Almost all current day DL work happens using Deep Learning frameworks. With so many around, it can get confusing what they are all trying to do. In this talk, I will be discussing how these frameworks work under the hood, how they all use the same concepts and how you could roll your own framework if you desire. I will be covering both high-level mathematical concepts as well as low-level implementation details in an accessible way.
Gokul is currently a Master's student in CS at ETH Zurich, focusing on Machine Learning and Deep Learning. Gokul did hisbachelor's from BITS Pilani. Currently a research assistant in the DS3Lab, his current work revolves around GANs, on how to use them in astronomy as well as methods to train them. Gokul have been a contributor to Theano as well. Gokul is currently looking for internships, if you like his talk and want to hire him, You can find more details about Gokul at :