Weihao Xuan

Proprietary MLLMs may achieve high diagnostic accuracy, but they still struggle with reliable clinical reasoning, revealing significant gaps in their practical utility.

VLMs struggle with object-level counterfactual reasoning, achieving only a fraction of human accuracy in spatial tasks.

SurfBind's innovative surface-centric approach outperforms traditional methods, revealing the critical role of molecular surface interactions in epitope prediction.

Code-switching can degrade information retrieval performance by up to 27%, revealing a critical blind spot in current multilingual models.