Artificial Intelligence: Question 5
Syllabus 18.1
An agricultural technology company is building three different AI systems. For each system, state which category of machine learning is being used, supervised learning, unsupervised learning or reinforcement learning, and give one reason for your choice.
(a) System 1 predicts a crop's expected yield, in tonnes per hectare, from historical records of rainfall and soil quality, where every past record already states the yield that was actually achieved. [2]
(b) System 2 sorts thousands of unlabelled satellite images of fields into groups of similar-looking images, without being told in advance what the groups should be or how many groups to form. [2]
(c) System 3 controls a crop-spraying drone that chooses its own flight actions. It receives a positive reward when it applies fertiliser evenly across a field, and a penalty when it wastes fertiliser or misses part of the field, without being told the correct action for any given situation in advance. [2]
Show worked solution Hide worked solution
Worked solution
Part (a): System 1
Supervised learning. Every past record already includes the input data (rainfall, soil quality) and the known, correct output (the yield actually achieved). This is labelled data, which is exactly what supervised learning trains on.
This is specifically a regression method, because the system predicts a continuous numeric value (tonnes per hectare) rather than choosing between a small set of named categories.
[2 marks]: [1] for correctly identifying supervised learning and justifying it with the presence of a known, correct yield label in every past record, [1] for identifying that this is a regression task, since the predicted yield is a continuous numeric value.
Part (b): System 2
Unsupervised learning. The satellite images are explicitly described as unlabelled. The system is not given any known, correct grouping in advance, and instead must find structure (similar-looking images) in the data by itself.
[2 marks]: [1] for correctly identifying unsupervised learning, [1] for justifying it with the fact that the images are unlabelled and no predefined groups are given.
Part (c): System 3
Reinforcement learning. The drone (acting as an agent) chooses its own actions and learns from reward and penalty feedback received as a result of those actions, through trial and error, rather than being trained on a fixed set of examples that already show the correct action for each situation.
[2 marks]: [1] for correctly identifying reinforcement learning, [1] for justifying it with the reward/penalty feedback received for the drone’s own chosen actions.
Final answers
- (a) Supervised learning (specifically regression). Every past record already has a known yield label, and the prediction is a continuous number
- (b) Unsupervised learning. The images are unlabelled with no predefined groups
- (c) Reinforcement learning, the drone learns from reward/penalty feedback on its own actions