• Smart Media Journal
• /
• v.10 no.1
• /
• pp.39-46
• /
• 2021

This study is about a method of extracting a summary from a news article in consideration of the importance of each sentence constituting the article. We propose a method of calculating sentence importance by extracting the probabilities of topic sentence, similarity with article title and other sentences, and sentence position as characteristics that affect sentence importance. At this time, a hypothesis is established that the Topic Sentence will have a characteristic distinct from the general sentence, and a deep learning-based classification model is trained to obtain a topic sentence probability value for the input sentence. Also, using the pre-learned ELMo language model, the similarity between sentences is calculated based on the sentence vector value reflecting the context information and extracted as sentence characteristics. The topic sentence classification performance of the LSTM and BERT models was 93% accurate, 96.22% recall, and 89.5% precision, resulting in high analysis results. As a result of calculating the importance of each sentence by combining the extracted sentence characteristics, it was confirmed that the performance of extracting the topic sentence was improved by about 10% compared to the existing TextRank algorithm.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.12
• /
• pp.569-578
• /
• 2021

Thanks to web development today, we can easily access online news via various media. As much as it is easy to access online news, we often face fake news pretending to be true. As fake news items have become a global problem, fact-checking services are provided domestically, too. However, these are based on expert-based manual detection, and research to provide technologies that automate the detection of fake news is being actively conducted. As for the existing research, detection is made available based on contextual characteristics of an article and the comparison of a title and the main article. However, there is a limit to such an attempt making detection difficult when manipulation precision has become high. Therefore, this study suggests using a verifying article to decide whether a news item is genuine or not to be affected by article manipulation. Also, to improve the precision of fake news detection, the study added a process to summarize a subject article and a verifying article through the summarization model. In order to verify the suggested algorithm, this study conducted verification for summarization method of documents, verification for search method of verification articles, and verification for the precision of fake news detection in the finally suggested algorithm. The algorithm suggested in this study can be helpful to identify the truth of an article before it is applied to media sources and made available online via various media sources.

• Annual Conference on Human and Language Technology
• /
• 2005.10a
• /
• pp.167-172
• /
• 2005

지금까지 자연언어처리에서의 품사태깅(parts-of-speech tagging) 기술에 대한 연구는 활발히 진행된 반면, 전문용어에 대한 처리 기술은 미비한 점이 많았다. 전문용어에 관련된 연구는 대부분 구축, 표준화, 추출 등에 대한 연구가 많았으나 전문용어 태그 설정과 태깅 기술 연구는 부족한 상황이다. 본 논문에서는 전문용어 태그를 (분야정보: 아이디) 순으로 설정하고 백과사전의 분류 체계를 이용하여 어떤 특정 분야 문서의 전문용어를 자동으로 태깅하는 시스템을 구축하였다. 전문용어 태깅 시스템은 형태소분석기를 사용하지 알고 문맥의 규칙과 조사 어미사전을 이용해 자동으로 태깅을 하게 된다. 이 시스템의 정확률 측정을 위한 정답말뭉치는 웹 상에 공개되어 있는 백과사전 html문서를 이용하였다. 우선 백과사전에 나와있는 용어는 전문용어라고 가정한다. 하나의 문서에는 '용어', '요약', '본문', '이미지', '분류', '참조항목' 등의 정보들이 있다. 이 중 '본문'에는 그 용어에 대한 자세한 설명이 있는데 특정 단어에는 태그로 백과사전 내에 있는 단어를 찾아 볼 수 있게 링크 되어있다. 이 정보를 이용해 태그로 되어있는 것을 설정한 태그로 바꾸고 단계별로 확장 태깅을 해서 정답말뭉치를 만든다. 태깅 시스템과 정답말뭉치를 비교해 정확률을 계산해서 시스템의 성능을 측정하였다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.382-387
• /
• 2023

