Before we start digging into the fundamentals of Deep Learning, let’s take a look at the tools we will be using going forward.

1. Google Colaboratory : You can run your code on GPUs hosted by Google for free.
2. Seedbank : Collection of interactive machine learning examples for Colab or Jupyter.
3. Create a profile on Kaggle, if are not already signed up for it.

Deep learning
|-- Supervised
|     |-- MLP:  Multilayer perceptron
|     |-- CNN:  Convolutional Neural Networks
|     |-- RNN:  Recurrent Neural Networks
|     |-- RNN:  Recursive Neural Networks
|     |-- LSTM: Long short-term memory
|     |-- SQS:  Sequence-to-sequence models
|     |-- SNN:  Shallow Neural Networks
|-- Unsupervised
      |-- SOM:  Self Organizing Maps
      |-- BM:   Boltzmann Machines
      |-- AE:   AutoEncoders

Supervised Deep learning

DL form where an output label exists for every input example. The labels are used to compare the output of Deep Neural Nets to the ground-truth values and minimize the cost function. Other forms of Deep learning are semi-supervised and Unsupervised.

Smartest and most complex thing on this planet is human brain and with deep leaning we try to mimic how human brain works by creating “Artificial Neural Networks”.

What is an ANN?

Artificial Neural Network (ANN) is

1. CNN: Convolutional Neural Networks
2. RNN: Recurrent Neural Networks

Unsupervised Deep Learning

1. SOM: Self Organizing Maps
2. BM: Boltzmann Machines
3. AE: AutoEncoders

Some useful resources

1. See in action: Tensorflow Playground
2. Flashcards of the interview: Deep Learning Dictionary