Building Generative Adversarial Networks in Tensorflow and Keras

Abstract: Generative Adversarial Networks are a promising modern application of Deep Learning that allows models to *generate* examples. However, GANs are complex, difficult to tune, and limited to small examples. We will explore recent GAN progress with a model that generates faces conditional on desired features, like 'smiling' and 'bangs'.

This workshop is designed for Data Scientists, researchers, and software developers familiar with keras, tensorflow, or similar recent Deep Learning tools. It is expected that most in the audience will be able to build models and begin to train them on a local machine. Such students will not leave the tutorial with fully trained models. While students are not expected to have remote access to a machine configured with CUDA and tensorflow-gpu, the instructor will.
After attending, students in the target audience should be able to - Identify and explain the essential components of Generative Adversarial Networks including Deep Convolutional versions. - Modify existing GAN implementations. - Design a GAN for a novel application. - Understand and explain recent improvements in GAN loss functions.

Bio: Sophie is a Senior Data Scientist at Metis where she is a bootcamp instructor and leads curriculum development. Sophie works in deep learning and data science ethics. Through t4tech she helps provide free trans-centered classes in programming and data science. She holds masters degrees in Electrical and Computer Engineering and Psychology, and her writing has appeared in Information Week. Sophie is passionate about teaching, both in theory and in practice, and about making sure that data science is primarily a tool that is used to improve people's lives.