The authors propose Physical Registry, a set of registry slots in video Diffusion Transformers (DiTs) that serve as explicit predictive states.
Explicit predictive states boost physical plausibility in video world models
Adding temporally grounded registry slots that co-evolve with video tokens and are supervised by future frame representations improves physical plausibility across multiple backbones.
Big Tech
Shinseong Kim · Jaeseok Jeong · Jianlong Fu · Youngjung Uh
Yonsei University · Microsoft Research Asia
Research Digest··3 min read
The authors introduce Physical Registry, dedicated slots inside a video Diffusion Transformer that encode predictive dynamics.
Why this paper
From Microsoft Research Asia and Yonsei University
In one line
Physical Registry adds explicit predictive states to video DiTs, improving physical plausibility with minimal parameters.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§