Peter Steinbach: Deep Learning in a nutshell - a tour de force

Deep Convolutional Neural Networks (CNNs) have seen an unbelievable upspur of uptake in academia and industry since 2012 when a Canadian team of scientists use them to beat competitors by a 10% margin of detection accuracy in the ImageNet Large Scale Visual Recognition Challenge.

In this talk, I'd like to introduce the core concepts of CNNs and dissect the architecture of AlexNet from 2012 as an example. I will then turn to more recent scientific applications of deep learning and try to provide a glimpse the how and why they work. The talk will end with a performance discussion and critical view on the current hype in mass media about artificial intelligence.