search

XuGW-Kevin / DrM

DrM, a visual RL algorithm, minimizes the dormant ratio to guide exploration-exploitation trade-offs, achieving significant improvements in sample efficiency and asymptotic performance across diverse domains.
MIT License
61 stars 9 forks source link

atari montezuma's revenge? #1

Closed hlsafin closed 7 months ago

hlsafin commented 8 months ago

Has this been tested on Montezuma's revenge or Pitfall!, visual hard spare problems with discrete action space?

XuGW-Kevin commented 8 months ago

Hi! We're sorry that this work primarily concentrates on continuous control, so it has not been tested on tasks with discrete action space. However, the three key mechanisms we propose for DrM could well be adapted for discrete action tasks on DQN/Efficient Rainbow algorithms with some adjustments.