Bin Xiao

AdvDiff improves adversarial examples using diffusion models for better visual quality.

New method improves stealthy adversarial patches for face recognition systems.

A new method improves visual data processing efficiency and accuracy.

New method enhances adversarial attacks on 3D models using diffusion techniques.

A method to speed up text generation without losing quality.

Florence-2 and DBFusion redefine how machines interpret images and text.

Discover how Dynamic Ensemble Reasoning boosts language model performance effectively.

Discover how CustomTTT transforms video creation with unique motion and appearance.