Introduction to Deep Learning

Topics Covered

• Introduction to Neural Networks and Deep Learning
• Building a simple neural network from first principles
• Introduction to Backpropagation algorithm
• Multi-layer perceptron
• Convolution Neural Networks
  • Introduction to Convolution
  • Image Recognition using CNN
• Natural Language Processing :
  • Introduction to word2vec
  • Introduction to Recurrent Neural Networks
  • Text classification using RNN
  • Text generation using RNN
• Unsupervised learning using Autoencoders

Depending on time, some of the topics may not be covered during the workshop. But, please note that the entire content(data and source code in ipython notebook format) would be available in this repository.

Slides for the workshop

https://speakerdeck.com/bargava/introduction-to-deep-learning

Setup Guide

Pre-requisites: git, python 2.7.X, virtualenv, pip (7.1.X recommended)

• If you're using Ubuntu, here are all the packages you'll need before you can proceed

  $ sudo apt-get install python2.7 python-dev build-essential curl libatlas-base-dev gfortran
  $ sudo apt-get install libfreetype6-dev libpng-dev libjpeg-dev
  

• Clone the repo from GitHub

  $ git clone https://github.com/rouseguy/intro2deeplearning.git
  $ cd intro2deeplearning
  

• Create python virtual environment

  $ virtualenv env
  $ source env/bin/activate
  

• Install requirements using pip

  $ pip install -r requirements.txt
  

  Use requirements_linux.txt instead of requirements.txt if you're on linux

• When the requirements are being downloaded / installed, Fetch the datasets simultaneously

  $ sh download_data.sh
  

• Run check_env.py script to test the dependencies

  $ python check_env.py
  

  Output should look like this

  [ OK ] scipy version 0.15.1
  [ OK ] PIL version 1.1.7
  [ OK ] keras
  [ OK ] IPython version 4.0.0
  [ OK ] theano version 0.7.0
  [ OK ] numpy version 1.9.2
  [ OK ] pandas version 0.16.2
  [ OK ] gensim version 0.10.3
  [ OK ] sklearn version 0.16.1
  

  This means you have all the dependencies installed and you're ready to start.

• Run the notebook

  $ cd notebooks
  $ ipython notebook
  

  This opens your default browser which displays the list of notebooks in the current directory.

  Open 1. Introduction to Artificial Neural Networks.ipynb. Now, run the first cell with imports in the notebook (shift + enter). If you have all the dependencies installed, this should run without any errors.

Note: We only support Ubuntu Linux (Tested) & OSX environments. We strongly recommend Windows users to have a VM running Linux, and then install these requirements on that VM.