Machine learning means a system that improves at a task from data rather than from rules someone wrote. The guide names supervised, unsupervised, reinforcement and deep learning.
The same problem given to each kind of learning, showing what each one is handed and what it returns.
Interactive diagram
Explore the labelled diagram, charts and maps for this topic in full study mode.
Free preview
This is the free notes preview
You're reading the free notes. Aimnova Pro unlocks the full study experience — and you can try it with your first topic free to keep:
- FlashcardsLock in vocabulary and key terms with spaced repetition.
- Practice questionsAnswer exam-style questions and get instant AI marking.
- Mock exams & past-paper vaultSit full mocks and see exactly how examiners award marks.
- Personalised study planA daily plan built around your exam date and weak areas.
| Kind | Given | Returns | Where bias enters |
|---|---|---|---|
| Supervised | Examples with labels | A label for new cases | Whoever wrote the labels |
| Unsupervised | Data with no labels | Groups, unnamed | Whoever interprets the groups |
| Reinforcement | A goal and a score | A strategy | Whoever chose what to score |
| Deep learning | Any of the above | The same, less explainable | All of the above, harder to see |
Deep learning is not a fourth kind: “Deep” describes the shape of the network — many layers. A deep model can be supervised, unsupervised or reinforcement. Listing it as a fourth type is a common error.
Memorize terms 3x faster
Smart flashcards show you cards right before you forget them. Perfect for definitions and key concepts.
The uses the guide names
- Pattern recognition — finding structure in something messy.
- Facial and speech recognition — matching an input to an identity.
- Image analysis — reading a scan, a satellite photo, a production line.
- Natural language processing — translating, summarising, answering, generating.
Real-world examples you can name
Amazon's scrapped CV-screening tool — reported October 2018
An experimental hiring tool was trained on a decade of CVs submitted to the company, most of them from men. It learned to downgrade CVs containing the word 'women's' and graduates of two women's colleges. The project was abandoned.
Who it affected: Women applying for technical roles, and every organization that assumed historical data was neutral.
AlphaFold and the protein structure database — AlphaFold 2 published 2021; over 200 million predicted structures released July 2022
A machine-learning system that predicts a protein's three-dimensional shape from its amino-acid sequence, a problem that had resisted fifty years of work. The predictions were released openly rather than licensed.
Who it affected: Biologists worldwide, including laboratories that could never afford the equipment to determine structures experimentally.
San Francisco's facial recognition ban — May 2019
The first major US city to bar its own agencies, including the police, from using facial recognition — a rule about government use, not about the technology existing. Several other cities followed, and a few later relaxed theirs.
Who it affected: Residents, and police forces that had to find other methods.
How this is tested — you have to explain why a model needs a lot of data, and what follows from where that data came from. It comes up two ways:
Paper 1 — structured question
- Part b: explain why a model needs a large training set
- Part c: evaluate a system trained on historical decisions
Paper 2 — source-based question
- Q2: explain a claim about how a system was trained
- Q4: synthesise sources on an automated system, and Paper 3
The trap: “the algorithm is biased”: Say WHERE it came from: the labels, the missing group, the stand-in measure, or the target chosen. Naming the source is the difference between an assertion and an explanation.
Explain one reason why a machine-learning model needs a large amount of training data.
Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.
Discuss the use of machine learning to screen job applications.
Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.