• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2019년도 제31회 한글 및 한국어 정보처리 학술대회
• /
• pp.258-261
• /
• 2019

딥러닝을 이용한 Sequence to Sequence 모델의 등장과 Multi head Attention을 이용한 Transformer의 등장으로 기계번역에 많은 발전이 있었다. Transformer와 같은 성능이 좋은 모델들은 대량의 병렬 코퍼스를 가지고 학습을 진행하였는데 대량의 병렬 코퍼스를 구축하는 것은 시간과 비용이 많이 드는 작업이다. 이러한 단점을 극복하기 위하여 합성 코퍼스를 만드는 기법들이 연구되고 있으며 대표적으로 Back Translation 기법이 존재한다. Back Translation을 이용할 시 단일 언어 데이터를 가상 병렬 데이터로 변환하여 학습데이터의 양을 증가 시킨다. 즉 말뭉치 확장기법의 일종이다. 본 논문은 Back Translation 뿐만 아니라 Copied Translation 방식을 통한 다양한 실험을 통하여 데이터 증강기법이 기계번역 성능에 미치는 영향에 대해서 살펴본다. 실험결과 Back Translation과 Copied Translation과 같은 데이터 증강기법이 기계번역 성능향상에 도움을 줌을 확인 할 수 있었으며 Batch를 구성할 때 상대적 가중치를 두는 것이 성능향상에 도움이 됨을 알 수 있었다.

• 영어영문학
• /
• 제57권6호
• /
• pp.1173-1198
• /
• 2011

In this paper, I discuss Chaucer's Clerk's Tale by viewing the relationship between Walter and Griselda as that of a medieval translator and his translation. My major concern is how a medieval translation can serve power, more specifically the consolidation of power under particular historical circumstances. The motive and the process of Walter's creative translation of Griselda are closely examined to show that his translation, which includes a creation of a new Griselda as a pinnacle of wifely virtue of patience, is performed as a form of political propaganda, ultimately aimed at strengthening his governing power over his people and land. My discussion of the Clerk's Tale ends with the comparison of the two translators, Walter and the Clerk, the latter of whom is an example of an unsuccessful translator for his lack of creation in the translation.

• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2003년도 춘계 학술발표회 논문집
• /
• pp.219-224
• /
• 2003

Translation errors in the statistical funtions of the Korean Excel 2002 software are analyzed into three categories: Pure translation mistakes; Vague translation; Inconsistent translation. These errors have not been corrected at least for the last ten years and therefore, it is suspicious that the Korean agents for foreign software companies are more interested in marketing with monololy power than in improving quality of their products

• 한국언어정보학회지:언어와정보
• /
• 제8권1호
• /
• pp.101-116
• /
• 2004

This paper proposes a statistical model for the translation of prepositions in English-Korean machine translation. In the proposed model, statistical information acquired from unlabeled Korean corpora is used to choose the best translation from several possible translations. Such information includes functional word-verb co-occurrence information, functional word-verb distance information, and noun-postposition co-occurrence information. The model was evaluated with 443 sentences, each of which has a prepositional phrase, and we attained 71.3% accuracy.

• 한국수학교육학회지시리즈B:순수및응용수학
• /
• 제28권1호
• /
• pp.43-53
• /
• 2021

We study translation surfaces in the Euclidean 3-space ��3 and the Gauss map N with respect to the so-called Cheng-Yau operator ☐. As a result, we prove that the only translation surfaces with Gauss map N satisfying ☐N = AN for some 3 × 3 matrix A are the flat ones. We also show that the only translation surfaces with Gauss map N satisfying ☐N = AN for some nonzero 3 × 3 matrix A are the cylindrical surfaces.

• 한국언어정보학회지:언어와정보
• /
• 제2권1호
• /
• pp.79-105
• /
• 1998

This paper aims to describe a German-Korean machine translation system for nominal phrases. Besides, we have two subgoals. First, we are going to revea linguistic differences between two languages and propose a language-informational method fo overcome the differences. The method is based on an integrated model of translation knowledge, efficient information structure, and concordance selection. Then, we will show the statistical results about translation experiment and its evaluation as an evidence for the adequacy of our linguistic method and translation system itself.

