Student: Yichen Guo
Project: M3 - Learning Multifunctional Materials and Machine Learning
View: Research Poster (PDF) | Presentation (YouTube)
Department: Materials Science and Engineering
Advisor: Joshua Agar
Abstract
Multifunctional ferroic materials underpin technologies ranging from energy harvesting to sensing, to next-generation computing. Understanding these materials' functional responses requires advanced synthesis and characterization techniques across extended length and time scales. This has created a data deluge wherein a vast majority of the data collected is under-analyzer. This poster describes our group's efforts to couple advanced synthesis, multidimensional spectroscopy, and machine learning to uncover new functional ferroic physics. We explain how we can leverage epitaxial strain engineering to induce phase competition in PZT (PbZr(1-x)TixO3) using pulsed laser deposition. We describe multidimensional scanning probe spectroscopies are capable of extracting insight into mechanisms of ferroelectric switching. We develop and demonstrate how a long-short term memory neural network can rapidly extract information from these complex experiments. Finally, we provide a perspective of the role of real-time machine learning on the edge to accelerate scientific discoveries and enable real-time control of dynamic transformations in materials.
About Yichen Guo
Yichen Guo is currently a second year Ph.D. student in Materials Science and Engineering department supervised by Dr. Joshua Agar. His current research focus on expanding Machine Learning technique to Materials Science area by training model to understand material’s symmetry structure; he also working on the discovery of new strain induced polar phases by accelerating density functional theory (DFT) calculations using generative machine learning models.