Research map / Reinforcement learning / Visual and representation RL
Environment Agnostic Representation for Visual Reinforcement Learning
Hyesong Choi, Hunsang Lee, Seongwon Jeong, Dongbo Min
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add feature separation loss L_sep to SAC (ablation) | SAC · DMControl randomized colors, Walker walk | episode return: 412 -> 799 | Improved |
| add all three objectives instead of L_sep + L_recon only (ablation) | EAR · DMControl randomized colors, Walker walk | episode return: 891 -> 922 | Improved |
| add episode-aware state shift loss L_ss (ablation) | EAR · DMControl randomized colors | episode return std: Walker walk std 134 -> 37 | Improved |
| add reconstruction loss L_recon alongside L_sep (ablation) | EAR · DMControl randomized colors | episode return | No clear effect |
| apply ESS constraint to factorized EAF instead of non-factorized feature (ablation) | EAR · DMControl randomized colors | episode return: Cartpole swingup 805 -> 884; Cheetah run 389 -> 368 | Mixed |
| use environment-agnostic feature learning instead of test-time adaptation (PAD) | EAR · DMControl randomized colors, Walker walk | episode return: 468 -> 922 | Improved |
| EAR versus VAI on video backgrounds | EAR · DMControl video backgrounds | episode return: improves 7 of 8 tasks; Finger spin 764 vs 717 | Mixed |
| EAR on DrawerClose texture environments | VAI · DrawerWorld DrawerClose, Fabric | success rate: 74 -> 90 | Improved |
| EAR versus VAI on DrawerOpen Fabric | VAI · DrawerWorld DrawerOpen, Fabric | success rate: 99 -> 93 | Hurt |
| add EAR objectives to SAC encoder | SAC · Model complexity | parameters (M): 2.54 -> 2.55 | No clear effect |
Closest papers by meaning
- A Simple Framework for Generalization in Visual RL under Dynamic Scene Perturbations (NeurIPS 2024 · similarity 0.87)
- The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning (NeurIPS 2024 · similarity 0.86)
- Zero Shot Generalization of Vision-Based RL Without Data Augmentation (ICML 2025 · similarity 0.86)
- What Effects the Generalization in Visual Reinforcement Learning: Policy Consistency with Truncated Return Prediction (AAAI 2024 · similarity 0.86)
- Focus On What Matters: Separated Models For Visual-Based RL Generalization (NeurIPS 2024 · similarity 0.85)
- VIGOR: Zero-Shot Visual Generalization via Latent-Space Consistency in Model-Based Reinforcement Learning (NeurIPS 2026 · similarity 0.85)
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.
