The authors propose StateCraft, an architecture that decouples a persistent, updatable abstract state from the visual state in video world models.
Decoupling abstract state from visual state improves video world model consistency
StateCraft pairs a language model with a video diffusion model to track out-of-view changes, achieving high self-consistency over long trajectories.
Big Tech
Lorenzo Venturoli · Nedko Savov · Xi Wang · Marc Pollefeys · Luc Van Gool
INSAIT / Sofia University · ETH Zurich · University of Ulm · Technical University of Munich · Munich Center for Machine Learning
Research Digest··1 min read
Venturoli et al.
Why this paper
From INSAIT / Sofia University and 5 others
In one line
StateCraft pairs a language model with a video diffusion model to maintain a persistent abstract state for consistent and controllable video world simulation.
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
- ✓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.
§