Research map / Video / Video generation
InsActor: Instruction-driven Physics-based Characters
Jiawei Ren, Mingyuan Zhang, Cunjun Yu, Xiao Ma, Liang Pan, Ziwei Liu
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 hierarchical diffusion planning with low-level skill mapping instead of DReCon tracking | physics-based instruction-driven animation · KIT-ML test | R Precision Top 3: 0.470 -> 0.648 | Improved |
| use hierarchical diffusion planning with low-level skill mapping instead of DReCon tracking | physics-based instruction-driven animation · HumanML3D test | FID: 1.244 -> 0.566 | Improved |
| retain the hierarchical method under physical perturbations | InsActor · KIT-ML test with perturbation | R Precision Top 3: 0.447 -> 0.599 | Improved |
| apply waypoint inpainting while retaining the language condition | InsActor · HumanML3D waypoint heading | Success rate: 0.380 -> 0.907 | Improved |
| remove the language condition (ablation) | InsActor waypoint heading · HumanML3D | R Precision Top 3: 0.388 -> 0.089 | Hurt |
| remove the waypoint condition (ablation) | InsActor waypoint heading · HumanML3D | Success rate: 0.907 -> 0.081 | Hurt |
| add the low-level skill policy to high-level diffusion planning (ablation) | InsActor · KIT-ML test | R Precision Top 3: 0.188 -> 0.648 | Improved |
| add the low-level skill policy to high-level diffusion planning (ablation) | InsActor · KIT-ML test | FID: 1.106 -> 0.786 | Improved |
| add high-level diffusion planning to the low-level policy (ablation) | InsActor · KIT-ML test | R Precision Top 3: 0.460 -> 0.648 | Improved |
| increase the skill-space compactness weight (ablation) | InsActor · KIT-ML | instruction-motion alignment | Hurt |
| decrease the skill-space compactness weight (ablation) | InsActor · KIT-ML | motion plausibility | Hurt |
| smooth the planned motions (ablation) | DReCon tracking of generated plans · HumanML3D test | FID: 1.244 -> 1.271 | Hurt |
Closest papers by meaning
- InterControl: Zero-shot Human Interaction Generation by Controlling Every Joint (NeurIPS 2024 · similarity 0.87)
- AnySkill: Learning Open-Vocabulary Physical Skill for Interactive Agents (CVPR 2024 · similarity 0.87)
- PhysAnimator: Physics-Guided Generative Cartoon Animation (CVPR 2025 · similarity 0.87)
- InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions (CVPR 2025 · similarity 0.87)
- Enhanced Fine-Grained Motion Diffusion for Text-Driven Human Motion Synthesis (AAAI 2024 · similarity 0.86)
- HOIAnimator: Generating Text-prompt Human-object Animations using Novel Perceptive Diffusion Models (CVPR 2024 · similarity 0.86)
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.
