ML/AI for Optimizing Materials Synthesis and Device Operation

Prof. Julia Hsu from University of Texas at Dallas

@ CCMS/PHYSICS BUILDING R212

Abstract:

In this talk, I will show how my group uses Bayesian Optimization (BO) to accelerate the identification of optimal synthesis conditions to achieve materials with desired properties and the tuning of operation parameters for optimal performance in neuromorphic thin-film transistors (TFTs). BO is particularly useful in cases with a large number of inputs, costly experiments, or interconnected relationships between variables. Our work focuses on applying BO to fabricate high-performance solar cells and flexible metal oxide capacitors, and on finding the optimal operating parameter that correctly encodes time-series data. While mathematicians and computer scientists often test BO algorithms on synthetic data, applying these techniques to real-world experimental data requires accounting for the limitations of the tools used and for experimental failures that do not yield training data. Additionally, we are exploring transfer learning that leverages proxy experiments to save time on labor-intensive experiments. With the recent “AI Scientist”, the role of humans in scientific research is a thought-provoking question.

 

Short Biography:

Julia W. P. Hsu is a Professor of Materials Science and Engineering at the Erik Jonsson School of Engineering and Computer Science of the University of Texas at Dallas (UTD) and holds the Texas Instruments Distinguished Chair in Nanoelectronics. She received her B.S.E. degree in Chemical Engineering from Princeton University, her Ph.D. degree in Physics from Stanford University, and completed her postdoctoral training at Bell Labs. Prior to UTD, she worked at Sandia National Laboratories, Bell Labs, and the University of Virginia. 

Prof. Hsu’s research encompasses a wide range of materials, from organic materials to metal oxides to semiconductors. Her recent work focuses on the physics and applications of organic and perovskite solar cells, novel processing of flexible electronics, new resist materials for extreme UV lithography, and machine learning approaches to accelerate materials synthesis and processing. Prof. Hsu is a Fellow of the American Physical Society (APS), the American Association for the Advancement of Science, the Materials Research Society (MRS), and the Institute of Physics (IoP). She has served in key leadership roles in professional societies, including the Board of Directors and Treasurer of MRS, the DOE Basic Energy Sciences Advisory Committee, the National Academies Review Panels, and external advisory committees for research centers. She has published over 275 journal papers, has been granted 5 patents, and has given over 210 invited talks. (https://personal.utdallas.edu/~jxh101000/)

<< Back