Towards Accelerated and Reproducible AI-Driven Discovery

Eliu Huerta, University of Illinois at Urbana-Champaign
Huerta ADSP2020

This image is a snapshot from a visualization that shows the output of the penultimate layer of a deep neural network during training as it learns to classify spiral and elliptical galaxies. Credit: Janet Knowles, Joseph A. Insley, and Silvio Rizzi, Argonne National Laboratory


In this talk I will describe disruptive advances I have pioneered at the interface of domain-inspired AI and extreme scale computing that have led to new modes of data-driven discovery. I will present exemplars on how to scale disruptive advances in AI, distributed computing, and scientific data infrastructure to open new pathways to conduct reproducible, accelerated, data-driven discovery. I will also touch on the need to create and sustain a solid core of AI expertise that nimbly crosses academic and industry boundaries to transform AI innovation into tangible societal, economic and business benefits.  


Eliu Huerta is a theoretical astrophysicist, mathematician and computer scientist with broad research interests. He has done pioneering work at the interface of domain-inspired AI and extreme scale computing for Multi-Messenger Astrophysics, Cosmology, Observational Astronomy, and extreme scale simulations that describe multi-scale and multi-physics processes. He enjoys doing translational research in medicine, technology and industry. He is the founding director of the Center for Artificial Intelligence Innovation at the University of Illinois at Urbana-Champaign, and the Head of the Gravity Group at the National Center for Supercomputing Applications. He leads several NSF and DOE funded interdisciplinary and multi-institutional projects that focus on disruptive AI applications and advanced computing for big-data physics experiments. 

Please use this link to attend the virtual seminar:

Meeting ID: 384 806 131