Research map / Reinforcement learning / Visual and representation RL
Training Transition Policies via Distribution Matching for Complex Tasks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 a transition policy to immediate switching (ablation) | pre-trained policies · Repetitive picking up | success count: 1.49 $\pm$ 0.59 -> 3.55 $\pm$ 1.64 | Helped |
| add a transition policy to immediate switching (ablation) | pre-trained policies · Repetitive catching | success count: 4.58 $\pm$ 1.10 -> 4.75 $\pm$ 0.87 | Helped |
| add a transition policy to immediate switching (ablation) | pre-trained policies · Serve | success count: 0.05 $\pm$ 0.22 -> 0.99 $\pm$ 0.81 | Helped |
| add a transition policy to immediate switching (ablation) | pre-trained policies · Patrol | success count: 1.26 $\pm$ 0.44 -> 3.11 $\pm$ 1.58 | Helped |
| add a transition policy to immediate switching (ablation) | pre-trained policies · Hurdle | success count: 3.19 $\pm$ 1.61 -> 3.43 $\pm$ 1.60 | Helped |
| add a transition policy to immediate switching (ablation) | pre-trained policies · Obstacle Course | success count: 0.41 $\pm$ 0.52 -> 2.44 $\pm$ 1.63 | Helped |
| add Q-network switch timing to the transition policy (ablation) | hierarchical policies · Repetitive picking up | success count: 3.55 $\pm$ 1.64 -> 4.77 $\pm$ 0.91 | Helped |
| add Q-network switch timing to the transition policy (ablation) | hierarchical policies · Repetitive catching | success count: 4.75 $\pm$ 0.87 -> 4.76 $\pm$ 0.82 | No clear effect |
| add Q-network switch timing to the transition policy (ablation) | hierarchical policies · Serve | success count: 0.99 $\pm$ 0.81 -> 1.00 $\pm$ 0.00 | No clear effect |
| add Q-network switch timing to the transition policy (ablation) | hierarchical policies · Patrol | success count: 3.11 $\pm$ 1.58 -> 3.97 $\pm$ 1.38 | Helped |
| add Q-network switch timing to the transition policy (ablation) | hierarchical policies · Hurdle | success count: 3.43 $\pm$ 1.60 -> 4.84 $\pm$ 0.60 | Helped |
| add Q-network switch timing to the transition policy (ablation) | hierarchical policies · Obstacle Course | success count: 2.44 $\pm$ 1.63 -> 3.72 $\pm$ 1.51 | Helped |
Closest papers by meaning
- Flattening Hierarchies with Policy Bootstrapping (NeurIPS 2025 · similarity 0.88)
- Efficient Multi-task Reinforcement Learning with Cross-Task Policy Guidance (NeurIPS 2024 · similarity 0.87)
- Flow Matching for Offline Reinforcement Learning with Discrete Actions (NeurIPS 2026 · similarity 0.87)
- Foundation Policies with Hilbert Representations (ICML 2024 · similarity 0.87)
- Text-Aware Diffusion for Policy Learning (NeurIPS 2024 · similarity 0.87)
- QMP: Q-switch Mixture of Policies for Multi-Task Behavior Sharing (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.