Chao Shen

A new system improves AutoML performance by automatically fixing search bugs.

CleanSheet advances model hijacking without altering training processes.

A look at backdoor attacks and defenses in deep learning models.

Exploring challenges and solutions for safeguarding smart grids from data attacks.

Introducing the IC method to improve math problem-solving accuracy.

An overview of testing methods for deep learning libraries and their significance.

Introducing ProCo, a method for better self-correction in large language models.

This study evaluates LCGMs' performance on coding tasks in English and Chinese.

A new method finds performance bugs in DL frameworks efficiently.

A new framework tackles fairness in deep neural networks during model pruning.

BackdoorBench offers a unified approach to assess backdoor learning methods in deep neural networks.

A new method enhances jailbreaking efficiency and effectiveness in large language models.

Researchers tackle object hallucination in AI to improve accuracy and reliability.