RUAccent: Advanced System for Stress Placement in Russian with Homograph Resolution

Denis Andreevich Petrov


Abstract
This paper presents a novel approach to the problem of stress placement in Russian text, with a particular focus on resolving homographs. We introduce a comprehensive system that combines morphological analysis, context-aware neural models, and a specialized “Ё-fikator” to accurately place stress in Russian words, including those with ambiguous pronunciations. Our system outperforms existing solutions, achieving a 0.96 accuracy on homographs and 0.97 accuracy on non-homograph words.
Anthology ID:
2025.coling-main.444
Volume:
Proceedings of the 31st International Conference on Computational Linguistics
Month:
January
Year:
2025
Address:
Abu Dhabi, UAE
Editors:
Owen Rambow, Leo Wanner, Marianna Apidianaki, Hend Al-Khalifa, Barbara Di Eugenio, Steven Schockaert
Venue:
COLING
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6642–6648
Language:
URL:
https://aclanthology.org/2025.coling-main.444/
DOI:
Bibkey:
Cite (ACL):
Denis Andreevich Petrov. 2025. RUAccent: Advanced System for Stress Placement in Russian with Homograph Resolution. In Proceedings of the 31st International Conference on Computational Linguistics, pages 6642–6648, Abu Dhabi, UAE. Association for Computational Linguistics.
Cite (Informal):
RUAccent: Advanced System for Stress Placement in Russian with Homograph Resolution (Petrov, COLING 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.coling-main.444.pdf