ITPEC FE Morning October 2020 Question 5

Training on inputs paired with their correct outputs — the known answers act as the supervisor, and the model learns to predict outputs for unseen inputs.

In supervised learning each training example is labelled: an input and the output it should produce. The model is adjusted to reduce the gap between its predictions and those labels, so it generalises to new data. Classification and regression are the typical tasks.

Answer (d)

Why not others:

• •(a) — grouping data by its own statistical features with no correct answers given is unsupervised learning (clustering)
• •(b) — learning to maximise a score from rewards and penalties for actions is reinforcement learning
• •(c) — predicting the next action from how often past behaviour occurred is frequency-based prediction, as in recommendation or association analysis; no labelled input-output pairs are used

Key rule: Supervised learning has labelled answers, unsupervised learning has none, and reinforcement learning has only a reward signal for its actions.

Basic Theory · AI-written, may contain errors