Amazon and NSF give grant to UI researchers to make algorithms less discriminatory

A team of researchers and professors from the University of Iowa has received an $800,000 grant from the National Science Foundation and Amazon to make machine learning algorithms less discriminatory.

“Machine learning is used to make many high-stakes decisions, but it often discriminates against people who have protected characteristics,” said Qihang Lin, co-principal investigator of the grant, with Tianbao Yang, associate professor of computers in college. Liberal Arts and Sciences. “We want to help ensure that these decisions will not be discriminatory.”

Machine learning is the process of programming an algorithm to analyze huge amounts of data so that it learns to perform tasks it is programmed to do. Once additional data is added, the algorithm learns and changes the same way a human learns and reacts.

Analysts have found that these algorithms can often discriminate against people based on their race, gender, health status, etc., according to a press release.

Lin said algorithms can learn discriminatory things based on data. For example, an algorithm might conclude that blacks are less susceptible than whites to a certain disease based on data showing that fewer blacks are tested for the disease than whites or are hospitalized less often for the disease. But what the algorithm didn’t say was that fewer black people have access to health care, so even if they have the disease, they’re less likely to see a doctor.

Researchers will use the three-year grant to follow up on research already underway to define fairness and examine different measures of risk to help business leaders balance fairness and risk when making decisions. in health care.

Mingxuan Sun of Louisiana State University is the project’s third co-principal investigator.

Sharon D. Cole