신경망 기계번역(NMT) 기반 번역 품질 비교: Garden-path 예문을 중심으로
Comparison of Translation Qualities Based on Neural Machine Translators: Focusing on Garden-Path Sentences
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
The purpose of this paper is to compare the rates of accurate translations among neural machine translators such as Papago, Google, DeepL, and Copilot and to make suggestions for improving their accuracy. The average accuracy of four neural machine translators for Garden-path sentences was 16.25%. However, the average accuracy of the four neural machine translators for sentences that were slightly modified from the original Garden-path sentences was 61.25%, 3.77 times higher than 16.25%. The average accuracy of 75.36% for papers dealing with the accuracy of Google and DeepL was 4.64 times higher than the average accuracy of 16.25% for Garden-path sentences. It was also 1.26 times higher than the average accuracy of 60% for modified Garden-path sentences. To improve the accuracy of neural machine translators for Garden-path sentences or similar sentences, it is necessary to introduce a verification process by human expert translators, hybrid processing that combines existing machine translation methods, or improvements to the neural network system's algorithms to enable self-review and feedback.
Abstract
The purpose of this paper is to compare the rates of accurate translations among neural machine translators such as Papago, Google, DeepL, and Copilot and to make suggestions for improving their accuracy. The average accuracy of four neural machine translators for Garden-path sentences was 16.25%. However, the average accuracy of the four neural machine translators for sentences that were slightly modified from the original Garden-path sentences was 61.25%, 3.77 times higher than 16.25%. The average accuracy of 75.36% for papers dealing with the accuracy of Google and DeepL was 4.64 times higher than the average accuracy of 16.25% for Garden-path sentences. It was also 1.26 times higher than the average accuracy of 60% for modified Garden-path sentences. To improve the accuracy of neural machine translators for Garden-path sentences or similar sentences, it is necessary to introduce a verification process by human expert translators, hybrid processing that combines existing machine translation methods, or improvements to the neural network system's algorithms to enable self-review and feedback.
참고문헌 (26)
- 최동익, "모바일 기계번역기와 기존의 기계번역기의 영한 번역 비교", 동남보건대학교 논문집, 35(1), pp.55~69 (2017)
- Aiken, M, "An Updated Evaluation of Google Translate Accuracy", Studies in Linguistics and Literature, 3(3), pp.253~260 (2019)[DOI]
- Aiken, M., "An Analysis of Google Translate Accuracy", Translation Journal, 16(2) (2011)
- Baharuddin, S., "The Analysis of Google Translate Accuracy in Translating Procedural and Narrative Text", Journal of English Education Forum, 2(1), pp.7~11 (2022)
- Bever, T. G., "Cognition and Language Development", pp.279~352 (1970), Wiley & Sons, Inc
- Du, J-l., "A Computational Linguistic Approach to Natural Language Processing with Applications to Garden Path Sentences Analysis", International Journal of Advanced Computer Science and Applications, 3(9), pp.61~75 (2012)
- Eszenyi R., "Artificial Intelligence, Machine Translation & Cyborg Translators : A Clash of Utopian and Dystopian Visions", Orbislinguarum, 21(2), pp.102~113 (2023)[DOI]
- Frazier L., "On Comprehending Sentences: Syntactic Parsing Strategies", Doctor, pp.1979
- Frazier, L., "Making and Correcting Errors During Sentence Comprehension : Eye Movements in the Analysis of Structurally Ambiguous sentences", Cognitive Psychology, 14(2), pp.178~210 (1982)[DOI]
- Gao, R., "Machine Translation of Chinese Classical Poetry: a Comparison among ChatGPT, Google Translate, and DeepL Translator" (2024)[DOI]
- Multidisciplinary Analysis of the Phenomenon of Phraseological Variation in Translation and Interpreting, "Análisis multidisciplinar del fenómeno de la variación fraseológica en traducción e interpretación", pp.2020, Hidalgo-Ternero, C. M.
- Hockett, C, "Grammar for the Hearer", Proceedings of Symposia in Applied Mathematics, Structure of Language and Its Mathematical Aspects, 12 (1961), American Mathmatical Society
- "Internet Garden-path Sentences"[URL]
- Mercan, H., "The Evolution of Machine Translation: A Review Study", International Journal of Language and Translation Studies, 4(1), pp.104~116 (2024)
- Mohamed, Y, "The Impact of Artificial Intelligence on Language", IEEE Access (2024)
- Moneus, A. M, "Artificial Intelligence and Human Translation: A Contrastive Study Based on Legal Texts", Heliyon (2024)
- Patil, S., "Use of Google Translate in Medical Communication: Evaluation of Accuracy", British Medical Journal (2014)[DOI]
- Tabor, W., "Effects of Merely Local Syntactic Coherence on Sentence Processing", Journal of Memory and Language, 50(4), pp.355~370 (2004)[DOI]
- Wan, W., "Studies on Garden Path Phenomenon in English", US-China Foreign Language, 18(6), pp.178~182 (2020)
- Wang, L., "The Impacts and Challenges of Artificial Intelligence Translation Tool on Translation Professionals", SHS Web of Conferences (2023)
- Wang, L. H, "Explorations on the Triggering Factors of Garden Path Phenomenon", Journal of Southwest Agricultural University, 11, pp.93~95 (2013)
- Yoo, H., "Aging Effects and Working Memory in Garden-Path Recovery", Clinical Archives of Communication Disorders, 2(2), pp.91~102 (2017)[DOI]
- Yulianto, A., "Google Translate vs. DeepL: A Quantitative Evaluation of Close-language Pair Translation (French to English)", The Asian Journal of English Language & Pedagogy, 9(2), pp.109~127 (2021)[DOI]
- "Kwukminilbo Internet News" (2017)[URL]
- "Yonhap Internet News" (2017)[URL]
- "Chosun Internet News" (2017)[URL]