Excluding the Irrelevant: Focusing Reinforcement Learning through Continuous Action Masking.
Roland StolzHanna KrasowskiJakob ThummMichael EichelbeckPhilipp GassertMatthias AlthoffPublished in: CoRR (2024)
Keyphrases
- continuous action
- policy search
- reinforcement learning
- continuous state
- action space
- continuous state and action spaces
- partially observable markov decision processes
- reinforcement learning algorithms
- state space
- state dependent
- function approximation
- dynamic programming
- markov decision processes
- finite state
- function approximators
- optimal policy
- machine learning
- markov decision problems
- reinforcement learning methods
- learning algorithm
- model free
- planning problems
- multi agent
- temporal difference
- state action
- robot navigation
- dynamic environments
- linear programming