DrM: Mastering Visual Reinforcement Learning through Dormant Ratio Minimization.
Guowei XuRuijie ZhengYongyuan LiangXiyao WangZhecheng YuanTianying JiYu LuoXiaoyu LiuJiaxin YuanPu HuaShuzhen LiYanjie ZeHal Daumé IIIFurong HuangHuazhe XuPublished in: ICLR (2024)
Keyphrases
- reinforcement learning
- visual information
- dynamic programming
- state space
- databases
- markov decision processes
- function approximation
- machine learning
- multi agent
- low level
- model free
- optimal policy
- digital rights management
- visual representation
- reinforcement learning algorithms
- visual data
- genetic algorithm
- neural network