Introduction to Week 3

In the first two weeks we were mainly concerned with data acquisition and getting the data in suitable shape for use in machine learning and deep learning models.

This week we will look more closely at the processes involved in implementing such models. We won’t talk in any detail about the algorithms themselves – that is covered in other courses (see the links below). Instead, we will talk in more general terms about things to think about at key stages of machine learning and deep learning projects.

The week has been divided into the following activities:

• software, model selection and training
  • choosing a software environment
  • model selection and transfer learning
  • training and inference
• improving performance
  • troubleshooting and debugging
  • hyperparameter tuning
  • multitask learning.

We will begin in the next step with a video which discusses what software environments are most commonly used for the machine learning and deep learning tasks we are interested in.

Week 3 learning outcome

• Compare and select machine learning and deep learning models for use with your experimental data