Lakshmi Narke is motivated by the illness that killed her good friend and claims the lives of 1.5 million folks yearly. Syntiche Musawu desires to assist folks like her brother, who struggles with isolation resulting from a situation that disrupts how his mind understands what he’s listening to. Tyler Kim is aware of the problem of studying a brand new language and needs to make training extra accessible.
All three college students are among the many hundreds of members on this yr’s Think about Cup, Microsoft’s international competitors for college students to dream and construct new methods of utilizing know-how to enhance lives. In its 21-year historical past, the annual occasion has drawn greater than 2 million college students from 160 nations to compete for prizes together with money, mentorship and coaching.
This yr, a lot of the innovation vitality revolves round AI, with the overwhelming majority of scholars utilizing it to spark concepts, from coaching AI fashions to detect ailments to utilizing cognitive APIs to assist folks with disabilities.
“College students need to make the world higher by addressing and ultimately fixing critical issues, and the supply of AI is empowering them to do it,” says Charlotte Yarkoni, Microsoft president of Commerce and Ecosystems, which incorporates scholar developer applications.
“The rise in AI-powered initiatives in Think about Cup this yr exhibits us the potential of AI — the scholars you see working with AI now might quickly create the following innovation we are able to’t reside with out.”
Right here’s a have a look at a number of the groups utilizing AI, together with the prime three groups — TAWI, Eupnea and CS-M Software — who made it to the 2023 Think about Cup World Championship. That occasion was held at Microsoft Construct 2023 with TAWI named because the winner.
TAWI
Rising up in Congo, Musawu watched her youthful brother wrestle with an auditory processing dysfunction that affected his capability to understand speech, resulting in difficulties in class and social isolation.
“My brother doesn’t really feel assured to speak if there’s background noise, however I feel we will help,” says Musawu, a pc science scholar at america Worldwide College-Africa in Kenya.
She and her staff at the moment are constructing an app referred to as TAWI to assist folks with the dysfunction talk extra simply. It leverages speech recognition instruments from Azure Cognitive Providers and OpenAI Whisper to boost speech, scale back background noise and transcribe speech to textual content in actual time.
Customers can join the app to common earphones, making the answer extra reasonably priced, accessible and discreet than conventional listening to aids. They’ll additionally learn transcribed speech on their cellphone.
Centered on inclusivity, the staff is working with AI fashions to course of completely different languages, accents and background noise, from crosstalk and air conditioners to wind, rain and rural animals.
“We need to serve folks in every kind of environments and circumstances,” whether or not they’re in school, in an workplace or dwelling on a farm, says staff member Zakariya Hussein Hassan.
Eupnea
Most tuberculosis (TB) deaths happen in nations with much less accessible well being care, inflicting delayed testing and therapy. A standard pores and skin check is especially difficult because it includes two visits for an injection and a overview a couple of days later.
Lakshmi Narke and her staff hope their app, Eupnea, will encourage folks to see a physician earlier by shortly assessing a consumer’s danger for TB.
Customers can {photograph} their injection website with a coin on their arm and Eupnea, powered by Azure Cognitive Providers’ imaginative and prescient instruments, calculates danger by measuring the dimensions of the pores and skin’s response with the coin as a reference, the scholars say. One other AI algorithm is designed to research a couple of seconds of coughing.
“AI can scale back the price of making a number of journeys to the physician, who can overview the outcomes and save a life and restrict publicity to different folks,” says Chidroop Iyyhappan, a pc science graduate scholar at College of California, Irvine.
The staff additionally desires to assist folks full the lengthy therapy required for TB by way of a gamified system and a neighborhood that pairs folks for emotional assist with Azure Machine Studying.
The scholars grew up in India, which has the world’s highest burden of TB and is the place Narke’s good friend handed away.
“We now have misplaced family and friends to this illness and witnessed its influence on communities,” says Narke, additionally a pc science graduate scholar at UC Irvine. “With AI, we try to avoid wasting lives one breath at a time.”
CS-M Software
Cardiovascular ailments are the main reason behind dying globally, however most are thought of preventable, a problem in low- and middle-income nations the place well being care and early detection are much less accessible.
A staff of highschool college students in Thailand is constructing a screening software they are saying will help folks verify themselves for potential indicators of coronary heart illness. With a modified stethoscope related to a cellphone, the CS-M Software (brief for Cardiac Self-Monitoring Software) will be capable of analyze a consumer’s heartbeat to offer danger ranges for cardiovascular ailments, the staff says. The scholars additionally plan to include conversational AI to reply primary well being questions.
