Xiang Zhou

Forget painstakingly tuning time steps: this generative method learns Wasserstein gradient flows with stable training, achieving high accuracy with only a handful of discretization points.

MRI reconstruction can be made dramatically more robust to clinical domain shifts by eliminating the need for explicit coil sensitivity map estimation.

Sports expose surprising limitations in VLMs' spatial reasoning, as current models struggle to generalize from existing benchmarks despite fine-tuning gains on a new, large-scale dataset.