AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

Sample-Efficient and Safe Deep Reinforcement Learning via Reset Deep Ensemble Agents

Woojun Kim, Yongjae Shin, Jongeui Park, Youngchul Sung

NeurIPS 2023 · Reinforcement learning › Policy gradient and actor-critic · approach: Safe policy improvement · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use RDE instead of DQNDQN · Atari-100k, IQM, replay ratio 2IQM: 1.203 -> 1.377Improved
Use RDE instead of vanilla reset DQNDQN · Atari-100k, IQM, replay ratio 4IQM: 1.351 -> 1.422Improved
Use RDE instead of DQNDQN · MiniGrid FourRooms, replay ratio 1test return: 0 -> 0.159Improved
Use RDE instead of vanilla resetDQN · MiniGrid GoToDoor-8x8, replay ratio 1test return: 0.659 -> 0.944Improved
Set the DMC reset interval to 4×10^5 rather than use vanilla reset at replay ratio 1SAC · DMC, replay ratio 1normalized IQM: 1.02 -> 1.20Improved
Increase ensemble size from N=2 to N=4 (ablation)RDE+DQN · MiniGridperformanceImproved
Remove the reset mechanism, leaving ensemble learning only (ablation)RDE+DQN · MiniGrid FourRoomsperformanceHurt
Use reset-part (last two layers) instead of reset-all (ablation)RDE+DQN · MiniGrid FourRoom and SimpleCrossing9performanceHurt
Use reset-part (last two layers) instead of reset-all (ablation)RDE+DQN · MiniGrid easier tasksperformanceImproved
Use reset-1 (last layer only) instead of reset-all (ablation)RDE+SAC · DMC acrobot-swingup, fish-swim, humanoid-run, quadruped-runperformanceHurt
Use reset-1 (last layer only) instead of reset-all (ablation)RDE+SAC · DMC cheetah-run, finger-turn_hard, hopper-hop, swimmer-swimmer15, walker-runperformanceNo clear effect
Increase vanilla reset frequency to match RDE's number of reset operations (ablation)SR · MiniGridperformanceHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy