Enhanced Generalization through Prioritization and Diversity in Self-Imitation Reinforcement Learning over Procedural Environments with Sparse Rewards.
Alain AndresDaochen ZhaJavier Del SerPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- function approximation
- markov decision processes
- reinforcement learning algorithms
- multi agent environments
- state space
- machine learning
- sparse data
- learning problems
- learning algorithm
- high dimensional
- optimal policy
- multi agent
- model free
- real world
- partially observable
- transfer learning
- action selection
- temporal difference
- learning process
- sparse representation
- complex domains
- function approximators
- sparse coding
- optimal control
- neural network
- reinforcement learning methods
- hidden state
- supervised learning