Event Location: NSH 1507
Bio: Vicente Ordonez is a PhD student in the Department of Computer Science at the University of North Carolina at Chapel Hill. He holds an MS from Stony Brook University and an engineering degree from the Escuela Superior Politecnica del Litoral in Ecuador. His research interests are at the at the intersection of Computer Vision and Natural Language Understanding. He also works on big scale visual analytics by learning models that can perform high-level perceptual tasks. He is a recipient of the 2013 IEEE David Marr Prize in Computer Vision, a 2012 Yahoo! Key Scientific Challenges Award and a Renaissance Technologies Fellowship. His research has been published in both vision and language conferences and journals (ICCV, ECCV, CVPR, NIPS, ACL, EMNLP, TPAMI, IJCV, TACL).

Abstract: Traditional computer vision systems often explain the visual world by producing a set of labels, object locations or even annotations for every pixel in an image. People instead explain the visual word using natural language. As better representations for image and text data arise, there is an increasing opportunity to create systems that can provide visual explanations with an even higher level of reasoning that incorporates language and world knowledge. I present an extensive study of the connections between language, perception and vision at many levels. From naming objects to generating referring expressions for objects in natural scenes. I also discuss some of the intuitions behind these efforts and how they potentially connect to the larger goal of creating visual systems that can better learn from and communicate with people.