Talk Information 11/20/2025

Abstract: Red-teaming is a common practice for exploring the possible failure modes of deployed AI systems. As AI agents advance, this talk presents research on building novel agentic workflows to evaluate the safety of state-of-the-art large language models (LLMs) autonomously and at scale. Specifically, we will introduce a new red-teaming framework based on the composition of human-provided principles, called CoP. Additionally, we will discuss the extension of red-teaming practices from digital safety to physical safety, particularly in the context of LLM-driven drone control.

Bio: Dr. Pin-Yu Chen is a principal research scientist at IBM Thomas J. Watson Research Center, Yorktown Heights, NY, USA. He is also the chief scientist of RPI-IBM AI Research Collaboration and PI of ongoing MIT-IBM Watson AI Lab projects. Dr. Chen received his Ph.D. in electrical engineering and computer science from the University of Michigan, Ann Arbor, USA, in 2016. Dr. Chen’s recent research focuses on AI safety and robustness. His long-term research vision is to build trustworthy machine learning systems. He received the IJCAI Computers and Thought Award in 2023. He is a co-author of the book “Introduction to Foundation Models” and the book “Adversarial Robustness for Machine Learning”. At IBM Research, he received several research accomplishment awards, including IBM Master Inventor, IBM Corporate Technical Award, and IBM Pat Goldberg Memorial Best Paper. His research contributes to IBM open-source libraries including Adversarial Robustness Toolbox (ART 360), AI Explainability 360 (AIX 360), and In-Context Explainability 360 (ICX-360). He received the IEEE GLOBECOM 2010 GOLD Best Paper Award and UAI 2022 Best Paper Runner-Up Award. In 2025, he received the IEEE SPS Industry Young Professional Leadership Award.