• International Journal of Computer Science & Network Security
• /
• 제22권6호
• /
• pp.346-356
• /
• 2022

The jurisprudential legal rules govern the way Muslims react and interact to daily life. This creates a huge stream of questions, that require highly qualified and well-educated individuals, called Muftis. With Muslims representing almost 25% of the planet population, and the scarcity of qualified Muftis, this creates a demand supply problem calling for Automation solutions. This motivates the application of Artificial Intelligence (AI) to solve this problem, which requires a well-designed Question-Answering (QA) system to solve it. In this work, we propose a QA system, based on retrieval augmented generative transformer model for jurisprudential legal question. The main idea in the proposed architecture is the leverage of both state-of-the art transformer models, and the existing knowledge base of legal sources and question-answers. With the sensitivity of the domain in mind, due to its importance in Muslims daily lives, our design balances between exploitation of knowledge bases, and exploration provided by the generative transformer models. We collect a custom data set of 850,000 entries, that includes the question, answer, and category of the question. Our evaluation methodology is based on both quantitative and qualitative methods. We use metrics like BERTScore and METEOR to evaluate the precision and recall of the system. We also provide many qualitative results that show the quality of the generated answers, and how relevant they are to the asked questions.

• 인터넷정보학회논문지
• /
• 제23권2호
• /
• pp.71-77
• /
• 2022

본 연구에서는 음성 합성 모델을 감정에 따라 음성을 합성하는 모델로 확장하고 감정에 따른 얼굴 표정을 생성하기 위한 데이터 베이스를 수집한다. 데이터베이스는 남성과 여성의 데이터가 구분되며 감정이 담긴 발화와 얼굴 표정으로 구성되어 있다. 성별이 다른 2명의 전문 연기자가 한국어로 문장을 발음한다. 각 문장은 anger, happiness, neutrality, sadness의 4가지 감정으로 구분된다. 각 연기자들은 한 가지의 감정 당 약 3300개의 문장을 연기한다. 이를 촬영하여 수집한 전체 26468개의 문장은 중복되지 않으며 해당하는 감정과 유사한 내용을 담고 있다. 양질의 데이터베이스를 구축하는 것이 향후 연구의 성능에 중요한 역할을 하므로 데이터베이스를 감정의 범주, 강도, 진정성의 3가지 항목에 대해 평가한다. 데이터의 종류에 따른 정확도를 알아보기 위해 구축된 데이터베이스를 음성-영상 데이터, 음성 데이터, 영상 데이터로 나누어 평가를 진행하고 비교한다.

• 한국기록관리학회지
• /
• 제23권1호
• /
• pp.101-126
• /
• 2023

대학의 사진기록은 대학의 정체성을 확립하고 역사적 증거를 제공하는 매우 가치 있는 기록의 유형 중 하나이나, 텍스트와 달리 의미전달의 취약성을 갖고 있으므로 사진기록의 정보가 포괄적으로 기술되지 않으면 이용자의 검색과 활용이 어렵다. 본 연구에서는 대학기록관 사진 아카이브를 위해 사진기록의 분류체계를 구조화하고, 분류 내의 카테고리 특성을 반영한 메타데이터 셋 개발을 시도하였다. 이를 위해 국내와 미국 대학기록관의 사진기록 분류체계와 메타데이터 요소를 분석하고, 정보구조 모형을 제안하였다. 본 연구에서 제안한 정보구조 모형을 통해 대학기록관 사진기록의 데이터 품질을 향상시킬 수 있으며, 이용자에게는 사진기록에 대한 풍부한 디스커버리를 지원할 수 있다.

