• Journal of Information Technology Services
• /
• v.21 no.2
• /
• pp.85-95
• /
• 2022

This study analyzed the translation error factors of machine translation services such as Naver Papago and Google Translate through Self-Attention path visualization. Self-Attention is a key method of the Transformer and BERT NLP models and recently widely used in machine translation. We propose a method to explain translation error factors of machine translation algorithms by comparison the Self-Attention paths between ST(source text) and ST'(transformed ST) of which meaning is not changed, but the translation output is more accurate. Through this method, it is possible to gain explainability to analyze a machine translation algorithm's inside process, which is invisible like a black box. In our experiment, it was possible to explore the factors that caused translation errors by analyzing the difference in key word's attention path. The study used the XLM-RoBERTa multilingual NLP model provided by exBERT for Self-Attention visualization, and it was applied to two examples of Korean-Chinese and Korean-English translations.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.489-497
• /
• 2007

Named entity translation plays an important role in many applications, such as information retrieval and machine translation. In this paper, we focus on translating person names, the most common type of name entity in Korean-Chinese cross language information retrieval (KCIR). Unlike other languages, Chinese uses characters (ideographs), which makes person name translation difficult because one syllable may map to several Chinese characters. We propose an effective hybrid person name translation method to improve the performance of KCIR. First, we use Wikipedia as a translation tool based on the inter-language links between the Korean edition and the Chinese or English editions. Second, we adopt the Naver people search engine to find the query name's Chinese or English translation. Third, we extract Korean-English transliteration pairs from Google snippets, and then search for the English-Chinese transliteration in the database of Taiwan's Central News Agency or in Google. The performance of KCIR using our method is over five times better than that of a dictionary-based system. The mean average precision is 0.3490 and the average recall is 0.7534. The method can deal with Chinese, Japanese, Korean, as well as non-CJK person name translation from Korean to Chinese. Hence, it substantially improves the performance of KCIR.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2017.07a
• /
• pp.425-426
• /
• 2017

본 논문에서는 Google에서 지원하는 오픈소스인 Tesseract-OCR을 이용한 문자/음성 인식 및 번역 앱에 대해 제안한다. 최근 한국어를 포함한 외국어 인식과 번역기능을 이용한 다양한 스마트폰 앱이 개발되어 여행에 필수품으로 자리잡고 있다. 스마트폰의 카메라기능을 이용하여 촬영한 영상을 인식률을 높이도록 처리하고, Crop기능을 넣어 부분 인식기능을 지원하며, Tesseract-OCR의 train data를 보완하여 인식률을 높이고, Google 음성인식 API를 이용한 음성인식 기능을 통해 인식된 유사한 문장들을 선택하도록 하고, 이를 번역하고 보여주도록 개발하였다. 번역 기능은 번역대상 언어와 번역할 언어를 선택할 수 있고 기본적으로 영어, 한국어, 일본어, 중국어로 번역이 가능하다. 이 기능을 이용하여 차량번호 인식, 사진에 포함된 글자를 통한 검색 등 다양한 응용분야에 맞게 앱을 개발할 수 있다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.1041-1044
• /
• 2019

자동통번역의 성능을 평가하는데 가장 잘 알려진 자동평가 기술은 BLEU이다. 그러나 BLEU로는 자동통번역 결과의 어느 부분이 강점이고 약점인지를 파악할 수 없다. 본 논문에서는 자동통번역 시스템의 언어 현상별 자동평가 방법을 소개하고자 한다. 언어 현상별 자동평가 방법은 BLEU가 제시하지 못하는 언어 현상별 자동평가가 가능하며 개발자로 하여금 해당 자동통번역 시스템의 언어 현상별 강점과 약점을 직관적으로 파악할 수 있도록 한다. 언어 현상별 정확도 측정은 Google 과 Naver Papago 를 대상으로 실시하였다. 정확률이 40%이하를 약점이라고 간주할 때, Google 영한 자동번역기의 약점은 스타일(32.50%)번역이었으며, Google 영한 자동통역기의 약점은 음성(30.00%)인식, 담화(30.00%)처리였다. Google 한영 자동번역기 약점은 구문(34.00%)분석, 모호성(27.50%)해소, 스타일(20.00%)번역이었으며, Google 한영 자동통역기 약점은 담화(30.00%)처리였다. Papago 영한 자동번역기는 대부분 정확률이 55% 이상이었으며 Papago 영한 자동통역기의 약점은 담화(30.00%)처리였다. 또한 Papago 한영 자동번역기의 약점은 구문(38.00%)분석, 모호성(32.50%)해소, 스타일(20.00%)번역이었으며, Google 한영 자동통역기 약점은 담화(20.00%)처리였다. 언어 현상별 자동평가의 궁극적인 목표는 자동통번역기의 다양한 약점을 찾아내어 약점과 관련된 targeted corpus 를 반자동 수집 및 구축하고 재학습을 하여 자동통번역기의 성능을 점증적으로 향상시키는 것이다.

