The authors consolidated human-object interaction datasets and converted the captured motions into references for humanoid robots.
Humanoid tracker expands executable manipulation skills from sparse human demonstrations
InterMimicGen iteratively modifies, tests and retains simulated interaction motions, allowing its tracking policy and training archive to improve together.
Big Tech
Yucheng Zhang · Sirui Xu · Jinhong Li · Liuyu Bian · Anatulya Nandi · Derek Zhang · +5 more
University of Illinois Urbana-Champaign · NVIDIA
Research Digest··2 min read
Zhang et al.
Why this paper
From NVIDIA and University of Illinois Urbana-Champaign
In one line
InterMimicGen self-evolves motion imitation data and policies to scale humanoid loco-manipulation from sparse human demonstrations.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§