Research map / Vision-language and multimodal / Robot action and navigation
DiffVL: Scaling Up Soft Body Manipulation using Vision-Language Driven Differentiable Physics
Zhiao Huang, Feng Chen, Yewen Pu, Chunru Lin, Hao Su, Chuang Gan
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 language-guided actuator sampling constraints (ablation) | DiffVL · 20 short-horizon task stages, Total | success rate: 0.14 -> 0.88 | Improved |
| Add language-guided actuator sampling constraints (ablation) | DiffVL · 20 short-horizon task stages, Total | 3D IoU: 0.456 -> 0.557 | Improved |
| Use language-guided optimization constraints instead of shape-distance-only optimization (ablation) | DiffVL · 20 short-horizon task stages, Total | success rate: 0.48 -> 0.88 | Improved |
| Use language-guided optimization constraints instead of shape-distance-only optimization (ablation) | DiffVL · 20 short-horizon task stages, Total | 3D IoU: 0.529 -> 0.557 | Improved |
| Retain intermediate vision keyframes rather than optimize directly for the final goal (ablation) | DiffVL multistage solver · Multistage tasks | task completion | Improved |
| Keep actuator sampling at intermediate stages to switch actuators or manipulated objects (ablation) | DiffVL multistage solver · Multistage tasks | task completion | Improved |
| Retain vision subgoals within each stage (ablation) | DiffVL multistage solver · Multistage tasks | task completion | Improved |
Closest papers by meaning
- DiffSkill: Skill Abstraction from Differentiable Physics for Deformable Object Manipulations with Tools (ICLR 2022 · similarity 0.90)
- PlasticineLab: A Soft-Body Manipulation Benchmark with Differentiable Physics (ICLR 2021 · similarity 0.88)
- DIFFTACTILE: A Physics-based Differentiable Tactile Simulator for Contact-rich Robotic Manipulation (ICLR 2024 · similarity 0.86)
- CoDA: Coordinated Diffusion Noise Optimization for Whole-Body Manipulation of Articulated Objects (NeurIPS 2025 · similarity 0.86)
- ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulation (NeurIPS 2025 · similarity 0.86)
- PAC Bench: Do Foundation Models Understand Prerequisites for Executing Manipulation Policies? (NeurIPS 2025 · 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.
