Putting clear bounds on uncertainty

Researchers from MIT CSAIL predict semantically meaningful and calibrated uncertainty intervals in the latent space of a generative adversarial network to know the exact limits in our ability to clean up, and reconstruct, partly blurred images.

Related Keywords

Israel ,Yaniv Romano ,Stephen Bates ,Anastasios Angelopoulos ,Phillip Isola ,Artificial Intelligence Laboratory ,Simons Institute ,Us Air Force Artificial Intelligence Accelerator ,National Science Foundation ,Foundations Of Data Science Institute ,Israel Science Foundation ,Lincoln Laboratory Supercomputing Center ,University Of California At Berkeley ,Us Air Force Research Laboratory ,Israel Institute Of Technology ,Agreement Number ,Swami Sankaranarayanan ,Israel Institute ,Neural Information Processing Systems Conference ,New Orleans ,National Science ,Data Science Institute ,Career Advancement Fellowship ,Force Research Laboratory ,Force Artificial Intelligence Accelerator ,Cooperative Agreement Number ,Amit News ,Mit Csail ,Computer Vision ,Uncertainty Quantification ,Emantic Uncertainty ,Latent Spaces ,Generative Adversarial Networks ,Generative Models ,Istribution Free Statistics ,Stylegan ,Saliency Map ,