Bloomberg Distinguished Professor of Computational Cognitive Science Alan Yuille and his co-authors have received a 2026 Frontiers of Science Award from the International Congress for Basic Science (ICBS) for their paper “DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs,” published in IEEE Transactions on Pattern Analysis and Machine Intelligence in 2017.
The ICBS honors top original research of outstanding scholarly value; its Frontiers of Science Award was inaugurated in 2023 with the goal of encouraging young scholars to look to the frontiers of basic science, set goals to obtain breakthrough results as early as possible, and contribute to the study of the natural world. Sponsored by the City of Beijing and the Yanqi Lake Beijing Institute of Mathematical Sciences and Application, awards are given each year for basic and applied research in mathematics, physics, and theoretical computer and information sciences.
The director of Johns Hopkins’ Computational Cognition, Vision, and Learning group, Yuille researches computational models of vision, mathematical models of cognition, medical image analysis, and artificial intelligence and neural networks. He was joined on this paper by Liang-Chieh “Jay” Chen, George Papandreou, and Iasonas Kokkinos—all former graduate students or postdoctoral researchers in Yuille’s group at the University of California, Los Angeles—and Kevin Murphy, a collaborator at Google.
The team’s DeepLab system addresses the task of semantic image segmentation with deep learning; designed to classify every pixel in an image, it achieves high-accuracy segmentation by using atrous convolution and atrous spatial pyramid pooling to capture multi-scale context without losing resolution.
“DeepLab was just the start of this research,” Yuille says. “Orchestrated by Jay Chen, there was a sequence of papers which kept updating DeepLab to keep it state-of-the-art, including switching its architecture from convolutional neural networks to transformers. This means that DeepLab papers have been cited over 100,000 times total.”
A formal announcement of the team’s award took place on May 18.
Learn more about the Frontiers of Science Awards here.