Research map / 3D vision / 3D generation
TraceGen: World Modeling in 3D Trace Space Enables Learning from Cross-Embodiment Videos
Seungjae Lee, Yoonkyo Jung, Inkook Chun, Yao-Chih Lee, Zikui Cai, Hongjia Huang, Aayush Talreja, Tan Dao, Yongyuan Liang, Jia-Bin Huang, Furong Huang
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 cross-embodiment pretraining on TraceForge data (ablation) | TraceGen · Real robot, 5-video warm-up | overall success: 25.0% -> 80% | Improved |
| increase warm-up from 5 to 15 robot videos (ablation) | TraceGen · Real robot | overall success: 80% -> 82.5% | No clear effect |
| increase warm-up from 5 to 15 robot videos (ablation) | From scratch · Real robot | overall success: 25.0% -> 30% | No clear effect |
| add pretraining before human-demo fine-tuning (ablation) | TraceGen · Human->robot transfer | success: 0% -> 67.5% | Improved |
| pretrain on full cross-embodiment data instead of Agibot only (ablation) | TraceGen · 5-video warm-up | overall success: 45% -> 70% | Improved |
| pretrain on full cross-embodiment data instead of SSV2 only (ablation) | TraceGen · 5-video warm-up | overall success: 25% -> 70% | Improved |
| predict in 3D trace space instead of pixel-space video generation | TraceGen · Real robot | inference speed: over 50x faster | Improved |
| increase warm-up from 5 to 15 videos (ablation) | TraceGen · Real robot Block | success out of 10: 8/10 -> 6/10 | Hurt |
Later papers that compared with this method
1 later accepted paper reports TraceGen as a baseline in their results tables; 0 of them report a better number on most of the results they share with it.
- World Motion Models: Flexible Sequence Modeling of SE(3) Trajectories (NeurIPS 2026 · better on 0 of 1 shared result)
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
- RoboScape: Physics-informed Embodied World Model (NeurIPS 2025 · similarity 0.90)
- EnerVerse: Envisioning Embodied Future Space for Robotics Manipulation (NeurIPS 2025 · similarity 0.89)
- World Motion Models: Flexible Sequence Modeling of SE(3) Trajectories (NeurIPS 2026 · similarity 0.88)
- VidBot: Learning Generalizable 3D Actions from In-the-Wild 2D Human Videos for Zero-Shot Robotic Manipulation (CVPR 2025 · similarity 0.88)
- TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies (ICLR 2025 · similarity 0.88)
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (ICLR 2025 · similarity 0.88)
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.
