Research map / Reinforcement learning / Offline RL
SOLAR for Offline MARL: Plateau-Triggered Potential Shaping under World-Model Uncertainty
Jusheng Zhang, Yijia Fan, Ruiqi Chen, Jing Yang, Ziliang Chen, Yongsen Zheng, Yanxi Chen, Jian Wang, Kwok-Yan Lam, Liang Lin, Keze Wang
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 adaptive reward shaping (ARS) module (ablation) | SOLAR · SMAC 5m vs 6m (Medium data) | win rate (%): 72.3 -> 83.45 | Improved |
| use stagnation-triggered shaping instead of static shaping (ablation) | SOLAR · SMAC 5m vs 6m (Medium data) | win rate (%): 78.1 -> 83.45 | Improved |
| combine all three reward components (ct, dt, et) rather than any single one (ablation) | SOLAR · SMAC 5m vs 6m (Medium data) | win rate (%): 75.4 -> 83.45 (best single ct) | Improved |
| add the world model (WM) as training substrate (ablation) | SOLAR · SMAC 5m vs 6m (Medium data) | win rate (%): 64.5 -> 72.3 | Improved |
| use SOLAR instead of offline baselines on low-quality data | SOLAR · SMAC 5m vs 6m (Poor data) | win rate (%): 44.2 -> 58.17 | Improved |
| conservative CQL-style objective with SOLAR training versus offline IQL | SOLAR · SMAC 5m vs 6m (Medium data) OOD Q-values | OOD Q-value overestimation | Improved |
Closest papers by meaning
- Automatic Reward Shaping from Confounded Offline Data (ICML 2025 · similarity 0.90)
- SLOPE: Optimistic Potential Landscape Shaping for Model-based Reinforcement Learning (NeurIPS 2026 · similarity 0.90)
- Action-Dependent Optimality-Preserving Reward Shaping (ICML 2025 · similarity 0.88)
- BAMDP Shaping: a Unified Framework for Intrinsic Motivation and Reward Shaping (ICLR 2025 · similarity 0.87)
- Highly Efficient Self-Adaptive Reward Shaping for Reinforcement Learning (ICLR 2025 · similarity 0.87)
- ORSO: Accelerating Reward Design via Online Reward Selection and Policy Optimization (ICLR 2025 · similarity 0.87)
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.
