ENGSCI 304: Engineering Machine Learning II

Lecture notes

About these notes

These notes cover my half of ENGSCI 304: Engineering Machine Learning II. Here we first look at visual data, i.e. images and video, followed by text data, language models and agents. The material builds on previous machine-learning material in ENGSCI 205 to some extent but missing background can be filled in relatively quickly I think.

To begin, our main visual example is tracking players in basketball footage. We consider image representation, using pre-trained vision models, camera geometry, evaluation of predictions and, later, tracking.

Other resources

Lab notebooks, supplied files and assessment information are provided separately through Canvas.

A list of recommended readings is also provided.

Updates

These notes are still developing somewhat. Chapters will be updated and released as we progress; the contents list shows the material currently available.