Michael McCabe

Michael McCabe

I am a researcher at the Flatiron Institute working on deep learning for the physical sciences.

Talks and presentations

Data Science in the Healthcare Industry

February 27, 2020

Talk, Colorado State University, Data Science Seminar, Fort Collins, CO

From clinician-level analysis of medical imaging, to precision medicine powered by fine-grained genomic models, to automated drug discovery, it often seems like machine learning is only a few years away from completely transforming healthcare as we know it. While progress is certainly being made, in practice, the healthcare industry is an enormously complicated beast that is more resistant to change than one might expect from browsing arXiv. In this talk, I will discuss the structure of the healthcare industry, where organizations are currently deriving value from data science, and the complications that arise when trying to translate hype into improved care. We’ll explore a few use cases spanning both the clinical and business side of healthcare, focusing on problems shared by both large research institutions and smaller community hospitals and cover some specific technical and data challenges frequently faced by healthcare organizations as they start exploring data science.

Graph Neural Networks in the Physical Sciences

November 19, 2019

Talk, CU Boulder, Optimization and Machine Learning Seminar, Boulder, CO

Deep neural networks have contributed to wildly improved performance across a variety of tasks in fields like NLP and computer vision. One property of deep learning that gets credited for this progress is the ability to encoder prior knowledge of the problem structure into the architecture of the network. Unfortunately, the convenient low dimensional grid structure found in these problems that allows for straight-forward learning of locally compact filters is not present in the wide variety of non-Euclidean data used by other scientific fields which can often only be represented as graphs or point clouds. Graph Neural Networks, and in particular Graph Convolutional Neural Networks, are a family of architectures developed to address these types of problems. In this presentation, I’ll review the major ideas and milestones in the development of modern graph neural networks and highlight several recent applications in the physical sciences.