Machine Learning Basics
What is machine learning?
A type of AI that learns patterns from data without being explicitly programmed for every task.
Supervised learning uses __________ data.
Labelled
Does unsupervised learning use labelled or unlabelled data?
Unlabelled
Reinforcement learning learns through __________ and __________.
Rewards and Penalties
Name ONE app you probably used today that uses machine learning.
YouTube
TikTok
Spotify
Google Maps
Netflix
ChatGPT
Siri
Which comes first?
A) Prediction
B) Training Data
C) Feedback
B) Training Data
Which of these is supervised learning?
A) Gmail Spam Filter
B) Spotify Recommendations
C) Robot Learning to Walk
A) Gmail Spam Filter
What is the main goal of unsupervised learning?
A. To make predictions
B. To discover hidden patterns
C. To copy human decisions
D. To receive rewards
B. To discover hidden patterns
Which everyday example best explains reinforcement learning?
A) Teacher marking homework
B) Training a dog with treats
C) Organizing books by colour
B) Training a dog with treats
Google Maps suggests a faster route because traffic has changed.
What technology helps make these predictions?
Machine Learning
True or False:
The more high-quality data a machine learning model is trained on, the better its predictions are likely to be.
True
Your teacher shows you hundreds of pictures labelled "Cat" and "Dog."
What type of learning is this?
Supervised Learning
Instead of being given answers, what does unsupervised learning try to discover?
Hidden patterns or groups.
A robot is learning to play football.
After scoring a goal it receives +10 points.
What is the purpose of these points?
A. To store data
B. To tell the robot which actions are better
C. To increase the robot's speed
D. To organize information
B. To tell the robot which actions are better.
Face ID unlocks your phone by recognizing your face.
Which application of machine learning is this?
Facial Recognition
A machine learning model mistakes every wolf for a dog.
What is the MOST likely reason?
A) Poor training data
B) Internet connection
C) Computer virus
A) Poor training data
Why is Gmail's spam filter considered supervised learning?
Because it is trained using emails that have already been labelled as "Spam" or "Not Spam."
An online shopping website has information about customers' ages, locations and buying habits, but no labels describing the customers.
What useful insight could an unsupervised learning model provide?
It could group customers into similar segments (e.g., frequent buyers, luxury shoppers, bargain hunters) based on patterns in their behaviour.
Why does reinforcement learning usually improve over time?
Because the AI learns from the rewards and penalties it receives after each action.
KFC wants to know whether customers like its new burger by analysing thousands of online reviews.
What machine learning application is it using?
Sentiment Analysis
Why is machine learning more useful than manually programming every possible situation?
Because it can learn patterns from large amounts of data and adapt to new situations that programmers cannot predict individually.
A hospital is creating an AI to detect pneumonia from chest X-rays. Before the AI can make accurate predictions, what is the most important thing doctors need to provide?
A. More computers
B. Thousands of correctly labelled X-rays
C. Faster internet
D. More patients
B. Thousands of correctly labelled X-rays
Why do companies often use unsupervised learning before launching a marketing campaign?
To discover hidden customer groups so they can target each group more effectively with personalized products or advertisements.
Name one real-world application of reinforcement learning and explain why it uses this approach
Examples:
Because they learn the best actions through trial and error.
A company wants to create an app that automatically recommends the best restaurants for each user based on their previous searches, ratings, and dining history.
Name TWO types of machine learning the app could use and explain what each would do.