91°”Íű

INQUIRING MINDS

AI for Good

By Brett Bralley

Computer Engineering Professor Magdalini Eirinaki is showing students — and the world — what it means to be a socially responsible artificial intelligence engineer.

You know how once you've finished binge-watching a Netflix show, you’re immediately suggested another one that’s similar? That technology is called a recommender system, and it’s built using artificial intelligence and machine learning. Recommender systems gather data about you — the user — then make content suggestions based on that information.

“It fascinates me how people use these technologies,” says Magdalini Eirinaki, professor and associate chair of computer engineering at San JosĂ© State University — and an artificial intelligence and machine learning expert and researcher.

Eirinaki wants to make recommender systems even better, because while artificial intelligence and machine learning allow us to do amazing things, “it has a flip side, and we should be more critical of it,” she explains. She points to the Cambridge Analytica scandal that came to light in 2018, in which more than 87 million Facebook users’ data was collected without consent and used for political advertising. The propagation of fake news on social media is also a strong example of how such technologies can be used to reinforce social biases, she notes.

“We should be educating more about the dangers of AI, while promoting its good uses.”

Her mission is to harness the power of AI for good, and she’s dedicated to teaching her students how to make a difference. Whether it’s her own research or supporting student projects, “we can use AI as a tool to benefit others and as a method to solve different problems,” she says.

For example, she and her students built an app that helps translate American Sign Language into text, incorporating vocabulary often used in restaurant environments. She has also worked with students to create Viva — a virtual assistant that provides voice navigation, obstacle detection and a surrounding risk assessment for the visually impaired.

She has even developed an alternative to traditional recommender system algorithms leveraging the user’s social network. Instead of suggesting content based on what others similar to you like — how recommender systems usually work — she designed one that recommends based on your circle of influence, including your friends, and friends of friends. Putting it to the test, she found it works.

“We saw that introducing real-life relationships greatly improved the results when compared to more traditional recommender system algorithms,” she notes.

Recently, Eirinaki co-led the charge to launch the Charles W. Davidson College of Engineering’s new MS in Artificial Intelligence program, which accepted its first in-person cohort in fall 2021.

When she first joined 91°”Íű in 2007, Eirinaki was the main faculty member teaching AI-related courses. But soon, the college recognized a growing demand for AI education, as the industry began to need more experts. For the last two years, she and a group of faculty members conducted market research and met with business and technology leaders in Silicon Valley to build a program that would provide students with the education and experience to become socially responsible AI engineers.

“We can use AI as a tool to benefit others and as a method to solve different problems.”

— Magdalini Eirinaki

For Eirinaki, teaching is a passion. When she was a middle school student growing up in Athens, Greece, she wanted to be a teacher. As she got older, she developed an affinity for all things computer science. After earning both her bachelor’s and master’s in the field, she spent a short stint working for a software company. But she found herself working on the same tasks every day, and missed opportunities to be learning and trying new things. So she returned to academia to work on her PhD.

That’s when she fell in love with working with students. She admits part of that lies in instant gratification.

“You can start working with a student when they don’t know much and by the end of even one class they’ve learned something new. It’s transformative,” she says.

Zeeshan Pachodiwale, ’20 MS Computer Engineering, says he felt that transformative process when working with and learning from Eirinaki.

When he took one of her data mining courses, “the knowledge gaps I had had from a previous course were filled because of the way she teaches and helps you understand,” he says.

For that reason, he asked her for help in realizing the Viva project — the assistance app for the visually impaired.

“She guided [me and other students] through that process, was supportive and helped us aim high,” he says. Now, Pachodiwale works as an engineer for Amazon, and he hopes to continue pursuing machine learning and data mining throughout his career.

That’s what Eirinaki appreciates about the long term impact of teaching: She’s helping change lives. After 15 years at San JosĂ© State — an institution Eirinaki says lives up to its ranking by Money magazine as the “Most Transformative University” — she’s seen how her lessons have had an impact on students.

“I’ve had several students come back and tell me, ‘What you taught me helped get me an interview or land my dream job,’ many years later. That’s the rewarding part of teaching.”

A blue wavy line.

Top photo: David Schmitz

91°”Íű San JosĂ© State University

Washington Square: San José State University's Magazine © 2022. All Rights Reserved | Land Acknowledgement