• The Journal of Korean Physical Therapy
• /
• 제26권5호
• /
• pp.344-350
• /
• 2014

Purpose: The purpose of this study was to translate and culturally adapt the International Classification of Functioning, Disability and Health (ICF) into the Korean language. Methods: The process of translation and adaptation of the ICF used here followed the translation guidelines of WHO. Implementation of this procedure comprised of four steps; forward translation, expert panel back-translation, pre-testing and cognitive interviewing, and final adaptation. The translators included health professionals with knowledge of ICF and non-health professionals blinded to the ICF. Clinical academics with significant experience in the use of disability survey, medical doctors, special educators, related policy makers, clinicians, architecture professionals, and international experts in ICF were invited to integrate all versions of the ICF for testing; 151 clinicians volunteered from 19 medical institutes across the country. Four different core-sets and a questionnaire were used for testing its practical usability and adaptation. Results: All translations were reviewed and a consensus was reached on any discrepancy from the earlier versions. Over 90% of the newly translated version of K-ICF was found to be different from the 2004 K-ICF version in the ICF language. Understanding of K-ICF language was responded difficult and very difficult by 50% of participants, whereas its practical use was responded 'useful' by more than 50% of subjects. Conclusion: It can be suggested that the new version of K-ICF should be widely used for final adaptation in the field of areas. Future studies will be required for implementation of K-ICF.

• 정보처리학회논문지B
• /
• 제17B권4호
• /
• pp.327-332
• /
• 2010

보통 두 세 개의 어휘로 구성된 질문 분류에서 어휘의 다양한 표현으로 인한 어휘 불일치문제는 성능 저하의 주요 원인이다. 따라서 질문 분류에서 어휘 사이의 연관성을 반영하는 것이 필수적이다. 본 논문에서는 같은 범주의 질문-질문 쌍들에 대해 계산한 어휘 번역확률을 번역기반 언어모델에 반영하여 질문을 분류하는 방법을 제안한다. 실험에서 야후!앤써 질문대답 아카이브를 이용해서 전체 질문-대답 쌍들에 대해서 번역확률을 계산하는 것보다 같은 범주에 속하는 질문-질문 쌍들에 대해서 번역확률을 계산하는 것이 질문 분류에서 더 좋은 번역확률인 것을 증명한다.

• Journal of Advanced Marine Engineering and Technology
• /
• 제36권5호
• /
• pp.683-693
• /
• 2012

통계기반 자동 번역 시스템은 구현과 유지보수의 용이함으로 최근 많은 관심을 받고 있다. 본 연구의 목적은 MOSES[1] 시스템을 이용하여 통계기반의 한/일 양방향 기계번역시스템을 구축하는 것이다. 한/일 문장단위 병렬 코퍼스를 구축하여 번역모델 학습에 이용하였고, 한/일 각각 대량의 원시 코퍼스를 이용하여 언어모델 학습에 이용하였다. 시스템 구축 결과 기존의 규칙기반 번역 시스템의 성능에 근접하는 결과를 얻었으며, 발생하는 오류의 대부분은 각 처리 단계에서 발생하는 노이즈에 기인하였다.

• Communications for Statistical Applications and Methods
• /
• 제27권3호
• /
• 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.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2015년도 제27회 한글 및 한국어 정보처리 학술대회
• /
• pp.115-118
• /
• 2015

Neural Machine Translation (NMT) 모델은 단일 신경망 구조만을 사용하는 End-to-end 방식의 기계번역 모델로, 기존의 Statistical Machine Translation (SMT) 모델에 비해서 높은 성능을 보이고, Feature Engineering이 필요 없으며, 번역 모델 및 언어 모델의 역할을 단일 신경망에서 수행하여 디코더의 구조가 간단하다는 장점이 있다. 그러나 NMT 모델은 출력 언어 사전(Target Vocabulary)의 크기에 비례해서 학습 및 디코딩의 속도가 느려지기 때문에 출력 언어 사전의 크기에 제한을 갖는다는 단점이 있다. 본 논문에서는 NMT 모델의 출력 언어 사전의 크기 제한 문제를 해결하기 위해서, 입력 언어는 단어 단위로 읽고(Encoding) 출력 언어를 문자(Character) 단위로 생성(Decoding)하는 방법을 제안한다. 출력 언어를 문자 단위로 생성하게 되면 NMT 모델의 출력 언어 사전에 모든 문자를 포함할 수 있게 되어 출력 언어의 Out-of-vocabulary(OOV) 문제가 사라지고 출력 언어의 사전 크기가 줄어들어 학습 및 디코딩 속도가 빨라지게 된다. 실험 결과, 본 논문에서 제안한 방법이 영어-일본어 및 한국어-일본어 기계번역에서 기존의 단어 단위의 NMT 모델보다 우수한 성능을 보였다.

