Artificial Intelligence Seninar - Dr. Qing Li, University of Zayed - United Arab Emirates

Title: Building Trustworthy and Interpretable AI Systems

Abstract:

In recent years, large language models (LLMs) have emerged as powerful engines for reasoning and perception, enabling the development of increasingly complex intelligent agent systems. However, these advancements also raise critical concerns regarding reliability, safety, and control. In this talk, I will present my research on enhancing trustworthiness in AI, focusing on three major challenges: hallucinations, jailbreak attacks, and privacy leaks. I will introduce principled methods for hallucination detection, safety calibration, and robust defense mechanisms across both unimodal and multimodal systems. Building on this foundation, I will outline my vision for future research at the intersection of LLM reasoning, continual learning, and multi-agent collaboration.

This seminar will be followed by a lecture of Dr. Qing Li until 14:30 hours.