Discount Factor as a Regularizer in Reinforcement Learning.
Ron AmitRon MeirKamil CiosekPublished in: ICML (2020)
Keyphrases
- discount factor
- reinforcement learning
- markov decision processes
- optimal policy
- markov decision problems
- partially observable
- average reward
- state space
- infinite horizon
- reinforcement learning algorithms
- function approximation
- finite state
- decision problems
- dynamic programming
- average cost
- machine learning
- learning algorithm
- action space
- semi supervised
- policy iteration
- temporal difference
- dynamical systems
- model free
- long run
- supervised learning
- markov decision process
- action selection
- sufficient conditions
- learning process