Illinois ECE faculty awarded $2.2 million by DOE for NCSA project

8/11/2020 NCSA

Illinois ECE faculty members Zhizhen Zhao, Volodymyr Kindratenko, Daniel Katz, and Mark Neubauer are leading a project with the NCSA for Physics-Inspired Artificial Intelligence in High Energy Physics. Their project was awarded $2.2 million by the Department of Energy. 

Written by NCSA

The United States Department of Energy (DOE) awards $2.2 million to the FAIR Framework for Physics-Inspired Artificial Intelligence in High Energy Physics project led by accomplished faculty members acros the country including the following Illinois ECE faculty members: Assistant Professor Zhizhen Zhao, Adjunct Associate Professor Volodymyr Kindratenko, Research Associate Professor Daniel S. Katz and Faculty Affiliate Mark Neubauer

Zhizhen Zhao
Zhizhen Zhao

Spearheaded by the National Center for Supercomputing Applications' Center for Artificial Intelligence Innovation (CAII) and the University of Illinois at Urbana-Champaign, the primary focus of this project is to advance our understanding of the relationship between data and artificial intelligence (AI) models by exploring relationships among them through the development of FAIR (Findable, Accessible, Interoperable, and Reusable) frameworks. Using high-energy physics (HEP) as the science driver, this project will develop a FAIR framework to advance our understanding of AI, provide new insights to apply AI techniques, and provide an environment where novel approaches to AI can be explored.

This project is an interdisciplinary, multi-department, and multi-institutional effort led by Eliu Huerta, principal investigator, director of the CAII, senior research scientist at NCSA, and faculty in PhysicsAstronomyComputational Science and Engineering and the Illinois Center for Advanced Studies of the Universe at UIUC. "Innovative AI applications have enabled disruptive advances across disciplines. This realization has led to significant investments by DOE to create a rigorous framework that maximizes the impact of intuitive, AI-driven discovery," says Huerta. "As part of the mission of CAII, we create synergies across units and departments at U of I, and with collaborators across the U.S., to continue spearheading disruptive advances in AI, facilitating its use and adoption by new practitioners. This DOE award will enable us to further these activities."

NCSA's Center for Artificial Intelligence Innovation advances AI research, provides students with opportunities for career development in AI, and addresses industrial grand challenges through innovative use of AI by engaging the research community, students, and industry collaborators.

Volodymyr Kindratenko
Volodymyr Kindratenko

Alongside Huerta are co-PIs from Illinois: Zhizhen Zhao, assistant professor of Electrical & Computer Engineering (ECE) and Coordinated Science LaboratoryMark Neubauer, professor of physics, member of Illinois Center for Advanced Studies of the Universe, and faculty affiliate in ECE, NCSA, and the CAII; Volodymyr Kindratenko, co-director of the CAII, senior research scientist at NCSA, and faculty at ECE and Computer Science (CS); Daniel S. Katz, assistant director of Scientific Software and Applications at NCSA, faculty in ECE, CS, and School of Information Sciences.

In addition, the team is joined by co-PIs Roger Rusack, professor of physics at the University of MinnesotaPhilip Harris, assistant professor of physics at MIT; and Javier Duarte, assistant professor in physics at UC San Diego.

This project aligns with the DOE's initiative to advance FAIR data principles that focuses on making AI data and models more accessible and reusable by application developers and researchers to further accelerate AI research and development. Through this award, the interdisciplinary and multi-institutional team of experts will lead the definition and adoption of FAIR principles for AI models and data in the context of HEP.

"We are very excited about the DOE's support for our project on establishing a FAIR framework for physics-inspired AI in HEP," says Zhao. "Our efforts for constructing the FAIR benchmark data in HEP and building a theoretical framework for FAIR AI models will lead to new interdisciplinary collaborations at the intersection of physics, engineering, and computer science and advance DOE's objectives in data-driven scientific discovery."

Daniel S. Katz
Daniel S. Katz

"I was excited to see DOE supporting using the concept of FAIR beyond data, which is where it was first created," says Katz. "I've been working on expanding FAIR beyond data, to software and notebooks, and this gives us the opportunity to take that experience and combine it with the AI and physics strengths of the university and our collaborators, to build and share the datasets and models that will enable new discoveries."

The UIUC High Energy Physics Group is pursuing a deep understanding of fundamental particles and their interactions, and the nature of dark matter and dark energy. Neubauer is a member of the ATLAS group at the Large Hadron Collider (LHC) which contributed to the Higgs boson discovery that led to a Nobel Physics prize in 2013 and continues to search for new physics at the LHC. The collaboration between Illinois' CAII and HEP group will further the DOE's objective toward constructing a theoretical framework that makes the best use of AI in science and engineering.

"AI methods are key to maximizing the discovery potential in HEP as it enters the High-Luminosity Large Hadron Collider era," says Neubauer. "Through support from the DOE, we will develop a FAIR framework for physics-inspired AI that will increase the performance and transparency of AI models used to interpret HEP data and strengthen collaboration between physicists and computer scientists for both HEP and AI research innovation."

"Creation of a FAIR framework for physics-inspired AI is an important undertaking that will benefit a large community of AI practitioners well beyond HEP," says Kindratenko. "Physics-inspired AI models are becoming increasingly popular to model complex processes that so far have been addressed mostly by solving differential equations on large-scale computer systems. The proposed framework will be instrumental in enabling model sharing and reuse across many disciplines. This is a great example of how an investment made by DOE will lead to innovations beyond its original target."

Read the full award summary. View the award summaries.

Read the original article on the NCSA site.


Share this story

This story was published August 11, 2020.