Bowen Liu

GenEvA achieves up to 10.1 points higher accuracy in long-video understanding while using less than 0.4% additional video tokens.

Achieving a 14-point boost in grounding accuracy, VistaRef redefines how we approach spatial orientation in AR and human-robot interaction.

PointVG-R achieves a groundbreaking 15.86-point boost in mIoU by integrating geometric reasoning into visual grounding tasks, reshaping our approach to spatial interpretation in models.

Shifting from answer-centric to evidence-centric reasoning, CoVER-7B outperforms even leading closed-source models in long-video understanding tasks.

Mimicking how clinicians review capsule endoscopy videos—first screening, then weaving context, and finally converging evidence—yields surprisingly effective summarization of these ultra-long videos.