• Annual Conference of KIPS
• /
• 2023.11a
• /
• pp.514-517
• /
• 2023

본 연구는 인 컨택스트 러닝 (In-Context Learning)을 오디오-언어 작업에 적용하기 위한 멀티모달 (Multi-Modal) 딥러닝 모델을 다룬다. 해당 모델을 통해 학습 단계에서 오디오와 텍스트의 소통 가능한 형태의 표현 (Representation)을 학습하고 여러가지 오디오-텍스트 작업을 수행할 수 있는 멀티모달 딥러닝 모델을 개발하는 것이 본 연구의 목적이다. 모델은 오디오 인코더와 언어 인코더가 연결된 구조를 가지고 있으며, 언어 모델은 6.7B, 30B 의 파라미터 수를 가진 자동회귀 (Autoregressive) 대형 언어 모델 (Large Language Model)을 사용한다 오디오 인코더는 자기지도학습 (Self-Supervised Learning)을 기반으로 사전학습 된 오디오 특징 추출 모델이다. 언어모델이 상대적으로 대용량이기 언어모델의 파라미터를 고정하고 오디오 인코더의 파라미터만 업데이트하는 프로즌 (Frozen) 방법으로 학습한다. 학습을 위한 과제는 음성인식 (Automatic Speech Recognition)과 요약 (Abstractive Summarization) 이다. 학습을 마친 후 질의응답 (Question Answering) 작업으로 테스트를 진행했다. 그 결과, 정답 문장을 생성하기 위해서는 추가적인 학습이 필요한 것으로 보였으나, 음성인식으로 사전학습 한 모델의 경우 정답과 유사한 키워드를 사용하는 문법적으로 올바른 문장을 생성함을 확인했다.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.5
• /
• pp.1089-1100
• /
• 2024

Despite various measures being implemented by the government and related institutions to prevent voice phishing, incidents of such fraud continue to occur. This study analyzed 448 actual conversations between voice phishing scammers and potential victims using text mining techniques. The text analysis reveals that voice phishing scammers frequently use words emphasizing limited time frames such as now, soon, in progress, today, first. This indicates that scammers manipulate the victim's psychology through specific words, preventing them from making rational decisions. The results of this study can aid government and related institutions in formulating effective policies for preventing voice phishing and protecting consumers.

• Smart Media Journal
• /
• v.13 no.9
• /
• pp.18-26
• /
• 2024

The development of information and communication technology and advanced technologies has led to the full-fledged implementation of smart tourism in the tourism industry. Concurrently, both domestic and international research on smart tourism has been actively pursued. This study aimed to compare and analyze the trends and issues in smart tourism research between South Korea and China, to identify differences in issues related to smart tourism in both countries, and to suggest directions for future research in smart tourism. For this purpose, a text mining analysis was conducted on 69 Korean papers and 42 Chinese papers related to smart tourism published from January 2020 to June 2024. The analysis showed that Korean research on smart tourism focused heavily on improving the quality and satisfaction of tourists' experiences and understanding their behavioral intentions, while Chinese research emphasized provision of immersive contents and studies on sustainable tourism. Based on the analysis results, implications for smart tourism research were presented, along with limitations of this study and future research directions.

• Cartoon and Animation Studies
• /
• s.17
• /
• pp.37-51
• /
• 2009

Immersion is the key factor in determining success or failure for fictional narrative animations. It does matter to not only for the socio-economic achievements like box-office success but to the asthetical achievement, immersion is very important prerequisite character that are raised in terms of research needs. Fictional narrative animation shares the audience's immersive method with the narrative content and Virtual Reality, since animation has the character which narrates itself with the visual image format (including sound) and added the fact that 3D format which emphasizes realism compared with the recent 2D animation has been ubiquitous phenomenon. In other words, the artistic for called 'animation' is located between the view point that text has the immersive point intrinsic and oppositely, specific function of the certain media which stays extrinsic of the text enforce the participants into the immersion. In any case, the subject who experience immersion is the audience, so the most useful theory to research the phenomenon 'immersion' is Constructivism cognitive narrotologic approach which lies the peader-spectator as the centric notion and is more useful than text analysis or research on the visual effects. In this study, research and review about the studies on the audience's immersive experience would be preceded particularly aroud the constructivism theories, and examine the uniqueness and naures of the animation which makes the audience immersive.

