Checklist ROAI 2024 County Stage · Task 2
Points Classifier
Romanian title: Problema 2
Learn an unknown axis-parallel rectangle from labelled 2D points in three scenarios with different training-set sizes.
The task
Points (X, Y) with 0 ≤ X, Y ≤ 100 are labelled 1 if they lie inside an unknown axis-parallel rectangle (boundary included) and 0 otherwise. From labelled training points, the contestant must build a classifier for any point of the square.
There are three independent scenarios, each hosted as a Kaggle competition and possibly using a different rectangle: 100, 1,000 and 5,000 training examples, each with 5,000 test examples.
Abridged and translated by SOTA from the official Romanian materials. The official statement has the exact rules, and it wins wherever this summary differs.
In English
This task was published in Romanian. SOTA translated it into English on 16 September 2026.
Read the task statement in English
Points Classifier
English translation by SOTA – AI Community of the Romanian original, published with the permission of Asociația Nitro. Organisers who would like this translation removed can email [email protected].
Problem 2
In this problem you will work with data representing points in the plane, specified by their coordinates (X,Y) with 0 <= X,Y <= 100. In the same region of the plane, [0, 100] x [0, 100], there is a rectangle with sides parallel to the axes, but whose coordinates are not known. Any point in the region [0, 100] x [0, 100] will be classified as positive (label 1) if it lies inside this rectangle (including on its sides) and as negative (label 0) if it lies outside that rectangle.
[Figure: see the original statement. It shows the square region with an axis-parallel rectangle inside it; the points inside the rectangle are marked "+" and the points outside it are marked "−".]
Your task is to start from a set of training examples (points labelled 0 or 1) and build a classifier that labels any point in the region [0, 100] x [0, 100] as positive (label 1) or negative (label 0) according to the mechanism described above. The performance of the classifier will be given by its accuracy on a set of test examples.
You will have to investigate three scenarios / situations, which will differ in the number of training examples. Each of the three scenarios will be a problem on Kaggle. The three scenarios have nothing in common except the mechanism described above, and the rectangle that performs the classification will not necessarily be the same.
You can find the technical details about the format of the training, test and solution files on Kaggle:
Scenario 1 (100 training examples, 5000 test examples): https://tinyurl.com/ROAI-scenario-1
Scenario 2 (1000 training examples, 5000 test examples): https://tinyurl.com/ROAI-scenario-2
Scenario 3 (5000 training examples, 5000 test examples): https://tinyurl.com/ROAI-scenario-3
Translated by SOTA. The Romanian original is the official version and wins wherever the two differ. The original is titled only "Problema 2". The data and file formats are on the three Kaggle pages behind the scenario links, which redirect to Kaggle invitation links. Published with the permission of Asociația Nitro, which runs ROAI (17 September 2026). If you organise this olympiad and would like the translation removed, email [email protected] and we will take it down.
At a glance
- You get
- Training and test point files per scenario, distributed through Kaggle (file formats described on Kaggle).
- You submit
- Predicted labels for the test points in the Kaggle submission format.
- Scoring
- Classification accuracy on the test set.
- Rules
- 4-hour test, common paper for all grades (per the stage page)
- Format
- ROAI 2024 county stage, 28 June 2024.