Action Selection for Composable Modular Deep Reinforcement Learning.
Vaibhav GuptaDaksh AnandPraveen ParuchuriAkshat KumarPublished in: AAMAS (2021)
Keyphrases
- action selection
- reinforcement learning
- basal ganglia
- robot soccer
- temporal difference
- state space
- action space
- function approximation
- human robot
- continuous state and action spaces
- decision making
- modular structure
- reinforcement learning algorithms
- total reward
- markov decision processes
- transfer learning
- action selection mechanism
- model free
- learning algorithm
- machine learning
- optimal control
- optimal policy
- learning process