• Annual Conference on Human and Language Technology
• /
• 2008.10a
• /
• pp.113-117
• /
• 2008

본 논문에서는 패턴기반 자동번역시스템의 효율적인 번역 성능 향상을 위해 병렬 코퍼스(parallel corpus)를 이용한 오류 자동 탐지 방법을 제안하고자 한다. 번역시스템에 존재하는 대부분 오류는 크게 지식 오류와 엔진 오류로 나눌 수 있는데 통상 이런 오류는 이중 언어가 가능한 훈련된 언어학자가 대량의 자동번역 된 결과 문장을 읽음으로써 오류를 탐지하고 분석하여 번역 지식을 수정/확장하거나 또는 엔진을 개선하게 된다. 하지만, 이런 작업은 많은 시간과 노력을 필요로 하게 된다. 따라서 본 논문에서는 병렬 코퍼스 중의 목적 언어(Target Language) 문장 즉, 정답 문장과 자동번역 된 결과 문장을 다양한 방법으로 비교하면서 번역시스템에 존재하고 있는 지식 및 엔진 오류를 자동으로 탐지하는 방법을 제안한다. 제안한 방법은 한-중 자동번역시스템에 적용하여 그 정확률과 재현률을 측정하였으며, 자동적으로 오류를 탐지하여 추출 할 수 있음을 증명하였다.

• English Language & Literature Teaching
• /
• v.13 no.2
• /
• pp.177-200
• /
• 2007

The language socialization approach states that novices are socialized into cultural norms through participating in routine, repeated interactional acts and sequences (e.g., Ochs & Schieffelin, 1984; Ochs, 1988; Schieffelin & Ochs, 1986a; 1986b; Watson-Gegeo & Gegeo, 1986). One of the cultural norms or dominant epistemological orientations in American culture is the tendency to avoid the overt display of power asymmetry in novice-expert relationship (Ochs & Schieffelin, 1984). This study examines how this cultural preference is reflected and encoded in ESL teachers' use of routine discourse patterns in corrective sequences. Eight hours of ESL classes taught by three Caucasian teachers born and educated in the U.S. were analyzed for the study. The analysis showed that the cultural tendency in question is keyed and indexed in the teacher's routine corrective discourse patterns in the form of various questioning, elicitation, and mitigation practices. Findings support that teachers' routine classroom discourse practices represent their cultural ideologies and transfer these cultural predispositions to second language learners and that they possibly socialize the learners into the target language-oriented beliefs.

• Journal of Korean language education
• /
• v.25 no.1
• /
• pp.29-52
• /
• 2014

The purpose of this paper is to suggest a discourse analysis activity that can be used in the stage after performing tasks in task-based Korean speaking class and show its pedagogical advantages. A discourse analysis activity is an metadiscourse activity in which learners speak what they have spoken. By analyzing discourse and performing tasks again, learners can enhance their fluency and accuracy, make their knowledges in target language more stable and extend them, and develop problem solving skills. Consequently, this facilitates learners' acquisition of Korean language. This paper reviewed theoretical background of proposing discourse analysis activity, suggested the pedagogical advantages of the analysis, and examined discourse analysis activity in Korean speaking class. And it included the discourse sample of learners in actual class.

• Asia Pacific Journal of Corpus Research
• /
• v.1 no.2
• /
• pp.15-26
• /
• 2020

This corpus-based study investigated third language acquisition by Taiwanese college students in learning Spanish clitic pronouns at beginning and intermediate levels. It examined the acquisition sequences of Spanish clitic pronouns of the Chinese-speaking learners whose second language was English and third language was Spanish. The results indicated that indirect object pronouns (OP) preceded direct OP (case), first person preceded third person OP (person), masculine preceded feminine OP (gender), and animate preceded inanimate OP (animacy). The findings presented similar patterns as those of previous studies on English-speaking learners of Spanish. In further comparisons of the target forms in Chinese, English, and Spanish, the results suggested that L1 Chinese had strong influence on L3 Spanish, which accounts for the challenges that Taiwanese learners of Spanish face as they learn the Spanish clitic pronouns in the beginning stage.

• ETRI Journal
• /
• v.46 no.1
• /
• pp.59-70
• /
• 2024

We introduce a high-performance named entity recognition (NER) model for written and spoken language. To overcome challenges related to labeled data scarcity and domain shifts, we use transfer learning to leverage our previously developed KorBERT as the base model. We also adopt a meta-pseudo-label method using a teacher/student framework with labeled and unlabeled data. Our model presents two modifications. First, the student model is updated with an average loss from both human- and pseudo-labeled data. Second, the influence of noisy pseudo-labeled data is mitigated by considering feedback scores and updating the teacher model only when below a threshold (0.0005). We achieve the target NER performance in the spoken language domain and improve that in the written language domain by proposing a straightforward rollback method that reverts to the best model based on scarce human-labeled data. Further improvement is achieved by adjusting the label vector weights in the named entity dictionary.