현재 GPT-4와 같은 거대한 언어 모델이 기계 번역, 요약 및 대화와 같은 다양한 작업에서 압도적인 성능을 보이고 있다. 그러나 이러한 거대 언어 모델은 학습 및 적용에 상당한 계산 리소스와 도메인 특화 미세 조정이 어려운 등 몇 가지 문제를 가지고 있다. In-Context learning은 데이터셋에서 추출한 컨택스트의 정보만으로 효과적으로 작동할 수 있는 효율성을 제공하여 앞선 문제를 일부 해결했지만, 컨텍스트의 샷 개수와 순서에 민감한 문제가 존재한다. 이러한 도전 과제를 해결하기 위해, 우리는 Super In-Context Learning (SuperICL)을 활용한 새로운 방법론을 제안한다. 기존의 SuperICL은 적용한 플러그인 모델의 출력 정보를 이용하여 문맥을 새로 구성하고 이를 활용하여 거대 언어 모델이 더욱 잘 분류할 수 있도록 한다. Super In-Context Learning for Generation은 다양한 자연어 생성 작업에 효과적으로 최적화하는 방법을 제공한다. 실험을 통해 플러그인 모델을 교체하여 다양한 작업에 적응하는 가능성을 확인하고, 자연어 생성 작업에서 우수한 성능을 보여준다. BLEU 및 ROUGE 메트릭을 포함한 평가 결과에서도 성능 향상을 보여주며, 선호도 평가를 통해 모델의 효과성을 확인했다.

• Journal of KIISE:Software and Applications
• /
• v.35 no.6
• /
• pp.392-403
• /
• 2008

User reviews are valuable information that can be used for various purposes. In particular, the product reviews on online shopping sites are important information which can directly affect the purchasing decision of the customers. In this paper, we present our design and implementation of a system for summarizing the customer's opinion and the features of each product by analyzing reviews on a commercial shopping site. During the analysis process, several natural language processing(NLP) techniques and the semantic dictionary were used. The semantic dictionary contains vocabularies that are used to express product features and customer's opinions. And it was constructed in semi-automatic way with the help of the tool we implemented. Furthermore, we discuss how to handle the vocabularies that have different meanings according to the context. We analyzed 1796 reviews about 20 products of 2 categories collected from an actual shopping site and implemented a novel ranking system. We obtained 88.94% for precision and 47.92% for recall on extracting opinion expression, which means our system can be applicable for real use.

• Journal of the Korea Convergence Society
• /
• v.12 no.2
• /
• pp.51-60
• /
• 2021

Due to the growth of digital media technology, there is increasing demand of personalized education based on context data of learners throughout overall education area. For music education, several studies have been conducted for providing appropriate educational contents to learners by considering some factors such as the proficiency, the amount of practice, and their capability. In this paper, a technique has been defined to recommend the appropriate music scores to learners by extracting and analyzing the practice data and chord data. Concretely, several meaningful relationships among chords patterns and learners were analyzed and visualized by constructing the learners' profiles of proficiency, extracting the chord sequences from music scores. In addition, we showed the potential for use in personalized education by analyzing music similarity, learner's proficiency similarity, learner's proficiency of music and chord, mastered chords and chords sequence patterns. After that, the chord practice programs can be effectively generated considering various music scores using the synthetically summarized chord sequence graphs for the music scores that the learners selected.

• Journal of KIISE:Software and Applications
• /
• v.34 no.6
• /
• pp.539-548
• /
• 2007

Mobile content adaptation is a technology of effectively representing the contents originally built for the desktop PC on wireless mobile devices. Previous approaches for Web content adaptation are mostly device-dependent. Also, the content transformation to suit to a smaller device is done manually. Furthermore, the same contents are provided to different users regardless of their individual preferences. As a result, the user has difficulty in selecting relevant information from a heavy volume of contents since the context information related to the content is not provided. To resolve these problems, this paper proposes an enhanced method of Web content adaptation for mobile devices. In our system, the process of Web content adaptation consists of 4 stages including block filtering, block title extraction, block content summarization, and personalization through learning. Learning is initiated when the user selects the full content menu from the content summary page. As a result of learning, personalization is realized by showing the information for the relevant block at the top of the content list. A series of experiments are performed to evaluate the content adaptation for a number of Web sites including online newspapers. The results of evaluation are satisfactory, both in block filtering accuracy and in user satisfaction by personalization.

