What happens when artificial intelligence doesn’t have access to millions—or even thousands—of data points? In this episode, Dr. Sarah Ostadabbas, Associate Professor of Electrical & Computer Engineering at Northeastern University and Director of the Augmented Cognition Lab, makes the case for why the future of A.I. depends on learning from small data.
She explains how her team tackles high-stakes problems in healthcare, defense, and autonomous systems by using advanced techniques like transfer learning, domain adaptation, simulation, and collaboration with domain experts. From diagnosing rare diseases with only a few samples to off-road military vehicle training in extreme weather, Sarah illustrates how small data solutions are essential when data is scarce, private, or expensive to label.
We also discuss her upcoming Machine Learning with Small Data course, launching on Coursera in Summer 2025, which teaches industry professionals and researchers how to apply data-efficient A.I. in real-world settings.
About the Guest:
Sarah Ostadabbas leads Northeastern University's Augmented Cognition Laboratory. Her research focuses on building intelligent systems that can learn effectively from limited or imperfect data, with applications ranging from healthcare to military and security domains. She holds a Ph.D. in Electrical and Computer Engineering from the University of Texas at Dallas and has received multiple awards for her theoretical and interdisciplinary contributions to machine learning, computer vision, and artificial intelligence.
