Generative Adversarial Networks using Apache MXNet

Abstract:
Generative Adversarial Networks (GANs) are deep learning based generative models which focus on creating data from scratch by pitting 2 neural networks against each other in a zero-sum game. GANs were introduced by Ian Goodfellow et al. in 2014, and since, have been used in a variety of applications ranging from generating anime characters to generating videos from captions.

In this session, we will go over the basics of GAN and how it works. We will also dive deep into one of the more popular designs for GAN, Deep Convolutional GAN (DCGAN) with the help of a code demonstration using Apache MXNet.

Bio:
Vandana Kannan is a Software Development Engineer with Amazon AI, contributing to Apache MXNet, one of the most scalable and developer-friendly Deep Learning frameworks. Recently, she has been working on providing support for ONNX on MXNet. She holds a Master of Science degree in Computer Science from San José State University, where her research focus was on Genetic Algorithms.

Piyush is an SDE at Amazon AI. He graduated with a master's in CS, focusing on Machine Learning and NLP from Ohio State University in Spring 2018. At Amazon, Piyush is focused on developing solutions to simplify the use and adoption of Deep Learning among novice to advanced practitioners of deep learning.

Agenda:
6pm – doors open & socializing
7-8pm – presentation
8pm – Q&A and socializing
9pm – closing