Overview

This part will focus on variants of Neural Network(NN) which are DAGs(Directed Acyclic Graph).

3 common types of network architectures will be covered. They are:

Feedforward NN:

1. Architecture and Computation;
2. Learning algorithms;
3. Regularization methods;

Convolutional NN:

1. Motivation and Architecture;
2. Interesting variants.

Recurrent NN:

1. Architecture and unfolding;
2. Learning problems and methods;
3. Memory.

For each section of a particular network architecture, there will be a following part "Historical Remarks" mentioned its history, some important workes and recent publications for future short term trend.