AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Does Self-supervised Learning Really Improve Reinforcement Learning from Pixels?

Xiang Li, Jinghuan Shang, Srijan Das, Michael S Ryoo

NeurIPS 2022 · Reinforcement learning › Visual and representation RL · approach: Robust visual encoders for RL · 9 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
Balance the two constituent prediction losses instead of averaging their concatenated outputs (ablation)two-loss RL context prediction methods · DMControlscore: Extract-AR: ↑75.7, ↓-10.6, ↑7.1, ↑98.5, ↑41.8, ↑75.4; Guess-AF: ↑588.6, ↑395.4, ↑190.3, ↓-37.1, ↑79.1, ↑12.3; Predict-FRMixed
Use a ResNet encoder backbone instead of the baseline CNN (ablation)RL methods evaluated in the backbone ablation · DMControlscoreImproved
Tune random-crop augmentation strength rather than using an arbitrary magnitude (ablation)SAC-based pixel RL · DMControl cheetah, runscoreMixed
Add a self-supervised loss alongside image augmentationSAC-Aug(100) · DMControlscoreMixed
Use SimSiam or ShuffleCLS self-supervisionSAC with image augmentation · DMControlscoreHurt
Search combinations of multiple SSL losses and augmentation magnitudesSAC · DMControlrelative scoreMixed
Add random-crop augmentation to Efficient Rainbow (ablation)Efficient Rainbow · Atarirelative scoreHurt
Use DrQ-style image augmentation with Efficient Rainbow (ablation)Efficient Rainbow · Atarirelative scoreHurt
Add autoencoder reconstruction lossEfficient Rainbow · Atarirelative scoreImproved
Use image augmentation (ablation)SAC · uArm reacherscoreImproved
Remove image augmentation (ablation)SAC · uArm reacherscoreHurt
Balance losses when manually combining two context-prediction objectives (ablation)Extract-AR, Guess-AF, and Predict-FR · DMControlscoreMixed

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