Plannable Approximations to MDP Homomorphisms: Equivariance under Actions.
Elise van der PolThomas KipfFrans A. OliehoekMax WellingPublished in: AAMAS (2020)
Keyphrases
- markov decision processes
- reward function
- action sets
- initial state
- partially observable
- decision theoretic
- state and action spaces
- decision theoretic planning
- state space
- reinforcement learning
- action space
- markov decision process
- plan recognition
- state transitions
- optimal policy
- action selection
- reasoning about actions
- probability distribution
- machine learning
- utility function
- linear programming
- rough sets
- multi agent