Presented at the 25th International Conference on Artificial Intelligence in Education (AIED 2024), Late-Breaking Results track, Recife, Brazil.
The paper presents the first Turkish crossword puzzle generator based on LLMs, with two new datasets: over 180,000 answer-clue pairs and over 35,000 samples that link text, answer, category and clue. The tool is aimed at classroom use to support vocabulary, memory and problem solving.
Authors: Kamyar Zeinalipour, Yusuf Gökberk Keptiğ, Marco Maggini, Leonardo Rigutini, Marco Gori.
Read the paper | Code | Models and data
All the lab's language releases are listed in Language models and datasets.