Research map / Video / Video generation
InterSyn: Interleaved Learning for Dynamic Motion Synthesis in the Wild
Yiyi Ma, Yuanzhi Liang, Xiu Li, Chi Zhang, Xuelong Li
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 coordinator (REC) refinement (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | FID: 0.847 -> 0.417 | Improved |
| add coordinator (REC) refinement (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | MM Dist: 5.842 -> 3.707 | Improved |
| add coordinator (REC) refinement (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | R-Precision Top-1: 0.103 -> 0.298 | Improved |
| add relative coordination loss L_rela (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | R-Precision Top-1: 0.283 -> 0.298 | Improved |
| add relative coordination loss L_rela (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | FID: 0.537 -> 0.417 | Improved |
| add smoothness loss L_smooth (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | FID: 0.431 -> 0.417 | No clear effect |
| use two segments (s-i-s) instead of four (s-i-s-i) (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | R-Precision Top-1: 0.242 -> 0.298 | Improved |
| use two segments (s-i-s) instead of five (s-i-s-i-s) (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | R-Precision Top-1: 0.115 -> 0.298 | Improved |
| use two segments (s-i-s) instead of four (s-i-s-i) (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | FID: 0.469 -> 0.417 | Improved |
| random t_i with t_s=0 instead of fixed t_i=0, t_s=60 (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | MModality: 1.416 -> 1.583 | Improved |
| random t_i with t_s=0 instead of fixed t_i=0, t_s=60 (ablation) | InterSyn · Interleaved motion benchmark (HumanML3D+InterHuman) | R-Precision Top-1: 0.301 -> 0.298 | No clear effect |
| replace FreeMotion with InterSyn | FreeMotion · InterHuman test | FID: 6.740 -> 6.332 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- FreeMotion — FreeMotion: A Unified Framework for Number-free Text-to-Motion Synthesis (ECCV 2024) · better on 9 of 11 shared results
- InterGen — InterGen: Diffusion-Based Multi-human Motion Generation Under Complex Interactions (IJCV 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- InterControl: Zero-shot Human Interaction Generation by Controlling Every Joint (NeurIPS 2024 · similarity 0.88)
- InterDyn: Controllable Interactive Dynamics with Video Diffusion Models (CVPR 2025 · similarity 0.87)
- InterDreamer: Zero-Shot Text to 3D Dynamic Human-Object Interaction (NeurIPS 2024 · similarity 0.87)
- InterMask: 3D Human Interaction Generation via Collaborative Masked Modeling (ICLR 2025 · similarity 0.86)
- InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions (CVPR 2025 · similarity 0.86)
- Human Motion Diffusion as a Generative Prior (ICLR 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.
