Xuan Chen

Xuan Chen

CS Ph.D. Student

Purdue University

About

I am a Ph.D. student in the Computer Science Department at Purdue University, advised by Professor Xiangyu Zhang. I also work closely with Professor Wenbo Guo. Before joining Purdue, I obtained my master’s degree in Electrical and Computer Engineering at Carnegie Mellon University and my bachelor’s degree from University of Science and Technology Beijing.

My research interests lie in AL/ML security, especially backdoor attacks and defenses and using reinformcement learning to solve complex security problems.

Interests

  • AI Security
  • Reinforcement Learning

Education

  • MSc in Electrical and Computer Engineering, 2020

    Carnegie Mellon University

  • BSc in Electrical Engineering, 2019

    University of Science and Technology Beijing

Experience

 
 
 
 
 

Applied Scientist Intern

Amazon

May 2024 – Aug 2024 San Diego, CA
 
 
 
 
 

Research Assistant

Computer Science Department, Purdue University

Aug 2021 – Present West Lafayette, IN
 
 
 
 
 

Research Assistant

Electrical and Computer Engineering Department, Carnegie Mellon University

Dec 2019 – Dec 2020 Pittsburgh, PA

Publications

When LLM Meets DRL: Advancing Jailbreaking Efficiency via DRL-guided Search

Recent studies developed jailbreaking attacks, which construct jailbreaking prompts to fool LLMs into responding to harmful questions. …
Xuan Chen, Yuzhou Nie, Wenbo Guo, Xiangyu Zhang
PDF Code
When LLM Meets DRL: Advancing Jailbreaking Efficiency via DRL-guided Search

BIRD: Generalizable Backdoor Detection and Removal for Deep Reinforcement Learning

Backdoor attacks pose a severe threat to the supply chain management of deep reinforcement learning (DRL) policies. Despite initial …
Xuan Chen, Wenbo Guo, Guanhong Tao, Xiangyu Zhang, Dawn Song
PDF Code
BIRD: Generalizable Backdoor Detection and Removal for Deep Reinforcement Learning

ParaFuzz: An Interpretability-Driven Technique for Detecting Poisoned Samples in NLP

Backdoor attacks have emerged as a prominent threat to natural language processing (NLP) models, where the presence of specific …
Lu Yan, Zhuo Zhang, Guanhong Tao, Kaiyuan Zhang, Xuan Chen, Guangyu Shen, Xiangyu Zhang
PDF Code
ParaFuzz: An Interpretability-Driven Technique for Detecting Poisoned Samples in NLP

Towards Behavior-Level Explanation for Deep Reinforcement Learning

While Deep Neural Networks (DNNs) are becoming the state-of-the-art for many tasks including reinforcement learning (RL), they are …
Xuan Chen, Zifan Wang, Yucai Fan, Bonan Jin, Pitor Mardziel, Carlee Joe-Wong, Anupam Datta
Preprint Project