Özet:

In recent years, significant developments have been in machine translation systems, but still these systems can produce erroneous or unnatural translations. In this study, in order to overcome the aforementioned problem, a innovative workflow is proposed that the machine translation systems and artificial intelligence-based chat robots can take advantage of the advantages of quality translations to the user. The proposed workflow consists of four basic steps: (1) the first translation performed using the machine translation system, (2) the art editing and final reading performed with the chat robot, (3) the man-oriented automatic language knowledge and writing control, and (4) the final evaluation performed using numerical standards. In the first stage, the source text is translated into the male language with the help of the machine translation system. In the second step, the machine translation output is reviewed by a chat robot. Translation can be developed using a chat robot with natural language understanding skills. This phase helps to solve the uncertainty of meaning that cannot be immediately understood by machine translation systems, cultural elements and problems related to terminology. In the third step, the language knowledge and writing control of the translation made by the art editing and final reading chat robot is carried out automatically. The final objective of this phase is to remove the translation from linguistic errors as far as possible and to improve its readability and overall quality by ensuring compliance with the male language rules. In the final step of the workflow, using the numerical criteria obtained from a writing control tool, the acceptability of the male text and the proximity to the human translation are assessed. These standards, which provide quantitative feedback to the translator about the quality of the translation, can be used for the purpose of determining the fields of improvement and guiding for future translation work. As the recommended translation workflow, the success of machine translation systems and the combination of the contextual awareness of chat robots and the natural language understanding skills can allow for higher quality translations to appear, it can reduce the effort of translation in the art editing and final reading processes. With this workflow, a comprehensive approach to producing high-quality and natural male texts is offered.

Anahtar Kelimeler: