Xin Yang

Tail classes in medical image classification can be effectively shielded from head-class encroachment, leading to significant improvements in performance.

DDVT achieves superior answer grounding in visual question answering by seamlessly integrating visual and textual features, outperforming existing methods on key benchmarks.

Unlearning in offline RL is more complex than previously thought, with common deletion methods showing environment-dependent privacy-utility behavior that can mislead evaluations.

Achieving accurate pose estimation in VIO while slashing data transmission needs by leveraging multi-resolution voxel maps could redefine edge device capabilities.