• Phonetics and Speech Sciences
• /
• v.10 no.4
• /
• pp.45-52
• /
• 2018

This study investigates the effect of Japanese (L1) on the production of the Korean rhythmic structure. Korean and Japanese have typologically different rhythmic structure as a syllable-timed language and mora-timed language, respectively. This rhythmic difference comes from the different phonological properties of the two languages. Due to this difference, Japanese speakers that are learning Korean may produce a different rhythm than native Korean speakers' rhythm. To investigate the influence of the native language's rhythm on the target language, we conducted an acoustic analysis using acoustic metrics such as %V, VarcoV, and VarcoS. Four Korean native speakers and ten advanced Japanese Korean learners participated in a production test. The analyzed material consisted of six Korean sentences that contained various syllable structures. The results showed that KS and JS's rhythms are different in %V as well as in VarcoV. In the case of VarcoS, significant rhythmic difference was observed in the VC and CVC syllable, in which the coda segment is nasal sound. This study allowed us to observe the influence of L1 on production of L2 rhythm.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.391-395
• /
• 2021

텍스트 스타일 변환은 입력 스타일(source style)로 쓰여진 텍스트의 내용(content)을 유지하며 목적 스타일(target style)의 텍스트로 변환하는 문제이다. 텍스트 스타일 변환을 시퀀스 간 변환 문제(sequence-to-sequence)로 보고 기존 기계학습 모델을 이용해 해결할 수 있지만, 모델 학습에 필요한 각 스타일에 대응되는 병렬 말뭉치를 구하기 어려운 문제점이 있다. 따라서 최근에는 비병렬 말뭉치를 이용해 텍스트 스타일 변환을 수행하는 방법들이 연구되고 있다. 이 연구들은 주로 인코더-디코더 구조의 생성 모델을 사용하기 때문에 입력 문장이 가지고 있는 내용이 누락되거나 다른 내용의 문장이 생성될 수 있는 문제점이 있다. 본 논문에서는 마스크 언어 모델(masked language model)을 이용해 입력 텍스트의 내용을 유지하면서 원하는 스타일로 변경할 수 있는 텍스트 스타일 변환 방법을 제안하고 한국어 긍정-부정, 채팅체-문어체 변환에 적용한다.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.19 no.2
• /
• pp.73-81
• /
• 2024

Recently, mobile manipulators have been utilized in agriculture industry for weed removal and harvest automation. This paper proposes a domain adaptive fruit detection method for harvest automation, by utilizing OWL-ViT model which is an open-vocabulary object detection model. The vision-language model can detect objects based on text prompt, and therefore, it can be extended to detect objects of undefined categories. In the development of deep learning models for real-world problems, constructing a large-scale labeled dataset is a time-consuming task and heavily relies on human effort. To reduce the labor-intensive workload, we utilized a large-scale public dataset as a source domain data and employed a domain adaptation method. Adversarial learning was conducted between a domain discriminator and feature extractor to reduce the gap between the distribution of feature vectors from the source domain and our target domain data. We collected a target domain dataset in a real-like environment and conducted experiments to demonstrate the effectiveness of the proposed method. In experiments, the domain adaptation method improved the AP50 metric from 38.88% to 78.59% for detecting objects within the range of 2m, and we achieved 81.7% of manipulation success rate.

• English Language & Literature Teaching
• /
• v.9 no.spc
• /
• pp.99-112
• /
• 2003

This study investigated how different types of formal instruction affect the second language looming of English grammatical structure among Korean high-school students. The linguistic focus of the study was English present perfect, which often creates learning problems for Korean learners of English. Subjects were divided into a control group and an experimental group (Enhanced group). The input the subjects in the experimental group received was manipulated by visually enhancing (with highlighting of the target structures in a reading text). Learners' awareness of the rules throughout the treatment period, as well as accuracy of target structures was measured. Results indicated that subjects in the Enhanced group showed higher performance than the control group. Further, awareness of rules that learners developed over the treatment period did not provide any advantage in learning.

• MALSORI
• /
• no.66
• /
• pp.41-59
• /
• 2008

Knowledge in linguistic rhythm of the target language plays a major role in foreign language proficiency. This study attempts to discover valid rhythm features that can be utilized in automatic assessment of non-native English pronunciation. Eight previously proposed and two novel rhythm metrics are investigated with 360 English read speech tokens obtained from 27 Korean learners and 9 native speakers. It is found that some of the speech-rate normalized interval measures and above-word level metrics are effective enough to be further applied for automatic scoring as they are significantly correlated with speakers' proficiency levels. It is also shown that metrics need to be dynamically selected depending upon the structure of target sentences. Results from a preliminary auto-scoring experiment through a Multi Regression analysis suggest that appropriate control of unexpected input utterances is also desirable for better performance.