• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.698-703
• /
• 2023

최근 ChatGPT와 같은 사전학습 생성모델은 자연어 이해 (natural language understanding)에서 좋은 성능을 보이고 있다. 또한 코드 작업을 도와주고 대학수학능력시험, 중고등학교 수준의 문제를 풀거나 도와주는 다양한 분야에서 활용되고 있다. 본 논문은 사전학습 생성모델을 이용하여 영어 교육을 위해 말뭉치를 확장하는 프레임 워크를 제시한다. 이를 위해 ChatGPT를 사용해 말뭉치를 확장 한 후 의미 유사도, 상황 유사도, 문장 교육 난이도를 사용해 생성된 문장의 교육적 효과를 검증한다.

• Journal of the Korea Convergence Society
• /
• v.12 no.12
• /
• pp.365-370
• /
• 2021

The current study aims to investigate the relationship between working memory capacity, vocabulary size, and listening competence of Korean EFL (English as a Foreign Language) learners. 30 English education majors from a university in Korea were recruited. The backward digit span and the operation span tasks were used to measure the participants' working memory capacity, while the Listening Vocabulary Level Test (LVLT) and the Michigan English Test (MET) were employed to measure their vocabulary size and listening proficiency in English, respectively. The correlational analyses revealed that the bigger one's working memory storage was, the better the person processed incoming input. However, no statistically significant correlation was found between working memory capacity, English vocabulary size, and listening proficiency, possibly due to the small sample size and the homogeneous subjects.

• The Journal of the Korea Contents Association
• /
• v.20 no.5
• /
• pp.344-350
• /
• 2020

The purpose of this study was to investigate the cohesion-related text factors that affected the correct answer rate of the English reading section in the Korean CSAT. To this end, reading passages of the 2014-2018 CSAT and CSAT mock test were collected and analyzed in terms of cohesion. And then, correlation analysis was implemented between the results regarding cohesion and item difficulty. The conclusions were as follows. First, the correct answer rate tended to increase as the linguistic burden of learners was reduced, when specific arguments between sentences were repeated. Second, there was no statistically significant correlation between the 8 sub-area of latent semantic analysis and the correct answer rate. Finally, the pedagogical implications of this study and suggestions for further study were discussed.

• Journal of the Korean Geographical Society
• /
• v.48 no.5
• /
• pp.786-801
• /
• 2013

This study aims to elucidate features of the geography assessment of France in which geography has a high status in school curriculum. The main data for analysing are the geography examinations(1999-2013) of the Baccalaur$\acute{e}$at. In terms of objectives of the assessment, the exams put emphasis on promoting cultural knowledges of students and developing their competences in graphic languages. In terms of formal aspect, the exams take a simple layouts to raise a readability, use variable test tools, especially an essay and a 'croquis'. They also emphasize students self-directedly constructing and representing geography knowledges and their inquiry abilities with geography materials. In terms of contents aspect, the exams mainly cover subjects of political-economic geography, particularly a spatial organization, a development inequality and the world powers in global contexts or macro spatial scales. One of a distinguishing characteristics of Baccalaur$\acute{e}$at's History-Geography examinations is that it designs to keep balance between two parties; history and geography.

• Annual Conference of KIPS
• /
• 2023.11a
• /
• pp.21-22
• /
• 2023

영어 공인 시험에서 영어 어휘는 가장 중요한 요소 중 하나이다. 그러나 기존 영어 어휘 학습 시스템은 단순히 어휘 교재의 콘텐츠를 디지털화한 것에 그치고 있다. 영어 어휘 학습에서 동의어에 대한 학습은 언어 능력을 향상시키고 다양한 상황에서 효과적으로 의사소통할 수 있는 능력 향상에 도움이 된다. 본 연구에서는 학습자들이 영어 어휘 학습을 효과적으로 할 수 있도록 동의어 및 반의어를 활용할 수 있는 학습 시스템을 제안한다. 본 시스템은 동의어와 반의어에 대한 정보가 필요한 영어 어휘를 입력으로 받아 생성형 AI에서 동의어 및 반의어에 대한 정보를 생성한 후에 반환하는 기능을 수행한다. 본 시스템은 학습자들이 어휘 평가에서 오답으로 답한 어휘에 대해 동의어 및 반의어에 대한 추가적인 어휘와 예문을 제공하여 영어 어휘 학습의 효율성을 높일 수 있을 것이다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.10
• /
• pp.6014-6028
• /
• 2014

