Multidisciplinary research team at Purdue included in NSF $15M AI award
10-07-2021
As scientific data sets become progressively larger, algorithms to process the data quickly become proportionally more complex. New processor types, such as graphics processing units (GPUs) and field-programmable gate arrays (FPGAs), allow Artificial Algorithms (AI) algorithms to be greatly accelerated.
The newly-created $15 million NSF HDR Institute of Accelerated AI Algorithms for Data-Driven Discovery (A3D3) aims to incorporate AI algorithms with new processors to analyze unprecedented data sets. A3D3 provides the opportunity for algorithm/system experts and domain scientists to closely work together to solve problems with extremely large data sets.
A team of interdisciplinary researchers at Purdue University were selected as part of the A3D3 award. The team will be led by Professor Mia Liu of the Department of Physics and Astronomy. The team includes Professors Maria Dadarlat of the Weldon School of Biomedical Engineering and Pan Li of the Department of Computer Science.
“This grant will fuel collaborative work between computer science and domain scientists,” says Liu. “For example, Professor Li and I have been working on solving challenging problems in particle physics with graph neutral networks. We have presented our results in recent conferences (Semi-supervised machine learning for pileup per particle identification at the LHC with Graph Neural Networks) and have a few papers in the pipeline.”
Professor Li attributes graph-based machine learning (ML) techniques as the most suitable way to process sparse and irregular data that is prevalent in scientific experiments.
“Many scientific experiments generate large-scale data that needs to be processed properly and efficiently,” said Li. He added, “Developing powerful and scalable machine learning algorithms to process such data and combining the algorithms with dedicated hardware/system support are the main ideas behind A3D3.”
Li shared, “The Purdue CS focus in A3D3 is on the algorithm design, where our team will develop scalable graph-based ML solutions to the scientific problems encountered in high energy physics, astronomy and neuroscience.”
A3D3 targets fundamental problems in three fields of science: high energy physics, multi-messenger astrophysics, and systems neuroscience. A3D3 works closely within these domains to develop customized AI solutions to process large datasets in real-time, significantly enhancing their discovery potential.
The ultimate goal of A3D3 is to construct the institutional knowledge essential for real-time applications of AI in any scientific field. A3D3 will empower scientists with new tools to deal with the coming data deluge through dedicated outreach efforts.
This article first appeared with the Department of Physics and Astronomy.