Lesson 2.2 · 21 min · video lesson
Labeled vs Unlabeled: Two Ways Machines Learn
If we have a million customer records but nobody has labelled a single one, can a machine still learn something useful from them?
In short: Supervised learning learns from examples that come with the correct answer (a label) and then predicts that answer for new inputs. Unsupervised learning gets no answers at all and instead discovers structure on its own, such as groups, unusual points, or simpler representations. The choice depends mostly on whether we have labels and whether we know exactly what we want to predict.
What you will learn
- Supervised Learning
- Unsupervised Learning
- Differences Between Supervised and Unsupervised Learning
In this lesson
- Two ways to learn from data
- Supervised learning
- Unsupervised learning
- Code: the same data, both ways
- Differences between supervised and unsupervised learning
- The space in between: semi-supervised and self-supervised
- Real-world use and how to choose
- Common mistakes and how to spot them
- Practice: try it yourself
Key terms in this lesson
- Supervised learning
- Learning a mapping from inputs to outputs using examples that include the correct output.
- Unsupervised learning
- Learning patterns or structure from data that has no labels.
- Label
- The correct answer attached to a training example.
- Classification
- A supervised task where the output is a category.
- Clustering
- Grouping examples so that similar ones end up in the same group.
- k-means
- A clustering algorithm that alternates assigning points to the nearest of k centres and moving centres to the mean of their points.
- Self-supervised learning
- Training where labels are generated automatically from the data itself, such as predicting a hidden next word.
This lesson is free. Sign in or create a free account to open it with its video and quiz.
← 2.1 Machine Learning from First Principles · 2.3 Predicting Numbers vs Categories: Regression Compared →