Voice tech project for Type 2 diabetes

UIC researchers (from left) Alex Leow, Faraz Hussain, Olu Ajilore, Ben Gerber and Jun Ma will present their project, DiaBetty, to a panel of judges in New York City next week.

UIC researchers will present DiaBetty, a virtual diabetes educator, to a panel of judges in New York City on Monday during the final stage of the Alexa Diabetes Challenge.

Developed by UIC researchers Faraz Hussain, Olu Ajilore, Jun Ma, Ben Gerber and Alex Leow and Melvin McInnis at the University of Michigan, DiaBetty uses Amazon’s Alexa platform as a diabetes coach that is sensitive and responsive to the patient’s mood. It provides users with context-dependent, mood sensitive and emotionally aware education and guidance.

DiaBetty tracks changes in the user’s mood by using emotion-related artificial intelligence and by recording the acoustic features of the user’s speech — volume, speed and pitch. Responses are based on analysis of current and past emotional states.

“Given the issue of diabetes-related stress and the impact of depression on outcomes for Type 2 diabetes, it is vital that we address the mental and emotional needs of patients alongside their physical needs,” said Ajilore, associate professor of psychiatry at UIC.

Hussain, the lead developer, says the technology was inspired by previous work on an app called BiAffect, which tracks mood based on keyboard strokes and won the Mood Challenge for Apple’s ResearchKit in May.

“Our entry into the Mood Challenge encouraged us to try to harness the power of voice—our deepest barometer for gauging emotion — and use it to arrest the vicious cycle between diabetes and depression,” Hussain said.

The Alexa Diabetes Challenge launched in April and called on innovators to create Alexa voice-enabled solutions to improve the lives of people newly diagnosed with Type 2 diabetes. DiaBetty is one of five projects selected as finalists.

Ninety-six projects were submitted to the challenge, which is sponsored by Merck & Co., supported by Amazon Web Services, and powered by Luminary Labs.

Print Friendly, PDF & Email


Campus, Research