自然言語処理学研究室のJustin VASSELLIさん(博士後期課程2年)らが、AmericasNLPワークショップのシェアードタスク2で優勝しました。(2025/5/4)
Justin VASSELLI et al. of Natural Language Processing lab won the first prize in the Shared Task 2 at the Workshop on NLP for Indigenous Languages of the Americas (AmericasNLP).
第5回AmericasNLPワークショップは、アメリカ大陸の先住民言語における自然言語処理、計算言語学、コーパス言語学、音声技術の研究を推進することを目的としています。このワークショップでは、研究者、マイノリティコミュニティの専門家、そしてネイティブスピーカーを機械学習およびNLP分野のコミュニティと結びつけ、リソースの少ない言語における多様な機械学習アプローチの推進、データセット作成の促進、そして倫理的配慮への取り組みを促進します。 | ![]() |
AmericasNLP 2025のシェアードタスク(共通課題)2「教材の作成」では、Justin VASSELLIらのチームがブリブリ語とナワトル語の先住民言語の練習問題を自動生成する優れたシステムを開発し優勝しました。 | |
The Fifth AmericasNLP Workshop aims to advance research in NLP, computational linguistics, corpus linguistics, and speech technologies for Indigenous languages across the Americas. It connects researchers, professionals from underrepresented communities, and native speakers with the machine learning and NLP communities, promoting diverse machine learning approaches for low-resource languages, fostering dataset creation, and addressing ethical considerations. The AmericasNLP 2025 Shared Task 2, "Creation of Educational Materials," awarded the winning team for their system that excels at automatically generating exercises for Indigenous languages in Bribri and Nahuatl. |
- 受賞者/著者 Awardees /Authors:
Justin Vasselli(D2), Haruki Sakajo(M2), Arturo Martinez Peguro(D1), Frederikus Hudi(D1), and Taro Watanabe
写真左からFrederikus Hudiさん、Justin Vasselliさん、坂上 温紀さん
- 受賞研究テーマ Research theme:
"Leveraging Dictionaries and Grammar Rules for the Creation of Educational Materials for Indigenous Languages"
This paper describes the NAIST submission to the AmericasNLP 2025 shared task on the creation of educational materials for Indigenous languages. We implement three systems to tackle the unique challenges of each language. The first system, used for Maya and Guarani, employs a straightforward GPT-4o few-shot prompting technique, enhanced by synthetically generated examples to ensure coverage of all grammatical variations encountered. The second system, used for Bribri, integrates dictionary-based alignment and linguistic rules to systematically manage linguisticand lexical transformations. Finally, we developed a specialized rule-based system for Nahuatl that systematically reduces sentences to their base form, simplifying the generation of correct morphology variants.
- 受賞者のコメント Awardee's voice:
We're honored to receive this award recognizing our work on generating educational materials for Bribri and Nahuatl. Supporting Indigenous languages through technology is deeply meaningful to us, and I hope our work can contribute in some small way to keeping these languages alive.
- 外部リンク Links to:
- AmericasNLP HP:https://turing.iimas.unam.mx/americasnlp/