Action discovery for reinforcement learning.
Bikramjit BanerjeeLandon KraemerPublished in: AAMAS (2010)
Keyphrases
- reinforcement learning
- action selection
- action space
- partially observable domains
- reward shaping
- function approximation
- state action
- markov decision processes
- transition model
- reinforcement learning algorithms
- temporal difference
- knowledge discovery
- state space
- optimal policy
- model free
- machine learning
- agent receives
- robotic control
- temporal difference learning
- reinforcement learning methods
- agent learns
- human actions
- multi agent
- discovery process
- partially observable
- optimal control
- pattern discovery
- stochastic approximation
- learning problems
- supervised learning
- mobile robot
- data mining
- neural network