|
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ИНХС РАН |
||
What can we gain from language models for morphological inflection?статья
- Автор: Sorokin Alexey
- Сборник: Proceedings of the CoNLL SIGMORPHON 2018 Shared Task: Universal Morphological Reinflection
- Год издания: 2018
- Место издания: Association for Computational Linguistics
- Первая страница: 99
- Последняя страница: 104
- Аннотация: This paper investigates the attempts to augment neural-based inflection models with characterbased language models. We found that in most cases this slightly improves performance, however, the effect is marginal. We also propose another language-model based approach that can be used as a strong baseline in low-resource setting.
- Добавил в систему: Сорокин Алексей Андреевич