Checklist ROAI 2026 National Stage, Grades 11–12 · Task 1
What does the fox say?
Count the vocabulary of an invented language corpus and classify its messages into four hidden topics.
The task
A fox left the Little Prince a journal written in an invented 'vulpine' language. The messages address four hidden topics that can only be inferred from context and word repetition.
Subtask 1 asks for the number of distinct words in the unlabelled corpus. Subtask 2 asks for the topic (0–3) of each test message.
Abridged by SOTA from the official materials. The official statement has the exact rules, and it wins wherever this summary differs.
At a glance
- You get
train_data.csv(1,000 labelled messages),test_data.csv(500 messages) andunlabeled_data.csv(thousands of unlabelled sentences, supplied with the starter kit).- You submit
- CSV with subtaskID, datapointID (0 for subtask 1, message ID for subtask 2) and answer.
- Scoring
- Subtask 1 (10 points): exact count. Subtask 2 (90 points): F1; 90 points at F1 ≥ 0.92, linear from 5 to 90 for 0.25 ≤ F1 < 0.92, 5 points below 0.25.
- Rules
- On-site at Genesis College, Bucharest, through the Nitro Contestant Cloud
- Only the packages listed in the round rules may be used
- Only submissions selected by the contestant count towards the final ranking
- Format
- ROAI 2026 national stage, grades 11–12 section, 7 April 2026 (07:15–11:25 UTC on the judge).