation
- social norms
- optimal control
- pedagogical agents
- virtual characters