• (The)Study of the Eastern Classic
• /
• no.43
• /
• pp.305-328
• /
• 2011

This study examines a study of communication factor in the Chinese augury bone, as a communication text, a communication history text and a communication thought text, in terms of sender message channel audience effects, interdisciplinary research. In many respects, it is generally accepted that the Chinese augury bone have been the generic references of the Oriental culture. Based on various research results about The Chinese augury boney as a communication text, this paper explores factors of communication text related with meta-communication as the definition of communication and a structure of communication, a method of communication. The Chinese augury bone consider graphic element as communication tool, its manager as sender or communicator, the record as message, the bone as channel, its reader as audience, text understanding and behavior as effects, the augury bone producer as journalist, the tying bone in a bundle as old style book. In the Chinese augury bone, one property of communication thought is that the political ideology, which is related to the strong and weak of the graphic style and determines the naming method of kings.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.51 no.5
• /
• pp.44-56
• /
• 2023

This research explores the enhancement of park operation and management by analyzing the changing demands of park users. While traditional methods depended on surveys, there has been a recent shift towards utilizing social media data to understand park usage trends. Notably, most research has focused on text data from social media, overlooking the valuable insights from image data. Addressing this gap, our study introduces a novel method of assessing park usage using social media image data and then applies it to actual city park evaluations. A unique image analysis tool, built on Visual Question Answering (VQA) deep learning technology, was developed. This tool revealed specific city park details such as user demographics, behaviors, and locations. Our findings highlight three main points: (1) The VQA-based image analysis tool's validity was proven by matching its results with traditional text analysis outcomes. (2) VQA deep learning technology offers insights like gender, age, and usage time, which aren't accessible from text analysis alone. (3) Using VQA, we derived operational and management strategies for city parks. In conclusion, our VQA-based method offers significant methodological advancements for future park usage studies.

• Korean Journal of Ornithology
• /
• v.25 no.2
• /
• pp.126-132
• /
• 2018

For balanced development of ornithological research in Korea, it is important to review what birds and what research topics have been studied so far. We quantitatively investigated the trends of domestic ornithological research using text-mining and co-word analysis. As a result of studying 372 articles published in the Korean Journal of Ornithology, which is the most representative ornithological journals, words related to research topics such as population and community monitoring, first record of species and breeding ecology, and heavy metal pollution in birds have been widely used in research articles. Except for subjects such as monitoring and first record of species, studies have not been conducted widely. It was also found that research were concentrated on specific birds such as Anas platyrhynchos, Calidris alpina, and Anas poecilorhyncha. The present study, which analyzed the research topics and avian taxa that were relatively active until now and those which were insufficient, suggests what we should do in the future for the balanced development of ornithological research in Korea.

• Journal of the Korea Convergence Society
• /
• v.9 no.4
• /
• pp.15-22
• /
• 2018

The purpose of this study is to investigate the consumer perception of packaging using big data analysis. This study use text mining to extract meaningful words from text and semantic network analysis to analyze connectivity and propagation trends. Data were collected by dividing the 'packaging(Korean)' and 'packaging(English)'. This study visualized the word network structure of the two key words and classified them into four groups with similar meaning through CONCOR analysis. The group name was specified based on the words constituting the classified group. These groups are a major category of consumers' perception of packaging. Especially cosmetics and design have high frequency of words and high centrality. Therefore it can be expected that the packaging design is perceived as important in the cosmetics industry. This study predicts consumers' perception of packaging so it can be a basis for future research and industry development.

• Informatization Policy
• /
• v.23 no.4
• /
• pp.76-94
• /
• 2016

Recently in Korea, the importance of information security awareness has been receiving a growing attention. Attacks such as social engineering and ransomware are hard to be prevented because it cannot be solved by information security technology. Also, the profitability of information security industry has been decreasing for years. Therefore, many companies try to find a new growth-engine and an entry to the foreign market. The main purpose of this paper is to draw out some information security issues and to analyze them. Finally, this study identifies issues and suggests how to improve the situation in Korea. For this, topic modeling analysis has been used to find information security issues of each country. Moreover, the score of sentiment analysis has been used to compare them. The study is exploring and explaining what critical issues are and how to improve the situation based on the identified issues of the Korean information security industry. Also, this study is also demonstrating how text mining can be applied to the context of information security awareness. From a pragmatic perspective, the study has the implications for information security enterprises. This study is expected to provide a new and realistic method for analyzing domestic and foreign issues using the analysis of real data of the Twitter API.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.3
• /
• pp.586-599
• /
• 2017

This study attempts to understand the overall research trends in Vocational Studies using a text mining method, which is a means to analyze big data. The findings of the research show that Vocational Studies in Korea has been directly influenced by global economic crises, as evidenced by its exponential growth after the 1997 foreign exchange crisis that resulted in a bailout from the IMF. In addition, the topics of research have been shifting from such macro subjects as government policies and systems to such micro topics as individual career development. Moreover, the perspective of research is being moved from the socially vulnerable, including women and the disabled, to the economically marginalized, including retirees and the unemployed. As for the research targets, college students overwhelmingly outnumbered primary and secondary school students. However, few cases analyzed the clinical outcomes of career counseling or attempted to process job information and study the history of jobs. This research is limited in that it only analyzed journal abstracts. Nonetheless, it is meaningful because it used topic analysis, one of the text mining methods, to give a complete enumeration of all articles available for search, thereby crafting a framework of quantitative analysis methodology for Vocational Studies. It is also significant in that it is the first attempt to analyze themes in every stage of the development of Vocational Studies.