Researcher improving fairness in computer systems that process human language
November 13, 2020 / by Nanci Hellmich
"I like solving problems. At its core, computer science is about finding efficient solutions.”
-- Antonis Anastasopoulos, assistant professor of computer science
When you ask your Google Assistant a question or type one into the search engine of your computer, you expect to get a reasonable answer.
But your accent or dialect may get in the way, preventing the system from understanding you.
To improve these systems, Antonis Anastasopoulos, an assistant professor in the Department of Computer Science, received a Google 2020 Award for Inclusion Research for his project, "Modern NLP for Regional and Dialectal Language Variants." Anastasopoulos studies natural language processing (NLP), which is how computers attempt to process and understand human languages.
“The goal of the research is to promote fairness and equity in these computer systems, so they work well for everyone,” he says.
Anastasopoulos’ research combines his love of computer science with languages and linguistics. He speaks Greek (his native language), English, German, Swedish, Italian, and some Spanish.
His research will apply to all languages. It’s important to look deeply into languages for differences because languages are flexible and diverse, he says. “There are many regional variations that are different from the standard.”
For instance, in the United States, there are variations in dialects and accents across the country and among native and non-native English speakers. “We are going to identify how big the difference in performance is in these systems when you switch from standard to non-standard language and then try to mitigate the problems so the system is fair,” he says.
David Rosenblum, chair of the Department of Computer Science, says, “Antonis’s research attacks very important pieces of the NLP puzzle, which are the widespread dialectic variation in spoken language and the widespread variation in accent, especially among non-native speakers of a language. And his award from Google is a great reminder that diversity and inclusion require not just sensitivity to gender and color and ethnicity, but to language as well.”
In his spare time, Anastasopoulos reads about languages and linguistics. “I find language so fascinating. Now I build tools to understand and generate language.”
Computer science is the key to solving many problems, he says. “My father taught computer science in high school in Greece, and my mother taught English. Growing up, I always played with computers. I like solving problems. At its core, computer science is about finding efficient solutions.”