Research map / Reinforcement learning / RL from feedback
How to Fine-tune the Model: Unified Model Shift and Model Bias Policy Optimization
Hai Zhang, Hang Yu, Junqiao Zhao, Di Zhang, Chang Huang, Hongtu Zhou, Xiao Zhang, Chen Ye
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 phase-2 fine-tuning that jointly minimizes model shift and model bias | MBPO · MuJoCo continuous control tasks | maximum average return | Improved |
| Remove phase-2 fine-tuning (ablation) | USB-PO (MBPO variant) · MuJoCo continuous control tasks | sample efficiency | Hurt |
| Fine-tune using only the model-shift term (ablation) | USB-PO · MuJoCo continuous control tasks | sample efficiency | Hurt |
| Fine-tune using only the model-bias term (ablation) | USB-PO · MuJoCo continuous control tasks | average return | Hurt |
| Use adaptive unified fine-tuning rather than limiting update magnitude through MBPO learning-rate changes (ablation) | MBPO versus USB-PO · Walker2d and HalfCheetah | average return | Improved |
| Vary the phase-2 learning rate (ablation) | USB-PO · Walker2d and HalfCheetah | average return | No clear effect |
| Vary the CMLO model-shift threshold from the recommended setting (ablation) | CMLO · Walker2d | average return | Hurt |
| Increase the ensemble size (ablation) | USB-PO · MuJoCo continuous control tasks | performance | Improved |
Closest papers by meaning
- When to Update Your Model: Constrained Model-based Reinforcement Learning (NeurIPS 2022 · similarity 0.91)
- When to Trust Your Model: Model-Based Policy Optimization (NeurIPS 2019 · similarity 0.88)
- Model-based Policy Optimization with Unsupervised Model Adaptation (NeurIPS 2020 · similarity 0.88)
- Improving Model-Based Reinforcement Learning by Converging to Flatter Minima (NeurIPS 2025 · similarity 0.88)
- Differentiable Information Enhanced Model-Based Reinforcement Learning (AAAI 2025 · similarity 0.87)
- Model Selection for Off-policy Evaluation: New Algorithms and Experimental Protocol (NeurIPS 2025 · similarity 0.86)
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.
