What is this post about?
In this super gentle 101 talk given at the Insight Centre for Data Analytics at University College Dublin , I have gone through some of the most mportant aspects of Artificial Neural Networks (ANNs):
Here are what you will learn in this talk along with the corresponding time-stamps:
00:33 Table of contents
03:04 The biological motivation
03:32 The big picture of what we expect from an ANN
04:30 An actual nervous cell and an aritificial perceptron: Some similarities
10:37 A giant biological neural net and a giant artificial neural network: A face off!
14:28 ANNs as Truth approximators: What is the Truth?
16:21 ANNs as Truth approximators: Examples of such Truth approximations
17:49 How well/poorly is my ANN performing? The idea of Error functions
20:54 The training cycle of an ANN: Epochs, Backpropagation and tuning the parameters of an ANN
25:27 What isthe backpropagation algorithm? What is a gradient? How does an ANN use them to tune its parameters?
28:56 Concluding thoughts