Research map / Reinforcement learning / Offline RL
Dual-Robust Cross-Domain Offline Reinforcement Learning Against Dynamics Shifts
Zhongjian Qiao, Rui Yang, Jiafei Lyu, Xiu Li, Zhongxiang Dai, Zhuoran Yang, Siyang Gao, Shuang Qiu
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 |
|---|---|---|---|
| use DROCO instead of OTDF (second best) | DROCO · D4RL kinematic shift, 16 datasets | total normalized score: 969.8 -> 1105.2 | Improved |
| use DROCO under easy kinematic test perturbation | DROCO · halfcheetah-kinematic-expert | performance degradation: 67.4 -> 54.4 (19.3% drop vs over 50% for IGDF and OTDF) | Improved |
| use DROCO under easy morphology perturbation | DROCO · halfcheetah-kinematic-expert | performance degradation: 42.1% drop vs 62.4% (OTDF) and 78.9% (IGDF) | Improved |
| use DROCO under min-Q perturbation scale 0.2 | DROCO · halfcheetah-kinematic-expert | performance degradation: 37.9% drop vs 73.6% (OTDF) and 84.0% (IGDF) | Improved |
| reduce target data to 10% of hopper-expert | IGDF · hopper-kinematic test perturbations | robustness to perturbations | Hurt |
| set transition threshold delta=5 instead of 30 (ablation) | DROCO · medium-replay datasets | normalized score | Hurt |
| sweep penalty coefficient beta in {0.1,0.5,1.0,1.2} (ablation) | DROCO · kinematic medium datasets | normalized score | Mixed |
| use DROCO instead of BOSA | DROCO · half-e | normalized score: 84.3 -> 67.4 | Hurt |
| use DROCO instead of OTDF | DROCO · hopp-e | normalized score: 97.0 -> 89.3 | Hurt |
| use DROCO instead of IGDF | DROCO · walk-m | normalized score: 51.8 -> 70.8 | Improved |
Closest papers by meaning
- Beyond OOD State Actions: Supported Cross-Domain Offline Reinforcement Learning (AAAI 2024 · similarity 0.91)
- Localized Dynamics-Aware Domain Adaption for Off-Dynamics Offline Reinforcement Learning (NeurIPS 2026 · similarity 0.90)
- Minimax Optimal and Computationally Efficient Algorithms for Distributionally Robust Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.90)
- Contrastive Representation for Data Filtering in Cross-Domain Offline Reinforcement Learning (ICML 2024 · similarity 0.90)
- A Clean Slate for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.89)
- Towards Robust Offline Reinforcement Learning under Diverse Data Corruption (ICLR 2024 · similarity 0.89)
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.
