Q-Pensieve: Boosting Sample Efficiency of Multi-Objective RL Through Memory Sharing of Q-Snapshots.
Wei HungBo-Kai HuangPing-Chun HsiehXi LiuPublished in: ICLR (2023)
Keyphrases
- multi objective
- evolutionary algorithm
- reinforcement learning
- multi objective optimization
- memory usage
- optimization algorithm
- multiple objectives
- multiobjective optimization
- learning algorithm
- feature selection
- trade off
- particle swarm optimization
- pareto optimal
- ensemble methods
- memory requirements
- information sharing
- conflicting objectives
- model free
- neural network
- multi objective optimization problems
- knowledge sharing
- markov decision processes
- data structure
- genetic algorithm
- machine learning