As digital technologies permeate every aspect of modern life, safeguarding the security and resilience of our systems, both virtual and physical, has become a critical concern. The growing sophistication of cyber threats, including adversarial machine learning, attacks on autonomous vehicles, and challenges to industrial control systems, poses significant hurdles for researchers, practitioners, and policymakers alike. This summer school on “AI-Defense in the AI-Cyber-Physical Environment” brings together cutting-edge research and interdisciplinary insights to equip participants with the essential knowledge and tools to understand and address emerging threats. The curriculum will cover secure and robust AI, vulnerabilities in large language models, human-centered security design, and the protection of cyber-physical infrastructures. Special emphasis will be placed on the convergence of generative AI and cybersecurity, examining both the risks it creates and its potential as a defensive tool. Through lectures, discussions, and hands-on activities, participants will engage with the leading edge of cybersecurity research and practice.
Topic 1: Adaptive attacks and Defenses for Robust and Secure AI systems
Speaker: Ilias Tsingenopoulos
Bio Speaker:
I am a postdoctoral researcher at KU Leuven, working on adversarial machine learning and reinforcement learning. My research focuses on both practical and theoretical aspects of adversarial attacks and defenses, across a broad range of systems and modalities: from breaking web bot detection, to adversarial malware and hardening commercial antivirus. More generally, I explore the fundamentals of robust learning and its adversarial and counterfactual facets as essential components in achieving trustworthy and safe AI.
Topic 2: Security in generative AI and large language models (LLMs)
Speaker: Lea Schönherr
Bio Speaker:
Lea Schönherr is a tenure-track faculty at CISPA Helmholtz Center for Information Security since 2022. Her research focuses on information security, particularly adversarial machine learning, trustworthy generative AI, and ML security applications. She is especially interested in language as an interface to machine learning models, including their cognitive representations and code generation with LLMs. She has published several papers on threat detection and defense of speech recognition systems, generative models, and on preventing the misuse of generative AI. She obtained her PhD from Ruhr University Bochum, Germany, in 2021 and is a recipient of two fellowships from UbiCrypt (DFG Graduate School) and Casa (DFG Cluster of Excellence).
Topic 3: Human-centered security
Speaker: Maximilian Golla
Speaker Bio:
Maximilian Golla is a tenure-track faculty at CISPA Helmholtz Center for Information Security since 2023. His research focuses on the intersection of computer security, privacy, and human-computer interaction (HCI). In particular, he is working on reinforcing the security of passwords by driving the adoption of two-factor, risk-based, and passwordless authentication. In the area of usable privacy, he is exploring users’ perceptions of online behavioral advertising, data collection, and tracking, and is analyzing the privacy risks of disruptive technologies. He received his PhD from Ruhr University Bochum, Germany, in 2019 and did a postdoc at the Max Planck Institute for Security and Privacy. His research was covered by TV news, leading newspapers, a surprisingly enthusiastic TV game show, and international media outlets such as Forbes and WIRED.
Topic 4: Security and privacy of cyber-physical systems including autonomous vehicles, drones, and SCADA systems controlling the power grid and other critical infrastructures.
Speaker: Alvaro Cardenas
Topic 5
- Robust ML Classifiers for Android Malware Detection
- Generative Models for Android Malware Detection
- Deterring IP Theft by Generating Fake Data & Documents
- Combating Phishing with Generative AI