Jiaxing Li

A new hierarchical framework for image style reveals an expansive style space, enhancing both diversity and semantic richness in visual representation.

Achieving minute-long real-time interactive world generation while maintaining geometric consistency and dynamic subject identity is now a reality with Matrix-Game 3.5.

Staging social interactions in trajectory prediction leads to remarkable gains in accuracy and consistency, reshaping how we approach agent behavior modeling.

NodeJEPA reveals that masking structural information can significantly improve node-level representation learning without relying on traditional reconstruction methods.

DEFRAG narrows the accuracy gap between small and large language models while slashing costs by up to 98.4%, revolutionizing edge-based AI deployment.