A Ph.D. student at Wuhan University supervised by Prof. Weiwei Liu, majoring in machine learning theory. Prior to that, I obtained a Bachelor of Science degree from Wuhan University in 2020, majoring in Mathematics and Applied Mathematics.

My research experience includes

• 2020-2022: Trustworthy Machine Learning (Adversarial Robustness, Black-box Attacks)

• 2023-Present: Feature Learning Theory: developing theoretical frameworks for understanding how neural networks (of different structures) extract useful features (w.r.t. different tasks) from data (of varying distributions).

I am also serving as

• Reviewers for NeurIPS, ICML, ICLR.
• Commiter of the MindNLP project.
• A Huawei Ascend (Mindspore/Torch-NPU) developer.

1. A Theory of Transfer-Based Black-Box Attacks, Y. Chen and W. Liu, NeurIPS 2023 (Link)

2. Prevalence of simplex compression in adversarial deep neural networks, Y. Cao*, Y. Chen*, and W. Liu, PNAS 2024 (Link)

📧 Email

• yanboch [at] 126 [dot] com (國內, preferred)
• yanbo [dot] acad [at] gmail [dot] com (Academic)

📕 In case you know Mandarin...

• 我平时主要在小红书冲浪，可以搜索用户名 吃鱼的虎 找到我。

"What I cannot create, I do not understand." — Richard Feynman