| title | LIMMT: Less is More for Motion Tracking |
|---|---|
| authors | Yu Guan, Zekun Qi, Chenghuai Lin, Xuchuan Chen, Dairu Liu, Wenyao Zhang, Jilong Wang, Xinqiang Yu, He Wang, Li Yi |
| date | 2026-06-05 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2606.06953 |
| affiliation | galbot |
| upvotes | 14 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | We argue that high-quality motion data can steer tracking policies toward better optimization trajectories early in training. In this work, we introduce LIMMT (Less Is More for Motion Tracking). To our knowledge, this is the first data-centric study for physics-based humanoid motion tracking. We go beyond simply removing low-quality and erroneous clips, but define motion data quality through three dimensions: physics feasibility, diversity, and complexity. We show that even training with under 3% of AMASS yields better tracking performance than training with the full dataset. We further conduct data cleaning on the estimated web-sourced mocap data. Extensive experiments and analyses validate the effectiveness of our framework. |
cite
@misc{guan2026limmt,
title={LIMMT: Less is More for Motion Tracking},
author={Yu Guan and Zekun Qi and Chenghuai Lin and Xuchuan Chen and Dairu Liu and Wenyao Zhang and Jilong Wang and Xinqiang Yu and He Wang and Li Yi},
year={2026},
eprint={2606.06953},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2606.06953}
}