Research map / Reinforcement learning / Visual and representation RL
Wavelet Policy: Lifting Scheme for Policy Learning in Long-Horizon Tasks
Hao Huang, Shuaihang Yuan, Geeta Chandra Raju Bethala, Congcong Wen, Anthony Tzes, Yi Fang
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 |
|---|---|---|---|
| use causal convolution instead of non-causal convolution (ablation) | Wavelet Policy · Franka Kitchen | success T1: 0.494 -> 0.953 | Improved |
| use causal convolution instead of non-causal convolution (ablation) | Wavelet Policy · Franka Kitchen | success T4: 0.033 -> 0.339 | Improved |
| use causal convolution instead of non-causal convolution (ablation) | Wavelet Policy · Franka Kitchen | success T5: 0.002 -> 0.041 | Improved |
| use learnable lifting networks instead of fixed Haar wavelet (ablation) | Wavelet Policy · Franka Kitchen | success T1: 0.884 -> 0.953 | Improved |
| use learnable lifting networks instead of fixed Haar wavelet (ablation) | Wavelet Policy · Franka Kitchen | success T4: 0.265 -> 0.339 | Improved |
| use learnable lifting networks instead of fixed DB2 wavelet (ablation) | Wavelet Policy · Franka Kitchen | success T4: 0.225 -> 0.339 | Improved |
| use learnable lifting networks instead of fixed DB2 wavelet (ablation) | Wavelet Policy · Franka Kitchen | success T5: 0.029 -> 0.041 | Improved |
| replace MinGPT in BeT with wavelet network | BeT · CARLA self-driving | success rate: 0.832 -> 0.847 | Improved |
| replace MinGPT in BeT with wavelet network | BeT · Franka Kitchen | success T4: 0.275 -> 0.339 | Improved |
| replace MinGPT in BeT with wavelet network | BeT · Franka Kitchen | success T2: 0.773 -> 0.775 | No clear effect |
| replace Transformer decoders in Diffusion Policy with wavelet network | Diffusion Policy · Robomimic Transport-mh | success rate: 0.440 -> 0.497 | Improved |
| replace Transformer decoders in Diffusion Policy with wavelet network | Diffusion Policy · Push-T | success rate: 0.942 -> 0.958 | Improved |
Closest papers by meaning
- STRAP: Robot Sub-Trajectory Retrieval for Augmented Policy Learning (ICLR 2025 · similarity 0.86)
- Prediction with Action: Visual Policy Learning via Joint Denoising Process (NeurIPS 2024 · similarity 0.85)
- Two-Steps Diffusion Policy for Robotic Manipulation via Genetic Denoising (NeurIPS 2025 · similarity 0.85)
- Instant Policy: In-Context Imitation Learning via Graph Diffusion (ICLR 2025 · similarity 0.85)
- Self-Supervised Keyframe Discovery for Horizon-Invariant Behavior Cloning (NeurIPS 2026 · similarity 0.85)
- Scaling Proprioceptive-Visual Learning with Heterogeneous Pre-trained Transformers (NeurIPS 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.
