Papers

Trending research and the full catalog - each paper linked to the benchmarks, methods, and models it introduces.

Filtered by domain: image-editingClear

Xiaomi-Robotics-U0: Unified Embodied Synthesis with World Foundation Model

13 Jul 2026

Recent foundation image and video generation models offer strong generalization and controllability, but their direct application to embodied scenarios is limited by requirements for multi-view consistency, geometric coherence, and robot embodiment constraints.

Image editing Image generation Robotics Video generation

1660.1/h

Filtering here covers the 2,000 most recent papers, as much as one page can hold in memory. See the full index of 22,029 papers.