Daniela Ushizima
Host: Keith Butler, Scientific Machine Learning, SCD
Thursday 21st January 2021 at 16:00-17:00
This seminar will take place via Zoom webinar, see joining instructions below.
Machine Learning and Data Analytics: polymer films and beyond
Machine learning (ML) algorithms have been used to streamline analyses and discovery from experimental records, especially processing high-throughput data to enable quick selection of materials configurations, a key step for creating autonomous experiments. Together with scientists at the Advanced Light Source, we have explored the interplay between ML models trained on a relatively limited amount of polymeric properties and domain expertise. This talk will discuss computational methods, focused on Convolutional Neural Networks (CNN) to enable lattice structure classification using diffraction patterns. Such patterns have come from Grazing Incidence Small Angle X-ray Scattering (GISAXS), a surface-sensitive technique with increasingly usage growth in probing complex morphologies, such as conductive polymers. Resulting scattering patterns work as signatures, which depend on the size, shape, and arrangement of the nanostructured components. We will illustrate how we apply ML and AutoML methods to search for configurations in large materials databases, how this can scale analysis to many more samples by providing faster experimental sessions and accelerate scientific discovery. This research work is part of CAMERA, the Center for Advanced Mathematics for Energy Research Applications at Berkeley Lab.
