Research map / Label-efficient and robust learning / Self-supervised and contrastive
R2-Dreamer: Redundancy-Reduced World Models without Decoders or Augmentation
Naoki Morihira, Amal Nahar, Kartik Bharadwaj, Yasuhiro Kato, Akinobu Hayashi, Tatsuya Harada
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 |
|---|---|---|---|
| replace image decoder reconstruction loss with Barlow Twins redundancy-reduction loss | DreamerV3 · DMC Walker Walk | training time (hours): 7.0 -> 4.4 (1.59x speedup) | Improved |
| use redundancy reduction instead of DA-based DreamerPro | DreamerPro · DMC Walker Walk | training time (hours): 10.4 -> 4.4 (2.36x speedup) | Improved |
| add random-shift data augmentation to R2-Dreamer (ablation) | R2-Dreamer · 20 DMC tasks (mean/median) | task score | No clear effect |
| remove data augmentation from DreamerPro (ablation) | DreamerPro · 20 DMC tasks | task score | Hurt |
| halve batch size (B=16 to B=8) (ablation) | R2-Dreamer · 20 DMC tasks | task score: B=16 -> B=8 | No clear effect |
| add random-shift data augmentation to R2-Dreamer (ablation) | R2-Dreamer · DMC-Subtle (5 tasks) | task score | Hurt |
| use R2-Dreamer instead of DreamerV3, DreamerPro, TD-MPC2, DrQ-v2 | R2-Dreamer · DMC-Subtle (5 tasks) | task score | Improved |
| use R2-Dreamer instead of baselines | R2-Dreamer · 20 DMC tasks and Meta-World MT1 (50 tasks) | mean/median score or success rate | No clear effect |
Closest papers by meaning
- The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning (NeurIPS 2024 · similarity 0.86)
- VIGOR: Zero-Shot Visual Generalization via Latent-Space Consistency in Model-Based Reinforcement Learning (NeurIPS 2026 · similarity 0.86)
- Zero-shot World Models via Search in Memory (NeurIPS 2025 · similarity 0.86)
- Learning Latent Dynamic Robust Representations for World Models (ICML 2024 · similarity 0.85)
- Zero Shot Generalization of Vision-Based RL Without Data Augmentation (ICML 2025 · similarity 0.85)
- Focus On What Matters: Separated Models For Visual-Based RL Generalization (NeurIPS 2024 · 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.
