Deep Learning 101

Derrick Joseph (~derrick)


11

Votes

Description:

Most of the time we hear buzzwords like Recurrent Neural Nets, LSTM etc But don't know what is all that jargon and how or why these things work, This talk will give you a decent introduction to Deep Learning (No boring history lessons or the typical "neurons in brain" talk). The agenda is to start by building, a simple Deep Learning Neural Network with one hidden layer, in core python(No frameworks) then move on to give an introduction on Convolutional Neural Nets(these are really good for object detection in images) and finally, end by building a Recurrent Neural Net . The goal is to give you a soft launch into the Deep Learning, from there on you should be able to fly with your own imaginations.

Topics going to be covered are:

  • A single Neural Net learning to do an XOR.
  • A ConvNet.
  • An RNN learning to do a Sequence to Sequence operation(just the fancy term for Addition, Subtraction etc).

Prerequisites:

  • Since this talk is being delivered at Pycon It would be nice of you to know some Python.
  • Also some idea of the traditional ML could to be useful.

Speaker Info:

I am a Product Engineer at Profoundis(Now FullContact). We built Vibe, A B2B Data Intelligence solution for Sales and Marketing. I started off with Data Sciences during my Post Graduate Days and have been an ML enthusiast ever since.

Speaker Links:

LinkedIn

Section: Scientific Computing
Type: Talks
Target Audience: Beginner
Last Updated: