Rose Yu & Yadi Cao
Postdoctoral Scholar of Computer Science and Engineering
University of California San Diego
Seminar Information
Engineering Building Unit 2 (EBU2)
Room 479
Seminar Recording Available: Please contact seminar coordinator, Jake Blair at (j1blair@ucsd.edu)
While AI has shown tremendous success in many scientific fields, it remains a grand challenge to incorporate physical principles in a systematic manner into the design and training of these models. For the first part of this talk, I will introduce Physics-Guided AI, a framework that aims to integrate first-principled physical knowledge into data-driven methods. By combining the best of both worlds, we can significantly improve sample complexity, computational efficiency, prediction accuracy, and scientific validity of AI models.
For the second part of the talk, we will dive into the application of energy fusion. We will showcase a state-of-the-art deep learning surrogate model for fast simulation of turbulence transport in tokamak fusion. Specifically, our model uses principled feature engineering and physics-guided regularization to improve prediction accuracy. It further integrates Bayesian active learning to reduce the training data requirement.
Dr. Rose Yu is an associate professor at the University of California San Diego, Department of Computer Science and Engineering. Her research focuses on advancing machine learning techniques for large-scale spatiotemporal data analysis, with applications to sustainability, health, and physical sciences. She is a recipient of the Presidential Early Career Award (PECASE)- the highest honor given by the White House to early career scientists, Army ECASE Award, NSF CAREER Award, Hellman Fellow, several industry Faculty Research Award, Best Paper Awards, Best Dissertation Award at USC. She was named as MIT Technology Review Innovators Under 35 in AI.
Yadi Cao is a postdoctoral researcher in the CSE, UCSD, within Rose Yu's STL lab. He completed Ph.D. in 2024 from the Multiples lab in UCLA, co-advised by Professors Chenfanfu Jiang and Demetri Terzopoulos.
His work has been recognized by prestigious awards and mentions, including Best Paper Award at the DLDE workshop at NeurIPS 2023, Spotlight at NeurIPS 2022, and Editor's Pick in Physics of Fluids 2018.
Yadi is seeking tenure-track positions and is open to mentoring and collaboration opportunities.