Generalization to New Actions in Reinforcement Learning.
Ayush JainAndrew SzotJoseph J. LimPublished in: CoRR (2020)
Keyphrases
- dynamical systems
- partially observable
- reinforcement learning
- state space
- reinforcement learning methods
- perceptual aliasing
- partially observable domains
- partial observability
- action selection
- action space
- markov decision processes
- action sets
- function approximation
- state and action spaces
- reward function
- state action
- optimal policy
- learning algorithm
- learning agent
- temporal difference learning
- initially unknown
- multiagent reinforcement learning
- neural network
- behavioural cloning
- reasoning about actions
- markov decision process
- temporal difference
- model free
- decision theoretic
- situation calculus
- transfer learning
- supervised learning
- dynamic programming
- machine learning