• Nutrition Research and Practice
• /
• 제17권2호
• /
• pp.316-329
• /
• 2023

BACKGROUND/OBJECTIVES: This study investigated the association between dietary quality based on the Korean Healthy Eating Index (KHEI), and the prevalence of chronic conditions among middle-aged individuals (40-60 yrs of age) living alone. MATERIALS/METHODS: The participants were selected (1,517 men and 2,596 women) from the Korea National Health and Nutrition Examination Survey (KNHANES) 2016-2018 and classified into single-person households (SPH) and multi-person households (MPH). Nutrient intake, KHEI, and the prevalence of chronic conditions were compared according to household size. The odds ratios (ORs) of chronic conditions were analyzed according to the tertile levels of KHEI by gender within each household size category. RESULTS: Men in SPH had a significantly lower total KHEI score (P < 0.0001) and a lower prevalence of obesity (OR, 0.576) than those in MPH. For men, the adjusted ORs for obesity, hypertension, and hypertriglyceridemia in the first tertile (T1) of KHEI scores within SPH compared with the third tertile (T3) were 4.625, 3.790, and 4.333, respectively. Moreover, the adjusted OR for hypertriglyceridemia in the T1 group compared to the T3 group within the MPH was 1.556. For women, the adjusted ORs for obesity and hypertriglyceridemia in T1 compared to T3 within the SPH were 3.223 and 7.134, respectively, and 1.573 and 1.373 for obesity and hypertension, respectively, within MPH. CONCLUSIONS: A healthy eating index was associated with a reduced risk of chronic conditions in middle-aged adults. Greater adherence to a healthy eating index could lower the risk of chronic conditions in middle-aged adults living alone.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2021년도 춘계학술발표대회
• /
• pp.411-414
• /
• 2021

One of the elements of technology that has become extremely critical within the field of education today is Deep learning. It has been especially used in the area of natural language processing, with some word-representation vectors playing a critical role. However, some of the low-resource languages, such as Swahili, which is spoken in East and Central Africa, do not fall into this category. Natural Language Processing is a field of artificial intelligence where systems and computational algorithms are built that can automatically understand, analyze, manipulate, and potentially generate human language. After coming to discover that some African languages fail to have a proper representation within language processing, even going so far as to describe them as lower resource languages because of inadequate data for NLP, we decided to study the Swahili language. As it stands currently, language modeling using neural networks requires adequate data to guarantee quality word representation, which is important for natural language processing (NLP) tasks. Most African languages have no data for such processing. The main aim of this project is to recognize and focus on the classification of words in English, Swahili, and Korean with a particular emphasis on the low-resource Swahili language. Finally, we are going to create our own dataset and reprocess the data using Python Script, formulate the syllabic alphabet, and finally develop an English, Swahili, and Korean word analogy dataset.

• 한국인터넷방송통신학회논문지
• /
• 제22권6호
• /
• pp.113-119
• /
• 2022

소프트웨어 결함 예측 작업 시 단순히 결함 유무만을 예측하는 이진 분류 모델에 비해 결함의 심각도 범주를 예측하는 다중 분류 모델은 훨씬 유용하게 사용될 수 있다. 소수의 심각도 기반 결함 예측 모델들이 제안되었지만 딥러닝 기법을 사용한 분류기는 없었다. 본 논문은 3개, 5개의 은닉층을 갖고 은닉층 노드수가 고정된 구조와 변화하는 구조의 MLP 모델들을 제작하였다. 모델 평가 실험 결과 기존 기계학습 모델들 중 가장 좋은 성능을 보인 MLPs보다 MLP 기반 딥러닝 모델들은 Accuracy와 AUC 모두 유의미하게 더 우수한 성능을 보였다. 특히 노드수 고정 구조에서는 은닉 층수 3, 배치사이즈 32, 노드수 64인 모델 구조가 가장 좋은 성능을 보였다.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2020년도 학술발표회
• /
• pp.168-168
• /
• 2020

This study is to evaluate the drought-related bigdata characteristics published from South Korean by developing crawler. The 5 years (2013 ~ 2017) drought-related posted articles were collected from Korean internet search engine 'NAVER' which contains 13 main and 81 local daily newspapers. During the 5 years period, total 40,219 news articles including 'drought' word were found using crawler. To filter the homonyms liken drought to soccer goal drought in sports, money drought economics, and policy drought in politics often used in South Korea, the quality control was processed and 47.8 % articles were filtered. After, the 20,999 (52.2 %) drought news articles of this study were classified into four categories of water deficit (WD), water security and support (WSS), economic damage and impact (EDI), and environmental and sanitation impact (ESI) with 27, 15, 13, and 18 drought-related keywords in each category. The WD, WSS, EDI, and ESI occupied 41.4 %, 34.5 %, 14.8 %, and 9.3 % respectively. The drought articles were mostly posted in June 2015 and June 2017 with 22.7 % (15,097) and 15.9 % (10,619) respectively. The drought news articles were spatiotemporally compared with SPI (Standardized Precipitation Index) and RDI (Reservoir Drought Index) were calculated. They were classified into administration boundaries of 8 main cities and 9 provinces in South Korea because the drought response works based on local government unit. The space-time clustering between news articles (WD, WSS, EDI, and ESI) and indices (SPI and RDI) were tried how much they have correlation each other. The spatiotemporal clusters detection was applied using SaTScan software (Kulldorff, 2015). The retrospective and prospective cluster analyses were conducted for past and present time to understand how much they are intensive in clusters. The news articles of WD, WSS and EDI had strong clusters in provinces, and ESI in cities.

