Mengdi Wang, 11-12 Feb 2026

Mengdi Wang from Princeton University will visit Caltech on 11-12 Feb 2026. Please sign up here to meet with her:

Wednesday, 11 Feb:

• 7:45 am: Richard Murray, Ath
• 9:00 am: Open (Annenberg Treehouse lounge)
• 9:30 am: Open (Annenberg Treehouse lounge)
• 10:00 am: Open (Annenberg Treehouse lounge)
• 11:00 am: Open (Annenberg Treehouse lounge)
• 11:30 am: Jiachen Yao (Annenberg Treehouse lounge)
• 12:00 pm: Lunch with students (Ath)
• 1:30 pm: Open (meet at Ath, take to Aaron Ames when done)
• 2:00 pm: Aaron Ames (Richard to pick up at ~2:45)
• 2:45 pm: Seminar setup
• 3:00 pm: CDS tea
• 3:30 pm: Seminar
• 5:00 pm: Done for the day

Thursday, 12 Feb:

• 9:00 am: Eric Mazumdar (Zoom)
• 9:45 am: Soon-Jo Chung (235 Guggenheim)
• 10:30 am: Jason Yang (will pick you up in Guggenheim and drop off with Georgia)
• 11:15 am: Georgia Gkioxari
• 12:00 pm: Lunch with Yisong Yue group (Yuan Sui coordinating; take to Chen when done)
• 1:30 pm: Pietro Perona, Chen 3rd floor
• 2:15 pm: Yisong Yue, ANB 3rd floor
• 3:00 pm: Richard Murray, 109 Steele Lab (will pick up in Yisong's office)
• 3:45 pm: Done for the day

---

From Genome to Theorem—and Back to the Lab: Can AI Co-Scientists Do Science?

Professor Mengdi Wang
Princeton University

11 February (Wed), 3-4 pm
213 Annenberg

Large Language Models (LLMs) are increasingly used for scientific reasoning across mathematics, genomics, biology, and physics. This talk discusses recent advances in AI for science—including reasoning for math, physics and emerging science agents—while critically examining their limitations such as overestimated reasoning abilities. I then introduce LabOS, an AI-XR co-scientist that bridges computation and physical science by combining multimodal AI agents, extended-reality interfaces, and laboratory automation. By enabling AI systems to see experimental context, collaborate with humans, and assist in real-time execution, LabOS points toward a future where AI moves beyond analysis to active participation in scientific discovery.

Mengdi Wang is an Associate professor at the Department of Electrical and Computer Engineering and Center for Statistics and Machine Learning at Princeton University. Mengdi received her PhD in Electrical Engineering and Computer Science from Massachusetts Institute of Technology in 2013, where she was affiliated with the Laboratory for Information and Decision Systems. Mengdi received the Young Researcher Prize in Continuous Optimization of the Mathematical Optimization Society in 2016, an MIT Tech Review 35-Under-35 Innovation Award (China region) in 2018, and the American Automatic Control Council Donald P. Eckman Award in 2024.