|
ИСТИНА |
Войти в систему Регистрация |
ИПМех РАН |
||
Construction of multilingual parallel corpora of literary texts using automatic alignment and crowd sourcing facilitiesтезисы доклада
- Авторы: Kedrova G., Potemkin S.
- Сборник: Electronic lexicography in the 21st century (eLex 2021): Post-editing lexicography
- Тезисы
- Год издания: 2021
- Место издания: Brno: Lexical Computing CZ s.r.o Brno
- Первая страница: 24
- Последняя страница: 25
- Аннотация: Project’s goal: construction of “Multilingual parallel corpus of translations of A.P. Chekhov’s complete works”.Methodology: semi-automatic processing (automatic alignment, manual post-editing work, crowdsourcing).Current stage: correction of the results of automatic alignment of paragraphs and sentences in translation of A. Chekhov’s short stories into nine European languages. 150 students of the Linguistic Department of Lomonosov Moscow State University participated in in the work.Results: (I) each sentence/phrase in the translation texts was unambiguously aligned with the original one; (II) detailed analysis of the mismatches in the alignment and analysis of translation/linguistic causes producing alignment errors discussed through multiple multilingual examples. Types of the most frequent causes of erroneous alignment are presented.
- Добавил в систему: Кедрова Галина Евгеньевна