Graduate Seminar: AI-Driven Advancements in Battery Materials and Alzheimer's Biomarker Detection

Abstract

The advancement of AI has profoundly impacted scientific analysis, especially in materials science for battery development and computational pathology. In battery research, AI techniques are employed to enhance the safety and performance of Lithium Metal Batteries (LMBs). The study introduces batteryNet, a deep learning model deployed on NERSC's Perlmutter, which deciphers in-operando X-ray tomography data, detecting defects in LMBs and appraising new polymer electrolytes. In the realm of computational pathology, novel AI methodologies merge molecular signals from PET imaging with neural microstructures, allowing unprecedented detection of Alzheimer's biomarkers. This breakthrough enables non-invasive quantification of tau inclusions in live patients, synergizing postmortem studies with live patient imaging, potentially expediting PET tracer clinical approvals for neurodegenerative diseases.

Biography of Speaker

Dani Ushizima, PhD is a Staff Scientist @Applied Math and Computational Research Division, Lawrence Berkeley National Laboratory where she leads the CAMERA ML Team, focusing on computer vision and natural language processing projects. She's also worked as an Affiliated Faculty & Data Scientist for the Bakar Institute at UCSF and the Berkeley Institute for Data Science, UC Berkeley.