Materials Design at SCALE through Automation & Machine Learning

Prof Ong gave a plenary talk titled "Materials Design at SCALE through Automation & Machine Learning" at the NCI and Intersect Showcase on AI and ML in Materials Design and Discovery. Here, we demonstrate how state-of-the-art graph deep learning models can provide rapid accurate property predictions, which enables materials scientists to access a vastly expanded design space in the search of exceptional materials. It is also discussed how ML models can be designed to make the best use of diverse data sources, and how machines can learn chemistry that can further aid in materials design.

Content
-------------
0:00 - Introduction
6:04 - Building reliable ML models for materials science
15:12 - Designing novel materials with ML
24:09 - Addressing THE data problem
31:30 - Learning new chemistry from ML models
38:53 - What's next for ML in materials science?