СИСТЕМА АВТОМАТИЧНОГО КОРЕГУВАННЯ АНГЛІЙСЬКО- УКРАЇНСЬКОГО КОМП'ЮТЕРНОГО ПЕРЕКЛАДУ ДЛЯ ТЕХНІЧНИХ ТЕКСТІВ В ГАЛУЗІ АВТОМАТИЗАЦІЇ ТЕХНОЛОГІЧНИХ ПРОЦЕСІВ

Autor: Стопакевич, А. О., Тігарєв, А. М., Романюк, О. Р., Стопакевич, О. А.
Zdroj: Informatics & Mathematical Methods in Simulation / Informatika ta Matematičnì Metodi v Modelûvannì; 2023, Vol. 13 Issue 3/4, p322-334, 13p
Abstrakt: The purpose of the work is to develop a system of automatic correction of computer-translated texts with specific terminology, which is inherent in scientific and technical texts in the field of automation of technological processes. An analysis of the reasons why computer translators cannot achieve high-quality English-Ukrainian translation of technical texts in the specified field is given. It was concluded that within the limits of the approach used by modern computer translators, the quality of such translations cannot be improved. An analysis of the experience of translation correction by professional translators, available metrics for evaluating the process and results of correction of computer translations, and available software solutions for working with texts written in the Ukrainian language was carried out. It was concluded that for computer English-Ukrainian translation, the only practically significant approach to assessing its quality is measuring the amount of work that must be performed by a professional translator in order for the text to meet literary standards. The analysis of scientific texts that were translated by DeepL showed that the amount of such work can be significantly reduced, since the errors made by this translator are systematic. Thus, by analyzing the mistakes made by a computer translator, it is possible to form universal correction rules for all texts translated by a certain translator of the industry, which can be performed automatically by the software application. The effectiveness of the approach is demonstrated on the example of the development of rules resulting from the analysis of the results of the translation of two scientific articles. It is shown that replacing approximately 5% of words in computer translation significantly increases its quality. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index