Discord

Checklist AICC 2026 Round 10 · Task 3

Moment Match

Identify which of four candidate sentence sequences from a second annotator describes the same action as a target sentence, using fixed fastText word embeddings.

  • NLP
  • Multiple-choice text sequence matching
Read the official statement

The task

Two annotators independently described the same sequence of everyday activities as short sentences. The same action may be phrased differently, and the annotators describe actions at different levels of detail, so their sequences do not align exactly.

Each sample gives 5 consecutive sentences from annotator A, whose middle sentence (index 2) is the target, and 4 candidate sequences of 9 consecutive sentences from annotator B. Exactly one candidate contains annotator B's description of the target action, at any position among its 9 sentences. Because the target alone may be ambiguous, the surrounding sentences carry the information needed to find the match.

Contestants predict, for each test sample, the index (0–3) of the correct candidate. Only the supplied fastText embeddings may be used as pretrained word representations.

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
texts_train.json, texts_validation.json, texts_test.json (target: 5 sentences; candidates: 4 × 9 sentences); labels_train.npy and labels_validation.npy (integers 0–3); embedding_matrix.npy (float32, 300-dimensional fastText vectors) and vocab.json.
You submit
submission.csv with columns row_id,label; row_id follows the test order from 0 to 1969 and label is an integer 0–3.
Scoring
Accuracy. Baseline 0.43, reference solution 0.70.
Rules
  • No additional pretrained models may be used to generate or obtain embeddings (the provided embeddings were generated with fastText).
  • Individual participation (maximum team size 1); at most 60 submissions per day.
  • AICC contest rules (stated on each Kaggle rules page, not enforceable): no use of LLMs for writing code or getting task ideas; no internet use other than official library documentation and the contest platform; no communication with anyone during the contest; clarifications only via the #clarification-requests channel on the AICC Discord server.
Format
AICC Round 10, online on Kaggle, 28 Aug 2026 18:00 UTC – 30 Aug 2026 18:00 UTC.

Notebook

Competition

Solution

Details

Year
2026, Online (Kaggle)
Round
Round 10 · Task 3
Language
English
License
Varies by task: What I Want and What I Don't — MIT; Inverse Translation and Moment Match — CC BY-NC-ND 4.0 (Kaggle competition licences). Solutions repository: MIT., as stated by the source