• Journal of The Korean Association For Science Education
• /
• v.43 no.3
• /
• pp.307-319
• /
• 2023

This study aims to explain the key concepts and principles of text-based generative artificial intelligence (AI) that has been receiving increasing interest and utilization, focusing on its application in science education. It also highlights the potential and limitations of utilizing generative AI in science education, providing insights for its implementation and research aspects. Recent advancements in generative AI, predominantly based on transformer models consisting of encoders and decoders, have shown remarkable progress through optimization of reinforcement learning and reward models using human feedback, as well as understanding context. Particularly, it can perform various functions such as writing, summarizing, keyword extraction, evaluation, and feedback based on the ability to understand various user questions and intents. It also offers practical utility in diagnosing learners and structuring educational content based on provided examples by educators. However, it is necessary to examine the concerns regarding the limitations of generative AI, including the potential for conveying inaccurate facts or knowledge, bias resulting from overconfidence, and uncertainties regarding its impact on user attitudes or emotions. Moreover, the responses provided by generative AI are probabilistic based on response data from many individuals, which raises concerns about limiting insightful and innovative thinking that may offer different perspectives or ideas. In light of these considerations, this study provides practical suggestions for the positive utilization of AI in science education.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.6
• /
• pp.239-250
• /
• 2020

The purpose of this paper is to explore the influence of the "contextualization" effect of web celebrity live broadcast on the e-commerce platform on consumers' perception of product value, risk and purchase intention. Live in this paper, using Taobao shopping consumers as the research object, the survey method, questionnaire survey is adopted, the form through the questionnaire and distributed network, a live in order to further validation of web celebrity effect of contextualized actual influence on consumer purchase intention, questionnaire design the Likert scale, seven and recycling questionnaire analysis using the statistical software SPSS 23.0 and AMOS 22.0 after processing the data. After determining the reliability and validity of the questionnaire, the exploratory factor analysis was used to verify the hypothesis and calculate the actual adjustment degree of the "contextualization" effect of web celebrity live broadcasting on consumers' purchase intention. The research results of this paper are summarized as follows :(1) consumers' perceived value of products can significantly positively affect their purchase intention, while perceived risk has a significantly negative impact on their purchase intention; (2) consumers' trust and purchase intention to products are regulated by the "contextualization" of web celebrity live broadcast. Specifically, for web celebrity live broadcasting with good "contextualization" effect, the perceived value of consumer products has a positive impact on product trust, which is higher than that of web celebrity live broadcasting with poor "contextualization" effect. In terms of resolving consumers' perceived risks to products, web celebrity live broadcast with good "contextualization" effect is also significantly better than web celebrity live broadcast with poor "contextualization" effect. Based on empirical analysis, this paper concludes that web celebrity live broadcasting will become a new breakthrough for the sustainable growth of the e-commerce industry, and puts forward Suggestions on the e-commerce marketing mode and the transformation of web celebrity live broadcasting industry.

• Advances in pediatric surgery
• /
• v.8 no.2
• /
• pp.143-155
• /
• 2002

A survey on biliary atresia was made among 26 members of the Korean Association of Pediatric Surgeons. The members were required to complete a questionnaire and a case registration form for each patient during the twentyone-year period of 1980-2000. Three hundred and eighty patients were registered from 18 institutions. The average number of patients per surgeon was one to two every year. The male to female ratio was 1:1.3. The age of patients on diagnosis with biliary atresia was on average $65.4{\pm} 36.2$ days old. The national distribution was 32.8% in Seoul, 25.3% in Gyoungki-Do, 21.6% in Gyoungsang-Do, 9.27% in Choongchung-Do, etc. in order. The most common clinical presentation was jaundice (98.4%) and change of stool color (86.2%) was second. Two hundred eighty (74.7%) of 375 patients were operated by 80 days of age. Three hundred thirty six (9 1.9%) of 366 patients were operated on by the original Kasai procedure, and 305 (84.3%) of 362 patients were observed by bile-drainage postoperatively. The overall postoperative complication rate was 18.5% and the overall postoperative mortality rate was 6.8%. The associated anomalies were observed in 72 cases (22.5%). One hundred ninty five (64.7%) of 302 patients have been alive in follow-up and 49 (25.1%) have survived over 5 years without problem after operation. Ascending cholangitis, varices and ascites affected survival significantly, and the important long-term prognostic factor was the occurrence of complications.