Research map / Reinforcement learning / Visual and representation RL
Task-aware world model learning with meta weighting via bi-level optimization
Huining Yuan, Hongkun Dou, Xingyu Jiang, Yue Deng
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 TEMPO bi-level meta-weighted world model learning to DreamerV2 | TEMPO · 9 DMC tasks | return / training stability | Improved |
| add TEMPO to DreamerV2 | TEMPO · 6 Atari games | return / training stability | Improved |
| add TEMPO to DreamerV2 | TEMPO · DMC Finger Spin vs Walker Walk | return / training stability | Mixed |
| use naive loss-based sample weights instead of learned meta weights (ablation) | TEMPO · DMC Cartpole Swingup | return | Hurt |
| remove state values from meta weighter input (ablation) | TEMPO · DMC | return / training stability | Improved |
| remove deterministic state from meta weighter input (ablation) | TEMPO · DMC | return / training stability | Improved |
| reduce hidden dimension of meta weighter (ablation) | TEMPO · DMC | return / training stability | No clear effect |
| change activation function of meta weighter (ablation) | TEMPO · DMC | return / training stability | No clear effect |
| decrease number of dense layers in meta weighter (ablation) | TEMPO · DMC | return / training stability | Hurt |
| change normalization method of meta weighter (ablation) | TEMPO · DMC | return / training stability | Hurt |
| use batch normalization in meta weighter (ablation) | TEMPO · DMC | return / training stability | Improved |
| use weight = 0.5*sigmoid+0.75 instead of sigmoid+0.5 (ablation) | TEMPO · DMC | return / training stability | No clear effect |
Closest papers by meaning
- HarmonyDream: Task Harmonization Inside World Models (ICML 2024 · similarity 0.88)
- Task-Aware Virtual Training: Enhancing Generalization in Meta-Reinforcement Learning for Out-of-Distribution Tasks (ICML 2025 · similarity 0.87)
- Meta-DT: Offline Meta-RL as Conditional Sequence Modeling with World Model Disentanglement (NeurIPS 2024 · similarity 0.86)
- PWM: Policy Learning with Multi-Task World Models (ICLR 2025 · similarity 0.86)
- MetaCARD: Meta-Reinforcement Learning with Task Uncertainty Feedback via Decoupled Context-Aware Reward and Dynamics Components (AAAI 2024 · similarity 0.86)
- AMAGO-2: Breaking the Multi-Task Barrier in Meta-Reinforcement Learning with Transformers (NeurIPS 2024 · 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.
