人工智能背景下的韩译汉机器翻译质量对比研究

A Comparative Study on the Quality of Korean-Chinese Machine Translation under the Background of Artificial Intelligence

摘要: 当今时代，人工智能技术发展速度惊人，在各个领域的应用持续拓展且不断深入，机器翻译作为人工智能的一项关键应用，其性能和准确性也在快速提高，从早期的Papago翻译、有道翻译、谷歌翻译、DeepL以及各类网页翻译，到ChatGPT、Cloud、Kimi以及新出现的DeepSeek等人工智能，可与翻译相结合，对翻译起到关键作用。虽然机器翻译取得了一定进步，但仍存在一些局限性与不足，为更全面地了解这些翻译工具的性能和特点，本文选取DeepSeek、Kimi、DeepL这三个机器翻译对第三十六届韩素音国际翻译大赛朝译汉进行对比分析，利用MQM框架作为分析工具，发现这三款工具在词汇选择、句子结构、主语补充方面各有优缺点，而在实际应用中，机器翻译需与人工翻译相结合，为翻译行业发展提供新的思路和方法。

Abstract: In the current era, the development of artificial intelligence technology is astonishingly rapid, and its applications in various fields are continuously expanding and deepening. As a key application of artificial intelligence, machine translation has seen rapid improvements in performance and accuracy. From early tools like Papago, Youdao, Google Translate, DeepL, and various web-based translators, to newer AI models such as ChatGPT, Cloud, Kimi, and the emerging DeepSeek, these can be integrated with translation to play a crucial role. Despite the progress made in machine translation, there are still some limitations and shortcomings. To gain a more comprehensive understanding of the performance and characteristics of these translation tools, this paper conducts a comparative analysis of DeepSeek, Kimi, and DeepL on the Korean-to-Chinese translation task of the 36th International Translation Competition of Han Suyin. The study uses the MQM framework as the analysis tool. The study finds that each of these tools has its own strengths and weaknesses in terms of vocabulary selection, sentence structure, and subject supplementation. In practical applications, machine translation should be combined with human translation to provide new ideas and methods for the development of the translation industry.