Introduction to week 2

In the first week we had a broader look at the main differences between traditional machine learning and deep learning, and introduced PyTorch and Colab, the main software tools we will use during the course. This week we will look in detail at the network type at the heart of deep learning, convolutional neural networks (CNNs). We’ll begin by looking at the key components in turn, before seeing how they can be connected together to make a network.

The week has been divided into the following activities:

• network components
  • convolutional layers
  • max pooling layers
  • activation functions
• tensors and dimensionality – moving data between network layers
• a simple CNN – how to connect the components together to make a working network.

We will begin in the next step with a video giving an overview of the different layers found in CNNs.

Week 2 learning outcomes

• Explain each of the key components of convolutional neural networks
• Create a simple convolutional neural network using PyTorch