Unit overview
What you will learn in this unit
Unit 1 is entirely conceptual. There is no code to write and no model to train. The goal is to give you the mental model and the vocabulary you need before you start fitting algorithms to data — vocabulary that will be assumed from Unit 2 onward.
Learning outcomes
By the end of this unit you will be able to
- Define machine learning and distinguish it from classical rule-based programming.
- Describe the historical development of artificial intelligence and identify the shifts that led to modern statistical learning.
- Distinguish AI, machine learning, and deep learning, and explain how the three relate.
- Explain at a high level how a learning algorithm improves its performance from data.
- Identify which kinds of problems machine learning is well-suited to, and which it is not.
- Recognise common failure modes, including overfitting, distribution shift, and spurious correlation.
- Describe the stages of the machine learning lifecycle, from problem framing to deployment and monitoring.
No code in this unit
Unit 1 is about ideas. If you came here expecting to run your first model, that happens in Unit 3. The time you invest in the concepts here pays off immediately from Unit 4 onward, when algorithms are introduced and their behaviour only makes sense against the framing you build now.
Tutorials
Nine tutorials, to be worked through in order. Each follows the same nine-part structure: objectives, motivating example, concept explanation, visual illustration, worked example, interactive activity, self-check, summary, and further reading.
What is Machine Learning?
The core idea: instead of writing rules, we let a system infer them from examples.
T-02A Brief History of Artificial Intelligence
From Turing’s question through two AI winters to the deep learning era.
T-03From Symbolic AI to Statistical Learning
Why the field abandoned hand-coded rules in favour of learning from data.
T-04AI, Machine Learning, and Deep Learning
The nested relationship between three terms that are often used interchangeably.
T-05How a Learning Algorithm Works
Features, labels, parameters, and loss — the four ideas behind every algorithm.
T-06Applications of Machine Learning
Where the field is actually deployed, from recommendation to medicine to fraud detection.
T-07Limitations and Failure Modes
When machine learning fails, and why. Overfitting, drift, and the danger of correlation.
T-08The Machine Learning Lifecycle
The eight stages a real project passes through, and why most of them are not modelling.
T-09Unit Review and Self-Assessment
A structured review of the unit and a short quiz to check your understanding.
How to work through this unit
Unit 1 is reading-heavy. The habits that matter here are different from those you will need later in the course.
-
Read in order
Each tutorial builds on the one before. T-04 assumes you have read T-02 and T-03, and T-08 assumes you have read T-05.
-
Take notes on the vocabulary
The terms introduced here — feature, label, parameter, loss, overfitting — are used without re-definition from Unit 4 onward.
-
Answer the self-checks honestly
Each tutorial ends with a set of self-check questions. If you cannot answer without looking back, re-read that section before continuing.
-
Finish with the unit quiz
The quiz at the end of T-09 is short and diagnostic. Score it honestly before looking at the feedback.
Assessment in this unit
Unit 1 contributes one unit quiz to your final grade. There is no assignment in this unit — the first assignment is released in Module 2, after you have started building models.
- Quiz 1 Introduction to Machine Learning — 12 questions, unweighted diagnostic Go to T-09
Looking ahead
Unit 2 shifts from ideas to data. The vocabulary you build here — features, labels, and the distinction between training and testing — is used directly in the next unit, where you will work with real datasets for the first time.