UIC computer scientist will be ‘conscience’ of AI astronomy project

A new multi-institutional project will build on the latest AI technology to look deep into space for astronomy and astrophysics discoveries. And, as part of the project, a University of Illinois Chicago researcher will make sure these tools are reliable, fair and interpretable.

UIC is one of four core universities in the new NSF-Simons AI Institute for the Sky, or SkAI. The project, led by Northwestern University, includes 100 team partners across 25 organizations.  

Lu Cheng, assistant professor of computer science at UIC, will bring her expertise in socially responsible, reliable AI to the collaboration. Her work examines the algorithms underlying AI technologies to make sure they produce trustworthy results and don’t contain hidden biases.  

This growing area of computer science has largely focused on applications such as health care or hiring, where irresponsible AI use can harm humans. With SkAI, Cheng’s group will adapt those insights to astronomy and astrophysics to make sure the tools they design produce accurate scientific results. 

Astronomy and astrophysics research relies on complex datasets taken from observatories, satellites and other instruments surveying cosmic objects such as galaxies and black holes and probing the origins of the universe. Working with this multimodal data will require new AI tools — and new safeguards. 

Cheng’s group will help the researchers assess how AI algorithms make decisions with this data and clarify the level of uncertainty around the tools’ predictions. They’ll also create videos and workshops on socially responsible AI use and AI ethics for astronomy and astrophysics researchers. 

“I think this is a very exciting opportunity because nobody has done this research in astronomy and astrophysics before,” Cheng said. “We will collaborate with astrophysicists, computer scientists and AI ethicists to help define socially responsible AI principles in this specific context and domain.” 

For more on Cheng’s work and her Responsible and Reliable AI Lab, visit Cheng’s website.