Ziyuan Yang

GROW slashes the time to incorporate new grain varieties from over an hour to just 39 seconds, revolutionizing the way we approach grain recognition.

Counterintuitively, the most *unreliable* samples in medical imaging datasets—those with fluctuating confidence and frequent forgetting during training—are the *most* informative for building accurate decision boundaries.

By explicitly modeling the manifold structure induced by different CT scanning protocols, this work achieves superior generalization in unsupervised NICT enhancement, overcoming limitations of methods that assume homogeneous noise.

Think your LLM's code is anonymous? This paper shows you can fingerprint it with high accuracy, even across different programming languages.