The authors designed a controlled video self-supervised learning (SSL) protocol to isolate which architecture-objective combinations produce motion-prioritized representations.
Video pretraining decouples motion and appearance paths to boost motion tasks
A matched 4x6 sweep of architectures and objectives shows only the TT3D plus Diff Compression pairing yields a representation that simultaneously leads four motion-heavy fine-tuning benchmarks.
Big Tech
Shih-Ying Yeh · Daniel Z. Kaplan · Xuehai Wang · Fu-En Yang · Min-Hung Chen · Shang-Hong Lai
National Tsing Hua University · Comfy Org Research · realiz.ai · Karolinska Institutet · Stockholm University
Research Digest··3 min read
The authors introduce TT-VidT, a video self-supervised learning method that separates static appearance from temporal change.
Why this paper
From NVIDIA and 5 others
In one line
TT-VidT decouples appearance and motion in video pretraining, leading on motion-heavy benchmark fine-tuning with fewer FLOPs.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (2 noted)
- ✓Reports numbers on named benchmarks (5 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§