Research map / Label-efficient and robust learning / Domain adaptation and generalization
Transfer Value Iteration Networks
Junyi Shen, Hankz Hankui Zhuo, Jin Xu, Bin Zhong, Sinno Pan
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 |
|---|---|---|---|
| Transfer reward and transition knowledge from NEWS-15, mapping target features and adapting shared actions | VIN policy for Moore · Moore-15, 1k target data | %Opt: 77.3 -> 90.1 | Improved |
| Use TVIN instead of heuristic transferred initialization | VIN_i on Moore · Moore-15, 1k target data | %Opt: 88.6 -> 90.1 | Improved |
| Transfer policy knowledge from NEWS-15 | VIN policy for Moore · Moore-28, 10k target data | %Suc: 68.0 -> 91.8 | Improved |
| Set iteration count K = 20 rather than K = 10 (ablation) | TVIN · Moore-15, NEWS-9 source | %Opt: 78.0 -> 81.6 | Improved |
| Set iteration count K = 30 rather than K = 20 (ablation) | TVIN · Moore-15, NEWS-9 source | %Opt: 81.6 -> 80.1 | Hurt |
| Set kernel size F = 5 rather than F = 3 (ablation) | TVIN · Moore-15, NEWS-9 source | %Opt: 80.1 -> 88.3 | Improved |
| Set kernel size F = 7 rather than F = 5 (ablation) | TVIN · Moore-15, NEWS-9 source | %Opt: 88.3 -> 85.3 | Hurt |
| Transfer more common actions (ablation) | TVIN · Moore-15, 1k data; NEWS-9 source | %Opt: 82.0 -> 88.3 | Improved |
| Transfer policy knowledge from NEWS-9 rather than train VIN from scratch | VIN policy for Moore · Moore-15, 1k target data | %Suc: 81.7 -> 91.0 | Improved |
Closest papers by meaning
- Value Iteration Networks (NeurIPS 2016 · similarity 0.89)
- Learning the Target Network in Function Space (ICML 2024 · similarity 0.86)
- Model-Based Transfer Learning for Contextual Reinforcement Learning (NeurIPS 2024 · similarity 0.85)
- Highway Value Iteration Networks (ICML 2024 · similarity 0.85)
- Scaling Value Iteration Networks to 5000 Layers for Extreme Long-Term Planning (ICML 2025 · similarity 0.85)
- A Transfer Approach Using Graph Neural Networks in Deep Reinforcement Learning (AAAI 2024 · similarity 0.84)
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.
