Introduction to Neural Networks

Continuing from yesterdays' study day, where you learnt about Perceptron (a single neuron), its time to stack these neurons and form a network of neurons.

For the day, you will be going through a series of video tutorial to learn about the concepts of building a neural network. (The tutorial is quite basic, so feel free to skip through parts of video if you already know the topics…)

1. Understanding the concepts of Neural Networks

Go through this LinkedIn learning course : Neural Networks in Python (2-3 hours) - mostly only the video tutorial parts. You can skip the python coding related tasks from this tutorial, as we will be going through them in the afternoon and tomorrow in DataLab.

These are the required chapters from the tutorial:

• Introduction

1) Choosing a Neural Network

2) The building Blocks of Neural Network

3) Building your network

4) Training your network

5) Segment Display Network [Optional]

LUNCH !!

2. Build a NN (Neural Network) in Python

1) watch the video to better know the problem statement for this tutorial and how to visualize them.

2) Learn about predictions and activation function

3) watch the video on creating your first neural network in python. (For Today, you can follow the video only, tomorrow in the datalab you will be actually implementing them)

4) [Optional] watch the next 3 video from the video series to learn about squared error between predicted value and actual value. How to compute cost function, gradient of cost function to eventually reduce the error in predictions.

in case the maths regarding differential calculus on how to calculate a slope was difficult to follow, feel free to check out the next video. Luckily, we don't need to understand all the maths, sometimes we can get along by knowing how to use maths.

5) Linear regression - watch the next set of 3 videos on how you can apply the concepts.

6) Finally solve the Flower problem !! watch here: