by

LESSON

AI 015. What are supervised and unsupervised learning?

listen to the answer

ANSWER

Supervised and unsupervised learning are two core approaches in machine learning, each with its own methods and applications.

Supervised Learning

Think of supervised learning like learning to paint with the help of an instructor. You’re given a set of example paintings (data) and their styles (labels), such as impressionism, surrealism, or cubism. Your goal is to learn these styles so well that when presented with a new painting, you can identify its style correctly. In supervised learning, the “instructor” is the dataset with known labels. The model is trained on this dataset until it can accurately predict the label of new, unseen data. This method is used for classification tasks (where the output is a category, like spam or not spam) and regression tasks (where the output is a quantity, like the price of a house).

Unsupervised Learning

Unsupervised learning, on the other hand, is like exploring an art gallery without a guide. You’re not told what to look for; instead, you observe and group paintings based on your own observations of similarities and differences. There are no labels here; the goal is to uncover hidden patterns or groupings in the data. This could mean clustering paintings by similar themes or styles that you notice, without being told any categories upfront. Unsupervised learning is useful for exploratory analysis, clustering (grouping similar things together), and dimensionality reduction (simplifying data without losing its essence).

Read more

Quiz

What is supervised learning in machine learning?
A) Learning with data that is not labeled and finding hidden patterns
C) Learning without any data
B) Learning with labeled data to predict the label of new data
D) Learning only specific tasks that are programmed by humans
The correct answer is B
The correct answer is B
What does unsupervised learning primarily focus on?
A) Following explicit instructions to perform tasks
C) Exploring data to find hidden structures or patterns
B) Using labeled data to perform classification
D) Predicting specific outcomes based on past data
The correct answer is C
The correct answer is C
Which of the following is an application of unsupervised learning?
A) Predicting the stock market
C) Diagnosing diseases based on symptoms
B) Clustering data into meaningful groups without prior labels
D) Calculating loan eligibility
The correct answer is C
The correct answer is B

Analogy

Imagine machine learning as learning about art:

Supervised learning is your art class where you learn to paint under the guidance of your teacher, who shows you different styles and techniques, and you learn by trying to replicate them. Your progress is checked against known standards (labeled data), helping you refine your skills.

Unsupervised learning is your solo visit to an art museum, where you explore different artworks on your own, grouping them by perceived similarities or themes without a predefined rulebook. It’s a journey of discovery and insight, where you learn to see the art in a new light based on patterns and connections you observe yourself.

Read more

Dilemmas

Data Labeling Challenges in Supervised Learning: Given the significant reliance on accurately labeled data for training in supervised learning, what strategies can be adopted to handle incorrectly labeled data which could lead to poorly performing models?
Ethics of Pattern Discovery in Unsupervised Learning: Considering unsupervised learning can discover hidden patterns and groups within data, what ethical considerations should be made to avoid invasions of privacy or unintended discriminatory practices?
Balance of Control in Machine Learning: How can developers balance the autonomy given to unsupervised learning algorithms, which operate without explicit guidance, with the need for accurate and accountable AI systems?

Subscribe to our newsletter.