Constructivist Approach to State Space Adaptation in Reinforcement Learning.
Maxime GuériauNicolás CardozoIvana DusparicPublished in: SASO (2019)
Keyphrases
- state space
- reinforcement learning
- reinforcement learning algorithms
- markov decision processes
- optimal policy
- action space
- heuristic search
- function approximation
- markov chain
- learning process
- learning environment
- partially observable
- continuous state spaces
- markov decision process
- state abstraction
- reward function
- control problems
- learning systems
- collaborative learning
- adaptation process
- learning algorithm
- temporal difference
- machine learning
- problem based learning
- state variables
- particle filter
- action selection
- e learning
- dynamic programming
- case based reasoning
- dynamical systems
- planning problems
- learning theory
- partially observable markov decision processes
- learning agent
- learning capabilities
- continuous state
- state transition
- state and action spaces
- constructivist learning
- reward shaping