AIF-C01 ยท AI and ML basics
Three ways a model learns
How the three learning types differ, what data each needs and a typical example of each.
Key points
- Supervised learning trains on labeled examples to predict a label or a number (classification, regression), for instance spam detection.
- Unsupervised learning finds structure in unlabeled data, for instance clustering customers into segments.
- Reinforcement learning has an agent try actions in an environment and learn from rewards and penalties, for instance a robot learning to walk.
Practice questions
Which learning type trains on examples that each include the correct answer?
- Unsupervised learning
- Reinforcement learning
- Supervised learning
- Clustering only
Show answer
Answer: C. Supervised learning
Supervised learning uses labeled examples. Unsupervised works without labels, and reinforcement learns from rewards.
Which statements describe reinforcement learning?
- An agent learns by trying actions in an environment
- A reward or penalty signals how good an action was
- Every training example must have a correct label
- It is used for tasks like game playing or robot control
- It only groups similar records together
Show answer
Answer: A. An agent learns by trying actions in an environment, B. A reward or penalty signals how good an action was, D. It is used for tasks like game playing or robot control
Reinforcement learning uses an agent, an environment and a reward signal to learn a behavior over repeated trials. It does not need a labeled answer for every example.
True or false: unsupervised learning requires a label for every training example.
Show answer
Answer: False
False. Unsupervised learning works on unlabeled data and looks for structure such as clusters. Needing labels is the mark of supervised learning.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
