AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

Style-Agnostic Reinforcement Learning

Juyong Lee, Seokjun Ahn, Jaesik Park

ECCV 2022 · Reinforcement learning › Imitation learning · approach: Adversarial imitation learning · 4 citations (OpenAlex)

Read the paper

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 the adversarial style perturbation loss (ablation)SAR · Procgen Starpilottest score: 27.44 -> 28.92Improved
add value similarity regularization G_V (ablation)SAR · Procgen Starpilotscore std: ±7.79 -> ±4.60Improved
add value similarity regularization G_V (ablation)SAR · Procgen Starpilotmean test score: 29.28 -> 28.92No clear effect
add adversarial style perturbation on top of style mixing (ablation)MixStyle · Procgen Starpilottest score: 26.81 -> 28.92Improved
add random translation augmentationSAR · Procgen Starpilottest score: 28.92 -> 30.76Improved
add random color cutout augmentationSAR · Procgen Starpilottest score: 28.92 -> 33.72Improved
replace PPO with SAR (style mixing plus adversarial perturbation)PPO · Procgen (8 tasks)average rank: 4.9 -> 2.9Improved
replace SAC with SARSAC · Distracting Control Suite (4 tasks)average rank: 2.125 -> 1.25Improved
replace SAC with SAR under zero noiseSAC · Distracting Control Suite walker:walk (zero noise)return: 373 -> 325Hurt
delay the adversarial loss until after 10M stepsSAR · Procgen (8 tasks, 50M steps)test score: 30.7±5.3 vs 27.6±7.9Mixed
delay the adversarial loss until after 25M stepsSAR · Procgen (8 tasks, 50M steps)test score: 27.6±7.9 -> 34.0±13.9Mixed
delay the adversarial loss until after 300k stepsSAR · Distracting Control Suite (zero noise and distraction)return: walker zero noise 325 -> 420; moderate 139 -> 113Mixed

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