Faculty
2025: AI 2000 Most Influential Scholar Award Honorable Mention 2025
2025: Best Machine Learning and Security Paper in Cybersecurity Award 2025
2022: Busy Beaver Award für "Privacy of Machine Learning"
2021: Busy Beaver teaching award for seminar “Privacy of Machine Learning” at Saarland University (2021 Winter)
2019: Best paper award at NDSS
Dr. Yang Zhang is tenured faculty member at CISPA. His research concentrates on trustworthy machine learning (privacy, safety, and security). Moreover, he works on measuring and understanding misinformation and unsafe content like hateful memes on the Internet. Over the years, he has published multiple papers at top venues in computer science, including CCS, NDSS, Oakland, and USENIX Security. His work has received the NDSS 2019 distinguished paper award and the CCS 2022 best paper award runner-up.
Conference on Neural Information Processing Systems (NeurIPS) Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency
Conference on Neural Information Processing Systems (NeurIPS) Finding and Reactivating Post-Trained LLMs’ Hidden Safety Mechanisms
Conference on Empirical Methods in Natural Language Processing (EMNLP) Breaking Agents: Compromising Autonomous LLM Agents Through Malfunction Amplification
IEEE International Conference on Computer Vision (ICCV) Hate in Plain Sight: On the Risks of Moderating AI-Generated Hateful Illusions
ACM Conference on Computer and Communications Security (CCS) UnsafeBench: Benchmarking Image Safety Classifiers onReal-World and AI-Generated Images
IEEE Transactions on Dependable and Secure Computing Revealing the Risk of Hyper-parameter Leakage in Deep Reinforcement Learning Models
Usenix Security Symposium (USENIX-Security) Data Duplication: A Novel Multi-Purpose Attack Paradigm in Machine Unlearning
Usenix Security Symposium (USENIX-Security) Bridging the Gap in Vision Language Models in IdentifyingUnsafe Concepts Across Modalities
Usenix Security Symposium (USENIX-Security) On the Proactive Generation of Unsafe Images From Text-To-Image Models Using Benign Prompts
Usenix Security Symposium (USENIX-Security) Generated Data with Fake Privacy: Hidden Dangers of Fine-tuning Large Language Models on Generated Data