• 호남수학학술지
• /
• 제45권2호
• /
• pp.359-379
• /
• 2023

In the present paper, we study translation hypersurfaces in E⁴. In this context, firstly we obtain first, second and third Laplace-Beltrami (LB^I, LB^II and LB^III) operators of the translation hypersurfaces in E⁴. By solving second and third order nonlinear ordinary differential equations, we prove theorems that contain LB^I-minimal, LB^II-minimal and LB^III-minimal translation hypersurfaces in E⁴.

• Wind and Structures
• /
• 제26권3호
• /
• pp.179-190
• /
• 2018

Translation of tornadoes is an important feature in replicating the near-ground tornado flow field which has been simulated in previous studies based on Ward-type tornado simulators using relative motion of the ground plane. In this laboratory investigation, effects of translation on the near-ground tornado flow field were studied using the ISU Tornado Simulator that can physically translate over a ground plane. Two translation speeds, 0.15 m/s and 0.50 m/s, that scale up to those corresponding to slowly-moving tornadoes in the field were selected for this study. Compared with the flow field of a stationary tornado, the simulated tornado with translation had an influence on the spatial distribution and magnitude of the horizontal velocities, early reversal of the radial inflow, and expansion of the core radius. Maximum horizontal velocities were observed to occur behind the center of the translating tornado and on the right side of its mean path. An increase in translation speed, resulted in reduction of maximum horizontal velocities at all heights. Comparison of the results with previous studies that used relative motion of the ground plane for simulating translating tornadoes, showed that translation has similar effects on the flow field at smaller radial distances (~2 core radius), but different effects at larger radial distances (~4 core radius). Further, it showed that the effect of translation on velocity profiles is noticeable at and above an elevation of ~0.6 core radius, unlike those in studies based on the relative motion of the ground plane.

• Journal of Acupuncture Research
• /
• 제32권4호
• /
• pp.1-9
• /
• 2015

Objectives : The aim of this study is to develop the official Korean translation of the CARE guidelines. Methods : CARE guidelines were developed in 2013 though a consensus of experts led by the CARE group. It is the first reporting guideline for case reports that includes a 13-item checklist. We contacted the main author of the original manuscript and obtained permission to develop the Korean translation. The translation process followed a systematic protocol designated by the CARE group which included an independent back translation and confirmation from the original author. Results : The Korean translation of the CARE guidelines was completed with approval from the original author. The Korean checklist is posted on the official website (www.care-statement.org) and the final translation of the manuscript is introduced in this article as an appendix. Conclusions : We hope that this will help introduce the CARE guidelines to Korean authors and other stakeholders, induce a more active use of reporting guidelines, and eventually contribute to the overall improvement of the reporting quality of case reports published in Korean literature.

• 대한수의학회지
• /
• 제39권4호
• /
• pp.772-777
• /
• 1999

The uncaped genomic RNA of bovine viral diarrhea virus (BVDV) initiates translation by recruitment of eukaryotic translation initiation factors at the internal ribosome entry site (IRES). N-terminal protease ($N^{pro}$) is the first translation product of the open reading frame (ORF). By using the vaccinia virus SP6 RNA polymerase transient expression system, we showed previously that deletion of $N^{pro}$ region reduced translation by 21%. To better understand the biological significance of $N^{pro}$ for translation, we carried out a mutational analysis of the $N^{pro}$ region of BVDV cloned in the intercistronic region of a bicistronic reporter plasmid. We constructed a bicistronic expression vector in which the entire 5 UTR and the mutated $N^{pro}$ region (${\Delta}386-901$, ${\Delta}415-901$ and ${\Delta}657-901$) was cloned between two reporter genes, chloramphenicol acetyltransferase (CAT) and luciferase (LUC). In vivo translation analyses showed that $N^{pro}$ region was dispensible for efficient translation. The results indicate that the $N^{pro}$ region is not essential for BVDV RNA translation and the 3' boundary of BVDV IRES is expanded into $N^{pro}$ region, suggesting that $N^{pro}$ may not play a major role in BVDV replication.