Runqi Lin
 |
Runqi Lin PhD student
Sydney AI Centre
Address: J12/ 1 Cleveland St, Darlington, NSW 2008, Australia.
E-mail: rlin0511 [at] uni.sydney.edu.au; runqi.lin0403 [at] gmail.com |
Biography
I am currently a final-year PhD student at The University of Sydney, supervised by Prof. Tongliang Liu. Before this, I received my Bachelor's degree in Data Science from Beijing Information Science and Technology University in 2021, supervised by Prof. Wei Zhang. I was a visiting student at the University of Oxford, MBZUAI (Mohamed bin Zayed University of Artificial Intelligence), and Tsinghua University. My research interests lie in trustworthy machine learning, with a particular emphasis on robust generalization and human alignment within foundation models.
Publications & Preprints
Understanding and Enhancing the Transferability of Jailbreaking Attacks.
Runqi Lin, Bo Han, Fengwang Li, and Tongliang Liu.
ICLR 2025. [PDF] [CODE]Instance-dependent Early Stopping.
Suqin Yuan, Runqi Lin, Lei Feng, Bo Han, and Tongliang Liu.
ICLR 2025 (Spotlight, 5.1%). [PDF] [CODE]Layer-Aware Analysis of Catastrophic Overfitting: Revealing the Pseudo-Robust Shortcut Dependency.
Runqi Lin, Chaojian Yu, Bo Han, Hang Su, and Tongliang Liu.
ICML 2024. [PDF] [CODE]On the Over-Memorization During Natural, Robust and Catastrophic Overfitting.
Runqi Lin, Chaojian Yu, Bo Han, and Tongliang Liu.
ICLR 2024. [PDF] [CODE]Eliminating Catastrophic Overfitting Via Abnormal Adversarial Examples Regularization.
Runqi Lin, Chaojian Yu, and Tongliang Liu.
NeurIPS 2023. [PDF] [CODE]
Research Interests
Adversarial Robustness
- Adversarial Attack & Training.
- Jailbreaking Attack in Large Language Models.
- Robustness in Vision-Language Models.Generalization Capability
- Catastrophic & Robust Overfitting.
- Sharpness & Transferability.