Yu Xiao

Stored patch vectors from visual document retrieval systems can be inverted to reconstruct original pages, revealing sensitive information and enabling source page identification with high accuracy.

Distilling multi-billion-parameter multi-vector retrievers no longer requires caching terabytes of document tokens: entropic optimal transport allows a lightweight 149M text model to capture 95% of VLM teacher performance at 26x query speed using queries alone.

Audio-first triage cuts VLM calls by up to 20% while improving action coverage, revolutionizing the efficiency of processing egocentric videos.

A single-vector visual document retrieval system that is 15.6 times smaller and an order of magnitude faster than existing multi-vector models while retaining high accuracy.

Quoting evidence directly from documents boosts evidence recall by up to 39 points while halving hallucination rates, challenging the reliance on coordinate-based methods.

Achieving superior annotation efficiency and task success rates, AnnotateAnything revolutionizes how 3D assets are prepared for robot manipulation.

MagicSim revolutionizes robot learning by merging diverse world construction and execution into a single, efficient framework that enhances both evaluation and interaction capabilities.