• Asia Marketing Journal
• /
• 제14권3호
• /
• pp.169-184
• /
• 2012

All firms require new stimulus to spurt growth. Therefore it is necessary to successfully develop new products and to employ appropriate marketing practices for the new products to enter and settle in the market. Binggrae, a Korean company that specializes in dairy and processed dairy products, introduced a ready-to-drink (RTD) coffee product, 'a Café la' to expand its business into the coffee market in 2008. Binggrae was a latecomer in the RTD coffee market but a Café la has shown an impressive average sales growth rate of 115% as of 2011 since the launch. Moreover, it is a steady bestselling coffee brand among the Polyethylene terephthalate (PET)bottle category. Binggrae found potential and opportunity in the growing coffee market and made efforts to develop a new product that can be differentiated from the existing products. The result was PET bottle coffee, which was more portable and convenient to drink than coffee products offered in cups or cans. PET bottle coffee is produced through the patented Aseptic Filing System, thus the original coffee flavor stays fresh when combined with milk and has a longer shelf life than coffee products in cups. Moreover, as the taste of coffee consumers has become more sophisticated, Binggrae developed a premium product by differentiating the product processing method and by using higher-quality Arabica beans. After launching the new product, the company also employed a well-designed communication strategy. First, Binggrae was able to confirm the level of market demand and market potential for the product by employing BTL (Below the Line) marketing strategies through the consumers' word-of-mouth. Afterwards, the company invested its resources for a full-scale ATL (Above the Line) marketing campaign. Later a Café la's TV commercial effectively portrayed the product's characteristics, and succeeded in raising consumer awareness of the product. As a result, a Café la has become the bestselling brand in the PET bottle coffee market. The successful new product marketing strategy of Binggrae'sa Café la offers many valuable implications for companies planning to launch new products in the future.

• 한국의류산업학회지
• /
• 제25권5호
• /
• pp.578-592
• /
• 2023

This study identified and evaluated by deriving and categorizing concepts related to the user experience of metaverse fashion shows using grounded theory, which is a qualitative research method. Based on experiential marketing theory, in-depth interviews were conducted for 14 days with 14 males and females in their 20s and 30s. The research results and contents are as follows: The causal condition was the purpose of using metaverse fashion shows, and the action/interaction strategy caused by such a case was found to be establishing a system for metaverse fashion shows and promoting a positive brand image. The results included content evaluation of satisfaction, normal, or dissatisfaction. The contextual condition was a change in the form of consumption that emphasized experience, while the interventional condition was psychological distance. Based on this, the core category was defined as "consumption patterns that emphasized the purpose of use and experience affects the metaverse fashion shows and psychological distance appeared as a user experience evaluation through the establishment of a system of metaverse fashion shows and the promotion of a positive brand image". User types were classified as active or passive. Active users have the autonomy to select content according to their individual preferences, and accordingly, their experience preference tends to change. In contrast, passive users' preference for the technical quality of content is relatively low, but they have a high concentration of content diversity and audio-visual interest elements.

• 한국정보전자통신기술학회논문지
• /
• 제16권6호
• /
• pp.369-377
• /
• 2023

산업 제조 분야에서 품질 관리는 불량률을 최소화하는 핵심 요소로, 미흡한 관리는 추가적인 비용 발생과 생산 지연을 야기할 수 있다. 본 연구는 제조품의 텍스쳐 결함 감지의 중요성을 중심으로, 보다 정밀한 결함 감지 방법을 제시한다. DFR(Deep Feature Reconstruction) 모델은 특징맵의 조합 및 재구성을 통한 접근법을 채택하였지만, 그 방식에는 한계가 있었다. 이에 따라, 우리는 제한점을 극복하기 위해 통계적 방법론을 활용한 새로운 손실 함수와 스킵 연결구조를 통합하고 파라미터 튜닝을 진행하였다. 이 개선된 모델을 MVTec-AD 데이터세트의 텍스쳐 카테고리에 적용한 결과, 기존 방식보다 2.3% 높은 결함 분할 AUC를 기록하였고, 전체적인 결함 감지 성능도 향상되었다. 이 결과는 제안하는 방법이 특징맵 조합의 재건축을 통한 결함 탐지에 있어서 중요한 기여함을 입증한다.