91吃瓜网

Ozlem Garibay 鈥�01MS 鈥�08PhD earned聽her degrees from 91吃瓜网 in computer science.聽She instructs classes in the Department of聽Industrial Engineering and Management聽Systems. Yet here she is, using all this聽expertise to help a team of 91吃瓜网 researchers聽change medicine 鈥� specifically, how diseases聽are diagnosed and prescription drugs are聽developed. Garibay turns from her computer聽monitor for a few minutes to explain the聽potential impact of the team鈥檚 work and why聽they鈥檙e motivated to put in the hours to make聽it happen.

鈥淚鈥檓 glad we鈥檙e having this conversation,鈥澛燝aribay says. 鈥淚t allows me to step back from聽the micro-details and remember the reason聽we鈥檙e doing this.鈥�

In simple terms, how can artificial intelligence (AI) improve drug development?

The process is like finding a matching key for a lock.聽Every virus is extremely complex, with million聽of unique characteristics. Each compound has its聽own characteristics, too. With the human mind, it鈥檚聽impossible to match the right compounds with a聽virus. Through the best trial and error being used聽now, it can take up to 10 or 20 years to find a match.聽But with artificial intelligence, we can screen for聽suitable candidates of drugs that can be tested聽quicker and with fewer resources.

Give us an idea of the cost of savings.

We recently received a research grant from 91吃瓜网聽for $40,000. It sounds like a lot of money, until聽you consider that it can cost more than $20 billion聽annually for pharmaceutical companies to develop聽a drug using traditional methods. Those high costs聽are passed down to the consumer. We want to get to聽a point where those companies can use our research聽to make the process faster and cheaper.

How did you get involved in this?

It started during the early stages of the pandemic.聽I saw friends and colleagues getting sick, and I feared聽for family members all over the globe. That summer,聽I attended a workshop in Orlando and heard how AI聽can extrapolate volumes of data to discover drugs to聽fight viruses. It sounded brilliant. Then a few days聽later, an engineering professor here at 91吃瓜网, Sudipta聽Seal, told me he was working on a similar problem聽in the Research Park. It made perfect sense to put聽a team together and pursue it.

What do you credit your progress to?

We aren鈥檛 motivated by accolades or personal gain. We know聽this research can help protect precious lives around the world聽for generations. That鈥檚 what keeps us pressing forward.

Could the research be used for diagnosis and prevention, too?

Definitely. We鈥檙e looking at how AI might be used to identify聽鈥渕arkers鈥� in basic tests before a person even knows they have聽an affliction. Our goal is to make it possible for healthcare聽providers to understand each person completely by connecting聽data about lifestyle, environment, health history, genetic聽composition and other factors. That way, they can truly聽individualize care, whether there are symptoms or not.

Why do you have to be careful with this research?

Advancing AI research in an ethical and responsible way is聽near and dear to my heart. We want very much to help people聽have better lives, but the technology must be reliable, safe聽and secure. There can be no compromises. We have recently聽published a paper on human-centered AI that highlights the聽six challenges for the research community to guide responsible聽AI. These include centering human well-being, responsible聽design, respecting privacy, following human-centered design聽principles, appropriate governance and oversight, and聽interacting with individuals while respecting humans鈥櫬燾ognitive capacities.

Might this be a silver lining out of the pandemic?

Oh my goodness, this goes far beyond COVID. The applications聽in healthcare have the potential to slow down diseases like聽Alzheimer鈥檚, cancer and the next global virus. And if it does,聽then we can say it happened partly because a group of聽researchers came together, motivated by a singular mission聽to make lives better.