Comparative Analysis: Machine vs Human Translation of Educational Scientific Texts. Challenges and Solutions

Introduction. The article presents the analysis of inaccuracies and errors in machine translation and deals with challenges encountered while translating specialist educational scientific texts. A number of feasible techniques to eliminate the inaccuracies and errors have been offered. Methodology and sources. Several machine-translated (DeepL programme) sections of a textbook on astronomy have been evaluated by a human expert. Results and discussion. The research allowed to identify and classify typical errors DeepLtranslated texts. One of the findings was the conclusion about the need and feasibility of eliminating inaccuracies and errors in machine translation using embedded software techniques. There are also inaccuracies and errors that still require human post-editing. Conclusion. Integrating the proposed software techniques for eliminating errors into a machine translation programme can provide a qualitatively new level of machine translation. This is possible with close cooperation between linguists, specific area experts and programmers.

Authors: Nadezhda A. Antonova, Irina V. Kuzmich

Direction: Linguistics

Keywords: machine translation, attributive noun chains, modulation, expansion, topic-comment relations, typographics standards, a standard target language image, localization

---

View full article