The first SGD Edition day - a day long hackathon to learn, investigate, document and develop a deeper understanding of some technique or problem in Data Science. This session focused on Filters, Styles, and Domains.
The addictive technique of using deep learning to convolute images with the styles of famous paintings and artists has become synonymous with deep learning. But can we develop an understanding of the problem deeper than its mere application? Can the same techniques be used for un-bluring images or unfolding histograms? What are the possibilities for maping features from a learned space to an example space? Let’s find out together!
In this edition we explore the fundamental nature of a convolutional neural network, give some examples of processing images in python and discuss some research applications of this exciting field.