This study examined the levels of language anxiety and their causal factors in different English class environments. The findings show that the levels of language anxiety vary significantly in different types of class environments, indicating that both the class environments and individual characteristics are meaningful variables of language anxiety. The findings also show that unlike the levels of language anxiety, its causal factors are almost constant regardless of the class environment with communication apprehension being the most important factor. This study suggests that the English proficiency of students and professors needs to be considered when constructing specific class environments, in addition to promoting class environments favorable to students' active participation in class.

• KIPS Transactions on Computer and Communication Systems
• /
• v.12 no.10
• /
• pp.291-298
• /
• 2023

In this paper, we study whether the programming questions of the Informatics·Computer recruitment tests were suitable for selecting teachers with required programming skills. The average points of the programming questions constituted 38%(20.8 points) of the total scores for the entire curriculum based on the results from analyzing the previous questions in the past 5 years. Moreover, the distribution of points for each evaluation criteria within programming and data structure, two exam subjects which have a high proportion of programming questions, demonstrated a large deviation ranging from 0% to 47% and 0% to 53% respectively. In this study, a questionnaire survey was conducted on 31 teachers to examine if the previous programming questions were suitable for measuring teachers' competency in programming abilities required in the actual teaching experience. Computational thinking ability was ranked the highest at 58% in response to the area that needs to be evaluated in the recruitment test. In response to the relevance of previous questions, problem solving ability was ranked the highest at 2.84 on a 5-point scale, but the overall appropriateness was deemed low. C language and Python were regarded as the computer languages suitable to be tested for programming questions with each ranked 55% and 45%. The finding confirms that teachers preferred Python and the incumbent C language to others. Based on the results of the questionnaire, we recommend changes in the programming questions to improve the selection criteria.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2001.11a
• /
• pp.89-109
• /
• 2001

The purpose of this research was to develop Public Service Aptitude Test(PSAT) in order to select high performer in the public sector of the government. Based on both literature review and job analysis using the interview with job incumbents, the test which measured verbal ability, situational judgment ability, data interpretation ability, and knowledge on common sense was prepared. The items was administered in undergraduate students, job incumbents, and trainees of public service officers to obtain information about the goodness and appropriateness of items. Base on the results of the item analysis, new items were made and administered in another undergraduate students, job incumbents, and trainees of public service officers to select the final test items. The reliability and validity of the developed PSAT were acceptable level. Finally, the impact of the PSAT on potential applicants and future suggestions were discussed.

• English Language & Literature Teaching
• /
• v.9 no.spc
• /
• pp.79-98
• /
• 2003

This article reports that a testee's content schemata have detrimental effects on the validity of standardized reading comprehension tests. in case that the reading passages constitute his prior background knowledge framed by regular academic pursuits focused on a major field of study. The empirical research findings on such a reliability issue have shown that the Korean subjects' familiarity with the content domain of the reading materials may substantially enhance their test performance, operated as unfair advantages over the other participants at the same level of English proficiency who do not possess technical knowledge closely related to the contents of the test passages. Based on these results, it may be argued that the reading text should be composed of maximally neutral contents to every testee to avoid the possible overlapping between a test-taker's content schemata and the content domain of the materials. Therefore, this article claims that the tenets of schema theory with its long-standing foundation on culturally predetermined content and formal schemata need to be expanded to cover professional knowledge structures, like those acquired by higher education, as the hard-wired framework of an individual's original culture-specific background knowledge.

• Journal for History of Mathematics
• /
• v.20 no.1
• /
• pp.83-102
• /
• 2007

A math class in Korean university was taught in English for one semester and the students' improvement was measured in math content and English proficiency. Pre and post test in 9 week intervals showed that math content loaming in the immersion class was superior to the non-immersed class. Especially, the immersion class showed remarkable improvement in difficult problems among math content test problems. The immersion class improved in math-related English, but not in general English. It is discussed that the particular English expressions for math are hardly separable from the math content knowledge in English because understanding and using those expressions correctly means the students' understanding of math concept in English and thus the math concept itself.