National Science Foundation funding supports Rutgers professor’s collaboration to understand dark matter
Coupling technological advances in data science and artificial intelligence/machine learning with established scientific theories is the basis for a study in the detection of dark matter by Rutgers School of Engineering professor Waheed Bajwa and other university colleagues.
Waheed Bajwa, associate professor in the Department of Electrical and Computer Engineering at Rutgers, and a team of scientists have received a $1 million National Science Foundation (NSF) grant to expedite and support the discovery of new subatomic particles through advances in data science and artificial intelligence. The grant is part of ten research and process “big ideas” the NSF has designated as long-term research areas that will push forward new discoveries in science and engineering.
According to Bajwa, the overarching goal of this project is to lay the groundwork for incorporating scientific domain knowledge into data science and artificial intelligence methods, using the tens to hundreds of terabytes of data that are being produced through the physical sciences of biology, astrophysics, earth and materials sciences, oceanography, and others.
“An abundance of data in physical sciences is changing how we advance science,” says Bajwa who has expertise in signal processing, statistics, and machine learning. “Through data science and artificial intelligence, new computational techniques are being developed that can take advantage of the data and accelerate exciting new scientific discoveries.”
Bajwa’s co-investigators are Rice University astrophysicist and lead investigator Christopher Tunnell and Hagit Shatkay, a professor of computer and information sciences at the University of Delaware. The team formed at an Ideas Lab run by the NSF and Knowinnovation that brought together scientists and engineers to facilitate novel data science ideas that did not fit any disciplinary mold.
While dark matter comprises 85 percent of our universe—essentially binding it together—scientists have no lab-based experimental knowledge of its properties. In searching for dark matter, the researchers will use data science and machine learning algorithms to measure astroparticle interactions and measure faint dark matter signals. Probabilistic graphical models and inverse problem formulations will be employed within the XENON1T, a sophisticated dark matter detector located under a mountain in Italy.
“Scientists have traditionally been reluctant to adopt artificial intelligence due to its supposed black-box nature, but this trend is slowly changing,” says Bajwa. “Finding ways to marry physical models with data driven models offers the synergy to focus in on extremely important technological challenges, opening doors to exciting discovery.”
The grant also includes funds for educational outreach and engaging the broader scientific community in the use of domain-specific artificial intelligence techniques for scientific discovery.
September 18, 2019