• Communications for Statistical Applications and Methods
• /
• v.27 no.3
• /
• pp.313-326
• /
• 2020

Deep Learning is the most important key to the development of Artificial Intelligence (AI). There are several distinguishable architectures of neural networks such as MLP, CNN, and RNN. Among them, we try to understand one of the main architectures called Recurrent Neural Network (RNN) that differs from other networks in handling sequential data, including time series and texts. As one of the main tasks recently in Natural Language Processing (NLP), we consider Neural Machine Translation (NMT) using RNNs. We also summarize fundamental structures of the recurrent networks, and some topics of representing natural words to reasonable numeric vectors. We organize topics to understand estimation procedures from representing input source sequences to predict target translated sequences. In addition, we apply multiple translation models with Gated Recurrent Unites (GRUs) in Keras on English-Korean sentences that contain about 26,000 pairwise sequences in total from two different corpora, colloquialism and news. We verified some crucial factors that influence the quality of training. We found that loss decreases with more recurrent dimensions and using bidirectional RNN in the encoder when dealing with short sequences. We also computed BLEU scores which are the main measures of the translation performance, and compared them with the score from Google Translate using the same test sentences. We sum up some difficulties when training a proper translation model as well as dealing with Korean language. The use of Keras in Python for overall tasks from processing raw texts to evaluating the translation model also allows us to include some useful functions and vocabulary libraries as well.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.689-692
• /
• 2020

People have studied English using online English dictionaries when they looked for the meaning of English words or the example sentences. These days, as the AI technologies such as machine learning have been developing, documents can be translated in real time with Kakao, Papago, Google translators and so on. But, there has still been some problems with the accuracy of translation. The AI secretaries can be used for real-time interpreting, so this kind of systems are being used to translate such the web pages, papers into Korean. In this paper, we researched on the usage frequency of the combined English phrases from dictionaries by analyzing the number of the searched results on Google. With the result of this paper, we expect to help the people to use more English fluently.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.1
• /
• pp.121-128
• /
• 2024

With the recent rise in the use of AI-based online translation tools, interest in their methods and effects on education has grown. This study involved 30 prospective elementary school teachers who completed an English writing task using an AI-based online translation tool. The study focused on assessing the impact of these tools on English writing skills and their practical applications. It examined the usability, educational value, and the advantages and disadvantages of the AI translation tool. Through data collected via writing tests, surveys, and interviews, the study revealed that the use of translation tools positively affects English writing skills. From the learners' perspective, these tools were perceived to provide support and convenience for learning. However, there was also recognition of the need for educational strategies to effectively use these tools, alongside concerns about methods to enhance the completeness or accuracy of translations and the potential for over-reliance on the tools. The study concluded that for effective utilization of translation tools, the implementation of educational strategies and the role of the teacher are crucial.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.3
• /
• pp.229-236
• /
• 2011

In this paper we present light weighted English-to-Korean and Korean-to-English mobile translators on smart phones. For natural translation and higher translation quality, translation engines are hybridized with Translation Memory (TM) and Rule-based translation engine. In order to maximize the usability of the system, we combined an Optical Character Recognition (OCR) engine and Text-to-Speech (TTS) engine as a Front-End and Back-end of the mobile translators. With the BLEU and NIST evaluation metrics, the experimental results show our E-K and K-E mobile translation equality reach 72.4% and 77.7% of Google translators, respectively. This shows the quality of our mobile translators almost reaches the that of server-based machine translation to show its commercial usefulness.

• IEIE Transactions on Smart Processing and Computing
• /
• v.5 no.2
• /
• pp.77-84
• /
• 2016

In the field of natural language processing, Universal Networking Language (UNL) has been used by various researchers as an inter-lingual approach to automatic machine translation. The UNL system consists of two main components, namely, EnConverter for converting text from a source language to UNL, and DeConverter for converting from UNL to a target language. Currently, many projects are researching how to apply UNL to different languages. In this paper, we introduce the tools that are UNL's applications and discuss how to reuse them to encode a Vietnamese sentence into UNL expressions and decode UNL expressions into a Vietnamese sentence. The testing was done with about 1,000 Vietnamese sentences (a dictionary that includes 4573 entries and 3161 rules). In addition, we compare the proportion of sentences translated based on a direct method (Google Translator) and another one based on UNL.

• Journal of the Korea Convergence Society
• /
• v.11 no.6
• /
• pp.281-288
• /
• 2020

This study delves into acceptance on neural network machine translation (NMT) such as Google Translate and Papago that uses technical acceptance model. In conclusion, it turned out that perceived usefulness impacts translators' attitude towards NMT. In other words, if translators determine that NMT is related to their work and the quality of the deliverables is guaranteed, they were more positive towards it. Unlike the existing normative approach that translators feel threatened by NMT, empirical results tell us translators perceive NMT as a business tool and such perception was largely influenced by advices of their colleagues and friends and expectations for use.