You are here
Jian Pei: Developing Fair and Transparent Applications for Advances in Data Science
New faculty member Jian Pei connects data science research to real-world applications with an eye on efficiency, fairness and transparency
Jian Pei will join the faculty of the Department of Electrical and Computer Engineering in Duke University’s Pratt School of Engineering beginning July 1, 2022. An expert in developing the connections between data science and real-world applications, Pei is not only enthusiastic about gleaning useful insights from troves of data, but doing so in ways that are fair and transparent to all.
Pei shares joint appointments at Duke in the Department of Computer Science and the Department of Biostatistics and Bioinformatics. He joins the three departments from Simon Fraser University, where he has been a professor of computing science since 2004. He also earned his PhD in computing science from Simon Fraser in 2002 after earning undergraduate and master’s degrees in the field from Shanghai Jiao Tong University.
“Data scientists have to be able to build programs with the ability to learn on the fly and communicate complex results from many different applications to stakeholders and users,” Pei said. “And it is important for data scientists to connect people with data in ways that they can understand and that contribute to society and do social good.”
As the discipline of data science has exploded over the past decades, it has become more difficult to name fields that it hasn’t impacted than those that it has. But as advanced concepts like data mining and applied artificial intelligence become a normal part of more aspects of society, it is becoming increasingly important to make sure they are used in equitable and unbiased ways.
Pei has spent the better part of the past two decades doing just that in a wide variety of fields. His work touches on spatial data, medical data, educational data, ecommerce, bioinformatics, data marketplaces and social networks, just to name a few.
“I’m a big advocate for diversity and fairness in education and training, and particularly for women and other underrepresented groups in science and engineering. We need to have more balance to make sure the entire field is healthy and productive.”
One example of Pei’s work comes in concert with one of the largest social media platforms in the world, LinkedIn. Because executives tend to skew male—especially in technology-related companies—the platform’s algorithms must fight against any possible biases in the connections and opportunities they present to men versus women. The company recently worked with Pei to examine the bias in their data and correct it so that data propagation is fair amongst all its users, no matter their race, gender or other characteristic.
Whatever the field and problem at hand, Pei sees opportunities to customize existing techniques and invent new methods of collecting and parsing data. Potential collaborators can then look to see if the newly applied tools and techniques can be used to solve problems in their own research areas. He is excited by the wealth of future industry collaborations available in the Research Triangle, especially with the likes of Apple and Google set to soon launch office locations in the area.
“I meet all different kinds of people at conventions and meetings, and I’m always looking for opportunities to apply my work to new challenges,” Pei said. “I also often work with professional graduate students who bring new challenges from their businesses, which provides opportunities to then work with their companies.”
Pei is passionate about teaching students of all ages and experience levels—not just professional graduate students. He often teaches undergraduate courses on the basics of data science, data mining and databases, and always has a large cohort of PhD students—50% of which, he is proud to say, are women.
“I’m a big advocate for diversity and fairness in education and training, and particularly for women and other underrepresented groups in science and engineering,” Pei said. “We need to have more balance to make sure the entire field is healthy and productive.”
Besides the outdoor opportunities for jogging and mountain biking that the Triangle area provides, Pei says he was drawn to Duke because of its strong research teams in data science and opportunities for collaboration across disciplines. He’s particularly excited to work with Cynthia Rudin, one of the world’s leading proponents of creating interpretable machine learning algorithms; Robert Calderbank, who runs many data-science-driven programs through the Rhodes Information Initiative at Duke; and Yiran Chen, whose recently launched Athena center focuses on leveraging AI and edge computing technologies and next-generation networked systems in its quest to reimagine future mobile devices.
“I think in the next 10 to 20 years, change is going to happen even faster than it is now, and more dynamic versatility will be very important,” Pei said. “Data science is a key to sense changes, understand dynamics and strengthen versatility. Duke University will play an increasingly important role in leading research, education and societal services, and I am very excited to be part of it.”