On the Hidden Biases of Policy Mirror Ascent in Continuous Action Spaces.
Amrit Singh BediSouradip ChakrabortyAnjaly ParayilBrian M. SadlerPratap TokekarAlec KoppelPublished in: ICML (2022)
Keyphrases
- action space
- state space
- markov decision processes
- state and action spaces
- real valued
- control policies
- reinforcement learning
- continuous state
- action selection
- continuous state spaces
- stochastic processes
- state action
- continuous action
- markov decision problems
- markov decision process
- skill learning
- function approximators
- policy search
- reinforcement learning problems
- single agent
- optimal policy
- markov chain
- decision making
- dynamic programming
- search space