Research map / Reinforcement learning / Imitation learning
Solving Compositional Reinforcement Learning Problems via Task Reduction
Yunfei Li, Yilin Wu, Huazhe Xu, Xiaolong Wang, Yi Wu
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 task reduction and imitation to SAC | SAC · Push Unif. | hard-case success rate: 0.816 -> 0.939 | Improved |
| Add task reduction to self-imitation learning | SIL · Push Mixed 30% | average success rate: 0.930 -> 0.942 | Improved |
| Add task reduction and imitation to SAC | SAC · Stack (at most 3 boxes) | hard-case success rate: 0.294 -> 0.846 | Improved |
| Add task reduction and imitation to PPO | PPO · 4-Room maze | average success rate: 0.649 -> 0.695 | Improved |
| Add task reduction and imitation to PPO | PPO · 4-Room maze | hard-case success rate: 0.512 -> 0.427 | Hurt |
| Add task reduction and imitation to SAC | SAC · Push Unif. | average success rate: 0.952 -> 0.938 | Hurt |
Closest papers by meaning
- Robust Subtask Learning for Compositional Generalization (ICML 2023 · similarity 0.88)
- Composing Task-Agnostic Policies with Deep Reinforcement Learning (ICLR 2020 · similarity 0.88)
- DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning (NeurIPS 2025 · similarity 0.86)
- Hybrid Inverse Reinforcement Learning (ICML 2024 · similarity 0.86)
- Compositional Reasoning in Language Models under Reinforcement Learning Post-Training (NeurIPS 2026 · similarity 0.85)
- Imitation Learning from Observation with Automatic Discount Scheduling (ICLR 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.
