Katerina Fragkiadaki joins Machine Learning as Assistant Professor
Katerina is interested in Computer Vision, specifically how we understand the stories that videos narrate, and how we build agents that accurately sense their environment and continuously improve both their perception and actions. How do we, humans, infer a stable interpretation of the world despite continuous ego-motion and object motions and occlusions? She is trying to devise algorithms that achieve similar level of understanding, segment objects, infer their 3D motions and shapes, and predict what the future evolution of the scene with little or no human supervision.
Katerina received her Ph.D. from University of Pennsylvania and spent two years as a post doc in UC Berkeley, and one in Google before coming to CMU.