Speeding-up Reinforcement Learning with Multi-step Actions.
Ralf SchoknechtMartin A. RiedmillerPublished in: ICANN (2002)
Keyphrases
- multi step
- reinforcement learning
- perceptual aliasing
- action selection
- partially observable
- single step
- state action
- action space
- reward function
- lower bounding
- state and action spaces
- function approximation
- reinforcement learning algorithms
- k nearest neighbor
- state space
- partially observable domains
- optimal policy
- learning algorithm
- multiagent reinforcement learning
- temporal difference
- markov decision processes
- learning agent
- knn
- multi agent
- model free
- td learning
- situation calculus
- evaluation function
- data sets
- behavioural cloning
- tumor classification
- average reward
- multiple agents
- machine learning