Plannable Approximations to MDP Homomorphisms: Equivariance under Actions.
Elise van der PolThomas N. KipfFrans A. OliehoekMax WellingPublished in: CoRR (2020)
Keyphrases
- markov decision processes
- action sets
- initial state
- reward function
- state and action spaces
- state space
- decision theoretic
- decision theoretic planning
- partially observable
- action space
- optimal policy
- finite state
- reinforcement learning
- markov decision process
- state transitions
- neural network
- plan recognition
- planning under uncertainty
- utility function
- situation calculus
- multiple agents
- reasoning about actions
- state action
- partial observability
- human activities