New IGB center dedicated to machine learning and predictive modeling

A new Center for Artificial Intelligence and Modeling will be established at the Carl R. Woese Institute for Genomic Biology. It will be led by Sergei Maslov (CABBI), a professor of bioengineering and Bliss Faculty Scholar and Olgica Milenkovic (BSD/CGD/GNDP), a Donald Biggar Willett Scholar and Franklin Woeltge Professor of Electrical and Computer Engineering. The goal of CAIM is to provide biological groups with appropriate expertise in computational sciences.

“We were part of the biocomplexity theme, which historically introduced people with quantitative backgrounds, such as computer science, physics, and mathematics, with biologists so they could work on problems together,” Maslov said. “We wanted to do something along those lines and decided to focus on machine learning and computational modeling.”

CAIM has participants from across the campus including experts from electrical and computer engineering, bioengineering, physics, computer science, mathematics, statistics, and civil and environmental engineering. The biological projects will include members from food science and human nutrition, chemistry, physics, bioengineering, and cell and developmental biology.

“My group works in machine learning, and most of the methods are tailor-made for data sets in e-commerce or from areas that are not quite as exciting as molecular biology, for example,” Milenkovic said. “I thought it would be a really good idea to have more access to interesting biological problems through datasets that could motivate new research ideas and directions in machine learning.”

The goal of CAIM is to combine both modeling and learning in areas including microbiome analysis, single-cell data analysis, multiomics, spatial transcriptomics, and neural networks. Maslov and Milenkovic hope that the center will help generate more research topics based on the collaborations that are forged.

“CAIM is different from the regular themes at the IGB. We don’t have big, overarching projects. Instead, we want to use the center to provide a matchmaking service between biological groups that generate data amenable to modeling and learning with appropriate expertise,” Maslov said. “We want to establish equal collaborations that are of mutual interest to the biologists generating the data and computational people who are analyzing the data.”

Maslov and Milenkovic would like to invite more participants from across the campus to join CAIM. “It would be a great opportunity for people who work in modeling, statistical analysis and algorithm design and those that have specific biological problems. We can provide additional opportunities for them to meet collaborators and develop projects,” Milenkovic said.