Plan-Space State Embeddings for Improved Reinforcement Learning.
Max PfluegerGaurav S. SukhatmePublished in: CoRR (2020)
Keyphrases
- video sequences
- space time
- reinforcement learning
- state space
- action space
- low dimensional
- vector space
- search space
- markov decision processes
- plan recognition
- state abstraction
- hidden state
- markov decision process
- initial state
- higher dimensional
- dimensionality reduction
- machine learning
- euclidean space
- function approximation
- model free
- dynamical systems
- optimal policy
- plan execution
- data sets