By MARTY LEVINE
Malihe Alikhani has devoted the early years of her career to promoting diversity in computer research.
Joining Pitt in fall 2020 as an assistant professor of computer science in the School of Computing and Information (SCI), Alikhani says she has always looked for “research that has real impact.” It shows in the projects she has been involved in.
First, she is collaborating with Gallaudet University (which educates those who use sign language) on a natural language processing model “to make the life of deaf and hard-of-hearing students easier when it comes to instruction.” This project aims to use artificial intelligence (AI) to understand sign language or potentially generate the language for the 70 million people who use it.
Second, she is collaborating with colleagues at UPMC and elsewhere at Pitt on a natural language processing system that could one day listen to doctors and nurses converse and automatically fill in patients’ charts.
With linguistics experts at the University of Texas–Austin, she is also studying in-group versus out-group biases as they affect how one group communicates with another.
Last year, she received a Google grant to begin PittInclusion at SCI with a colleague. It’s a new paid undergraduate research program for minority students who have been historically underrepresented in computer science. The program offers hands-on computing research experience and hopes to prompt the students to apply to graduate school and pursue a career in research.
Why choose Pitt? “I really like the interdisciplinary research environment of the University of Pittsburgh in general but specifically the School of Computing and Information. We have a very energetic and exciting school and really interesting faculty, and I really enjoy collaborating with them.
“I like Pittsburgh,” she adds. “It's not too small or too large” — and she’s more-than-used to the hills, having grown up in Tehran, Iran, with its surrounding mountains. Her favorite spots on campus are everyone’s favorites: “Schenley Plaza and hanging around the Cathedral of Learning when the weather is nice.”
Coming to the U.S. for her graduate studies, she earned her Ph.D. from Rutgers alongside a graduate certificate in cognitive science with a focus on language and cognition. Today, she is teaching an undergraduate class, “Bias and Ethical Implications in AI.” It shows how AI — which is being developed by mere humans, after all — can be biased toward certain age groups, genders and other demographics. Facial recognition algorithms, infamously, were designed without the ability to deal accurately with darker skin tones. And machine translations, Alikhani points out, can assume the nurse is a she and the doctor is a he.
“Anyone who is in the AI business,” she says, “should be aware of the consequences the technology can have.”
Marty Levine is a staff writer for the University Times. Reach him at martyl@pitt.edu or 412-758-4859.
Have a story idea or news to share? Share it with the University Times.