Can Reinforcement Learning Learn Itself? A Reply to 'Reward is Enough'.
Samuel Allen AlexanderPublished in: SEFM Workshops (2021)
Keyphrases
- reinforcement learning
- learning agent
- function approximation
- learning algorithm
- state space
- function approximators
- markov decision processes
- agent receives
- reinforcement learning agents
- reinforcement learning algorithms
- agent learns
- hierarchical reinforcement learning
- machine learning
- complex domains
- model free
- action selection
- reward function
- learning capabilities
- optimal policy
- average reward
- eligibility traces
- multi agent
- learning problems
- sufficient conditions
- solving problems
- previously learned
- policy gradient
- dynamic programming
- transfer learning
- temporal difference
- reward signal
- data sets