Biophysics faculty members, Jun Li, associate professor in the department of Applied and Computational Mathematics and Statistics and Olaf Wiest, professor in the Department of Chemistry and Biochemistry, are two of four College of Science faculty members awarded a two-year grant through the National Science Foundation (NSF), that will establish the creation of a pool of new computer nodes dedicated to providing researchers with new high performance technology for quicker speeds.
The University of Notre Dame is bolstering cyberinfrastructure that will support greater access to machine learning.
“GPUs used to only let you play video games with a high rate of animation, but computer scientists found that these GPUs can be used for accelerating the speed of certain computations. One type of computation that can be dramatically accelerated with GPUs is machine learning,” said Kevin Lannon, associate professor of physics and one of the principal investigators for the project.Machine learning, a subfield of artificial intelligence, is an approach to computation in which the computer algorithm learns to recognize connections by exposure to many examples. This approach to computation is roughly inspired by the way the human brain works.
The upgrade will complement the University’s existing graphical processing unit (GPU) cluster at the Center for Research Computing. CAML will enable Notre Dame’s researchers to make more progress in individual runs and do more of the runs in parallel, thereby speeding up the process of machine learning.
Others principle investigators for the project include Paul Brenner, senior associate director of the Center for Research Computing; Jun Li, associate professor in the department of Applied and Computational Mathematics and Statistics; Geoffrey Siwo, research assistant professor in the Department of Biological Sciences, and Olaf Wiest, professor in the Department of Chemistry and Biochemistry.
As part of the $400,000 grant, each of the researchers plans to use the resources in a different way. Wiest, for instance, will use machine learning to discover new chemical compounds. Siwo will be studying the impacts of gene editing technologies, and Li will be using the resources for data science and general exploration of machine learning.
Though they each have specific plans for the resources, the broadened machine learning capabilities will be open to anyone.
“These funds are going to give us the ability to get the project off the ground, and allowing some exploration,” Lannon said. “This should be very transformative.”
The resources should be available to use by the end of the year.
Edited by Cheryl Schairer. Originally published by science.nd.edu on November 19, 2019.at