Latest COVID-19 Information

Gitner Lecture Dives into AI Bias

Associate Professor of Computer Science Kate Saenko spoke about her research as part of virtual Alumni Weekend

| in Community

As artificial intelligence becomes more prevalent, so too are concerns about the potential dangers of it. Machine learning algorithms, which power AI, pervade nearly every aspect of our lives, and just like humans, they can be biased. How are these algorithms held accountable? How can humans understand their decisions? As much as computer scientists are studying AI, they are also studying how to answer these questions and prevent bias in AI and machine learning, both implicit and explicit. 

Kate Saenko portrait
Kate Saenko, Associate Professor of Computer Science

The College and Graduate School of Arts & Sciences proudly calls one of those experts in this field our own. Associate Professor of Computer Science Kate Saenko spoke at the sixth annual Gitner Family Lecture, as part of BU’s 2020 virtual Alumni Weekend, about just this, in her talk “Overcoming Dataset Bias in Artificial Intelligence.” Members of the Gitner family, former students students in the Department of Computer Science, and others joined as Kate spoke about machine learning and where it’s applied (from self-driving cars to Google translate to facial recognition and talk search on our smartphones), dataset bias, and how her research fits into it.

Some bias in machine learning is benign: maybe Siri doesn’t remember your name correctly or a sports game was predicted inaccurately. Sometimes, however, it can reflect the dangerous prejudices we see amongst humans. “There have also been some failures that have been uncovered recently that are more subtle and also more insidious and can affect real people’s lives,” Kate said. Those include algorithms that favor a certain race, gender, or culture or algorithms in devices like self driving cars that can lead to physical harm. Kate then went into how these biases happen in the development phase and a few ways researchers are adapting to prevent them.

You can watch the full lecture and read its transcript on Zoom (Access Passcode: D$.J8S8N).

Kate Saenko leads the Computer Vision and Learning Group at Boston University and is the founder and co-director of the Artificial Intelligence Research (AIR) initiative. She received a Ph.D. from MIT and did her postdoctoral training at UC Berkeley and Harvard. Her research interests are in the broad area of Artificial Intelligence with a focus on dataset bias, adaptive machine learning, learning for image and language understanding, and deep learning.

The annual Gerald and Deanne Gitner Family College of Arts & Sciences Lecture is designed to highlight a current CAS faculty members, in any field, whose teaching and research addresses topics of major importance for the broad interest and benefit of the BU community. It is held in the fall, usually in conjunction with Alumni Weekend. 

Photo Credit (banner): Michael Dziedzic on Unsplash