Alumnus DAIS Seminar: Prof. Jingbo Shang, "Can LLMs Make Genuine Discoveries?"

- Sponsor
- Siebel School of Computing and Data Science
- Speaker
- Prof. Jingbo Shang
- Contact
- Jash Rajesh Parekh
- jashrp2@illinois.edu
- Views
- 162
- Originating Calendar
- Siebel School Speakers Calendar
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks, from competitive programming to scientific reasoning. However, a fundamental question remains: can LLMs genuinely discover new knowledge, or are they primarily recovering and recombining patterns from their training data? In this talk, I will explore this question from a data and supervision perspective, drawing on our recent research in two domains: competitive programming and scientific law discovery. I will discuss how we evaluate LLMs against expert human problem solvers, investigate their ability to generate novel and challenging problems, and examine whether they can uncover scientific laws through controlled experiments. Our findings highlight both the remarkable progress and the remaining limitations of current LLMs, particularly in producing results that are simultaneously novel and correct. Finally, I will share my perspective on the path toward artificial super-intelligence and discuss why learning beyond human supervision may be essential for enabling AI systems to make discoveries beyond human knowledge
Bio: Jingbo Shang is a Full Professor in the Department of Computer Science and Engineering and the Halıcıoğlu Data Science Institute at the University of California, San Diego. He received his B.Eng. degree from Shanghai Jiao Tong University in 2014 and his Ph.D. in Computer Science from the University of Illinois Urbana-Champaign, advised by Prof. Jiawei Han. His research focuses on natural language processing and artificial intelligence, with particular interests in weak supervision, large language models, and AI-driven discovery. He is a recipient of the NSF CAREER Award, Google Research Scholar Award, and MIT Technology Review Innovators Under 35 Asia-Pacific (2026).