AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Self-Supervised Discovering of Interpretable Features for Reinforcement Learning

Wenjie Shi, Gao Ming Huang, Shiji Song, Zhuoyuan Wang, Tingyu Lin, Cheng Wu

TPAMI 2020 · Reinforcement learning › Visual and representation RL · approach: Robust visual encoders for RL · 62 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 SSINet masks instead of perturbation-based maskspolicy trained on mask pixels · Enduroaverage return: 1741.10 -> 2755.42Improved
use SSINet masks instead of gradient-based maskspolicy trained on mask pixels · Enduroaverage return: 819.92 -> 2755.42Improved
use SSINet masks instead of perturbation-based maskspolicy trained on mask pixels · Seaquestaverage return: 1830.00 -> 2356.67Improved
use SSINet masks instead of perturbation-based maskspolicy trained on mask pixels · SpaceInvadersaverage return: 562.92 -> 740.45Improved
use SSINet masks instead of gradient-based maskspolicy trained on mask pixels · Seaquestaverage return: 836.00 -> 2356.67Improved
use SSINet masks instead of gradient-based maskspolicy trained on mask pixels · SpaceInvadersaverage return: 297.73 -> 740.45Improved
overlay states with SSINet attention (mask policy vs expert policy)Lane-following agent · Duckietown (ten maps)average returnImproved
use product of FOR and BER instead of FOR alone as predictor of returnmask metrics · Duckietown emptyR^2 with average return: 0.525 -> 0.655Improved
use AUC instead of F1 as predictor of returnmask metrics · Duckietown U-turnR^2 with average return: 0.148 -> 0.812Improved
use PPO instead of SAC/TD3 as training algorithmLane-following agent · Duckietown (all maps)average returnImproved
use more complex actor architectureLane-following agent · Duckietownaverage returnNo clear effect
increase regularization scale alpha (ablation)SSINet · Duckietown masksattended regionMixed

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