Universitšt Bonn: Autonomous Intelligent Systems Group   Computer Science Institute VI: Autonomous Intelligent Systems

Deep Learning 

Our work in Deep Learning is focused on Computer Vision with emphasis on object categorization, object-class segmentation, and action recognition tasks. Deep Learning methods produce a hierarchy of representations using multi-layered models, such as deep neural networks or deep graphical models. To accelerate calculations with GPUs, we make use of the Nvidia CUDA framework.

Neural Abstraction Pyramid Architecture   Pavel et al.: Semantic Segmentation of RGB-S Video     Titan CUDA computers with a total of 34.560 CUDA cores and 96GB DDR5 RAM