Zhipeng Wei

Biography

Zhipeng Wei, PhD, is a Postdoctoral Researcher at ICSI, where he works in Dr. Benjamin Erichson’s group on AI safety. His research focuses on enhancing the robustness, security, and reliability of foundation models, with the goal of enabling safe and dependable deployment of modern AI systems.

Dr. Wei received his PhD from the School of Computer Science at Fudan University in 2024. His doctoral research focused on the robustness of deep learning models, particularly in computer vision. He studied adversarial transferability across different modalities, such as images and videos, as well as across architectures including convolutional neural networks and Transformers. His work also examined targeted adversarial attacks and their implications for model reliability.

During his studies, Dr. Wei collaborated with multiple researchers and institutions, including work with collaborators at Fudan University and international research groups. He also spent time as a visiting student at the National University of Singapore.

His current research extends these interests to large-scale foundation models, with a focus on understanding how distribution shifts and adversarial perturbations impact model behavior. At ICSI, Dr. Wei investigates vulnerabilities in large language models, including multi-turn jailbreak attacks, and works toward developing more robust and secure AI systems.

Dr. Wei holds a B.E. and M.Sc. from Jilin University.

Publications

Areas of Expertise

• Machine Learning (Supervised, Unsupervised, Reinforcement Learning)
• Deep Learning and Neural Networks
• Generative AI and Foundation Models
• Planning, Decision-Making, and Intelligent Agents
• Adversarial AI and Robust Machine Learning
• AI Safety, Alignment, and Trustworthiness
• Agentic AI (Autonomous AI Agents)
• Multi-Agent Systems and Collective AI
• AI Ethics, Fairness, Accountability, and Transparency
• Multimodal NLP (Text + Vision, Speech, and Sensor Fusion)

ICSI Activity