• 한국언어정보학회:학술대회논문집
• /
• 한국언어정보학회 1998년도 Language, Information and Computation = Selected Papers from the 12th Pacific Asia Conference on Language, Information and Computation, Singapore
• /
• pp.248-254
• /
• 1998

• 한국언어정보학회:학술대회논문집
• /
• 한국언어정보학회 1996년도 Language, Information and Computation = Selected Papers from the 11th Pacific Asia Conference on Language, Information and Computation, Seoul
• /
• pp.43-52
• /
• 1996

• 한국중재학회지:중재연구
• /
• 제15권1호
• /
• pp.207-228
• /
• 2005

When closing an international contract, both contract parties endeavor to convey their intentions from the stage of negotiation to the moment of signing the contract. Of the many problems presently related to contract language, the first one to consider is which contract party will run the risk of the language deficiencies occurring as a result of the misunderstanding and misinterpretation between different languages. The second problem to consider is whether the interpretation and translation of the contract language is needed and, if so, which party is going to bear the expenses and assume responsibility of the misinterpretation in the translation of, the contract language. The third problem is related to the obligation of explaining to both contract parties the contents and details of the international contract written in different languages. The fourth issue is which language of both contract parties becomes the standard contract language in the procedure of arbitration. The fifth, but not the last problem, is how to solve the language defects in interpreting and translating the contract languages. These five problems can be easily solved by the approval of the contract parties in scrutinizing and selecting the contract languages. However, this research mainly focuses on which effects of the contract language and as how to define and select the contract language.

• International Journal of Advanced Culture Technology
• /
• 제6권4호
• /
• pp.165-171
• /
• 2018

As the artificial intelligence (AI) plays a crucial role in machine translation (MT) which has loomed large as a new translation paradigm, concerns have also arisen if MT can produce a quality product as human translation (HT) can. In fact, several MT experimental studies report cases in which the MT product called post-editing (PE) as equally as HT or often superior ([1],[2],[6]). As motivated from those studies on translation quality between HT and PE, this study set up an experimental situation in which Korean literature was translated into English, comparatively, by 3 translators and 3 post-editors. Afterwards, a group of 3 other Koreans checked for accuracy of HT and PE; a group of 3 English native speakers scored for fluency of HT and PE. The findings are (1) HT took the translation time, at least, twice longer than PE. (2) Both HT and PE produced similar error types, and Mistranslation and Omission were the major errors for accuracy and Grammar for fluency. (3) HT turned to be inferior to PE for both accuracy and fluency.

• 한국융합학회논문지
• /
• 제12권7호
• /
• pp.37-44
• /
• 2021

기계번역 품질 예측(Quality Estimation, QE)은 정답 문장(Reference sentence) 없이도 기계번역 결과의 질을 평가할 수 있으며, 활용도가 높다는 점에서 그 필요성이 대두되고 있다. Conference on machine translation(WMT)에서 매년 이와 관련한 shared task가 열리고 있고 최근에는 대용량 데이터 기반 Pretrained language model(PLM)을 적용한 연구들이 주로 진행되고 있다. 본 논문에서는 기계번역 품질 예측 task에 대한 설명 및 연구 동향에 대한 전반적인 survey를 진행했고, 최근 자주 활용되는 PLM의 특징들에 대해 정리하였다. 더불어 아직 활용된 바가 없는 multilingual BART 모델을 이용하여 기존 연구들인 XLM, multilingual BERT, XLM-RoBERTa와 의 비교 실험 및 분석을 진행하였다. 실험 결과 어떤 사전 학습된 다중언어 모델이 QE에 적용했을 때 가장 효과적인지 확인하였을 뿐 아니라 multilingual BART 모델의 QE 태스크 적용 가능성을 확인했다.