Yuchen Yang

I am a Ph.D. Candidate in Department of Computer Science at Johns Hopkins University, where I'm honored to be advised by Dr. Yinzhi Cao. I'm also working closely with Dr. Neil Gong from Duke University. Before that, I received my M.S. in Security Informatics at Johns Hopkins University and my B.E. in Software Engineer at Shandong University.

My research focuses on the intersection of security, privacy, machine learning (ML), and artificial intelligence (AI). My goal is to develop functional, trustworthy solutions for ML and AI systems. Currently, my work involves diagnosing, correcting, and steering AI behavior—addressing models that generate unsafe content, including sexually explicit, violent, or sensitive data, ensuring seamless real-world deployment aligned with societal values. I am also working on improving the functionality of privacy-preserving ML, such as accurate federated learning and differential privacy. My work has been featured in media outlets like MIT Technology Review and IEEE Spectrum.

yc [dot] yang [at] jhu [dot] edu / Google Scholar / GitHub / CV

I am on the job market for tenure-track faculty or postdoctoral positions.

News

10/2024, I gave an invited talk on Trustworthy AI at Monash University.

09/2024, Our paper on knowledge editing in LLMs has been accepted by EMNLP 2024.

07/2024, Our paper on video anomaly detection using LLMs has been accepted by ECCV 2024.

05/2024, Our paper on mitigating unsafe generation from text-to-image models has been accepted by CCS 2024.

11/2023, Our paper on jailbreaking text-to-image models has been accepted by S&P 2024.

Publications

Conference Papers

2024

SneakyPrompt: Jailbreaking Text-to-image Generative Models
Yuchen Yang, Bo Hui, Haolin Yuan, Neil Gong, Yinzhi Cao
In the Proceedings of the IEEE Symposium on Security and Privacy (S&P), 2024
Reported by MIT Technology Review and IEEE Spectrum.
paper | slides | code

Follow the Rules: Reasoning for Video Anomaly Detection with Large Language Models
Yuchen Yang, Kwonjoon Lee, Behzad Dariush, Yinzhi Cao, Shao-Yuan Lo
In the Proceedings of European Conference on Computer Vision (ECCV), 2024
paper | code

SafeGen: Mitigating Sexually Explicit Content Generation in Text-to-Image Models
Xinfeng Li*, Yuchen Yang*, Jiangyi Deng*, Chen Yan, Yanjiao Chen, Xiaoyu Ji, Wenyuan Xu
In the Proceedings of The ACM Conference on Computer and Communications Security (CCS), 2024
(* Co-first Authors)
paper | code

Ripplecot: Amplifying ripple effect of knowledge editing in language models via chain-of-thought in-context learning
Zihao Zhao, Yuchen Yang, Yijiang Li, Yinzhi Cao
In the Findings of Empirical Methods in Natural Language Processing (EMNLP), 2024
The first author finished the paper mainly under my mentoring.
paper | code

2023

PrivateFL: Accurate, Differentially Private Federated Learning via Personalized Data Transformation
Yuchen Yang*, Bo Hui*, Haolin Yuan*, Neil Gong, Yinzhi Cao
In the Proceedings of USENIX Security Symposium, 2023
Artifact Badges: Artifacts Available, Artifacts Functional, Results Reproduced.
(* Co-first Authors)
paper | code

Fortifying Federated Learning against Membership Inference Attacks via Client-level Input Perturbation
Yuchen Yang, Haolin Yuan, Bo Hui, Neil Gong, Yinzhi Cao
In the Proceedings of IEEE/IFIP International Conference on Dependable Systems and Networks (DSN) 2023
paper | code

2022

Addressing Heterogeneity in Federated Learning via Distributional Transformation
Haolin Yuan*, Bo Hui*, Yuchen Yang*, Philippe Burlina, Neil Gong, Yinzhi Cao
In the Proceedings of European Conference on Computer Vision (ECCV), 2022
(* Co-first Authors)
paper | code

2021

Practical Blind Membership Inference Attack via Differential Comparisons
Bo Hui*, Yuchen Yang*, Haolin Yuan*, Philippe Burlina, Neil Gong, Yinzhi Cao
In the Proceedings of Network & Distributed System Security Symposium (NDSS), 2021
(* Co-first Authors)
paper | slides | code

Preprints

Pseudo-Probability Unlearning: Towards Efficient and Privacy-Preserving Machine Unlearning
Zihao Zhao, Yijiang Li, Yuchen Yang, Wenqing Zhang, Nuno Vasconcelos, Yinzhi Cao
paper | code (coming soon)

Experiences

Research Assistant, at Johns Hopkins University, 2020.3 - Present

Research Intern, at Honda Research Institute, 2023.10 - 2024.2

Teaching Assistant, at Johns Hopkins University, 2020.9 - 2020.12, 2022.9 - 2022.12

Research Assistant, at Chinese Academy of Sciences, 2018.6 - 2018.9

Services

Conference/Journal Reviewing

Reviewer, International Conference of Learning Representations (ICLR), 2025

Program committee, ACM Workshop on Adaptive and Autonomous Cyber Defense (AACD), 2024

Reviewer, IEEE Transactions on Dependable and Secure Computing (TDSC), 2023/2024

Reviewer, IEEE Transactions on Information Forensics & Security (T-IFS), 2024

Artifact evaluation committee, IEEE/IFIP International Conference on Dependable Systems and Networks (DSN) 2024

External reviewer, IEEE Symposium on Security and Privacy (S&P), 2025

External reviewer, ACM ASIA Conference on Computer and Communications Security (ASIACCS), 2024

External reviewer, USENIX Security Symposium 2023/2024

External reviewer, The ACM Conference on Computer and Communications Security (CCS), 2022

External reviewer, IEEE Computer Security Foundations Symposium (CSF), 2022/2024

External reviewer, IEEE International Conference on Distributed Computing Systems (ICDCS), 2022

Organizing and Chairing

Session chair, IEEE Workshop on Deep Learning Security and Privacy (DLSP), 2024

Miscellaneous

Meet my two incredibly charming cats! First up is Go-Wha, whose name translates to "Puppy" in Chinese, because he's always as playful as a puppy. Click here to see the funniest and happiest cat you'll ever meet.

Next is Mao-Dan, whose name means "Snowball" in Chinese, as fluffy and soft as a freshly fallen snow. Discover Mao-Dan here and see why his name is so perfect!

What's more? How about a cheer leader (former), a off-roading adventurer (recent), and a hotpot lover (forever)!

Template