Kun Xu

Achieving a balance between pattern preservation and attribute retrieval, this method boosts retrieval performance while maintaining fine-grained contextual integrity.

InDex bridges the morphology gap, enabling VLA models to excel in dexterous manipulation with minimal data while preserving spatial reasoning.

MLT-Dedup slashes online video repetition rates by 91% while boosting indexing capacity fivefold, revolutionizing deduplication efficiency.

Reinforcement learning boosts multimodal performance, raising task scores and creating unexpected synergies between image generation and editing.