AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

MADE: Exploration via Maximizing Deviation from Explored Regions

Tianjun Zhang, Paria Rashidinejad, Jiantao Jiao, Yuandong Tian, Joseph E. Gonzalez, Stuart Russell

NeurIPS 2021 · Reinforcement learning › Visual and representation RL · approach: Novelty and impact-driven exploration · 11 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
Add MADE intrinsic reward to IMPALAIMPALA · MiniGrid procedurally generated taskssamples to maximum reward: 2-5 times fewer samplesImproved
Add MADE intrinsic reward to RADRAD · DeepMind Control Suite taskssample efficiency: 2.6 times over RAD+RE3, 3.3 times over RAD+RND, 3.8 times over RADImproved
Add MADE intrinsic reward to RADRAD · Pendulum-Swingupreturn: around 800 in only 30K stepsImproved
Add MADE intrinsic reward to DreamerDreamer · Hopper-Hopreturn: more than 100 higher return than RE3 and 250 higher return than DreamerImproved
Replace the Hoeffding bonus with the MADE bonusValue iteration, PPO with a model, and Q-learning · Bidirectional LockperformanceImproved
Replace the Bernstein bonus with the MADE bonusValue iteration, PPO with a model, and Q-learning · Bidirectional LockperformanceMixed
Add MADE regularization to policy gradientPolicy gradient · Chain MDPconvergence rateImproved
Vary the recent-data buffer length (ablation)MADE · DeepMind Control Suite tasksperformanceMixed

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