• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2022년도 제65차 동계학술대회논문집 30권1호
• /
• pp.411-412
• /
• 2022

본 논문에서는 코로나19 뉴스와 코로나19 가짜뉴스의 데이터셋을 활용하여 입력 받은 뉴스가 가짜뉴스일 확률을 예측한다. 가짜 뉴스 본문에는 코로나19, 대통령, 정부, 가짜, 언론 등의 키워드의 빈도가 높았다. 위의 키워드를 토대로 나이브 베이즈 모델링을 하여 이를 적용해 가짜 뉴스를 가려내는 웹페이지를 개발하였다.

• 한국융합학회논문지
• /
• 제10권7호
• /
• pp.15-21
• /
• 2019

가짜 정보를 탐지하기 위한 연구는 2016년 미국 대통령 선거 이후 본격적으로 시작되었다. 정확한 출처를 알 수 없는 정보들이 뉴스 형식으로 생산되고, 이는 자극적이고 흥미로운 소재에 많은 관심을 보이는 대중의 특성에 따라 빠른 속도로 확산되고 있다. 또한, 소셜 네트워크 서비스 등 정보를 전달하기 쉬운 플랫폼의 대중화는 이러한 현상을 더욱 악화시킨다. Poynter는 IFCN(International Fact Checking Network)를 만들어 숙련된 전문가들이 사실 여부를 판단할 수 있는 가이드라인을 제시하고, 팩트 체크 기관을 위한 강령을 제공하고 있다. 하지만 이러한 접근 방법은 하나의 기사에 대한 진위 여부를 검증하기 위해 다수의 전문가 인력이 투입되어야 하므로 시간 및 금전적 비용이 크다. 따라서 지속적으로 증가하는 가짜 뉴스에 효율적으로 대응할 수 있는 자동화된 가짜 뉴스 탐지 기술에 대한 연구가 주목받고 있다. 본 논문에서는 최근 딥러닝 기술의 접목으로 인해 빠르게 발전하고 있는 가짜 뉴스 탐지 시스템과 연구들을 정리 및 분석한다. 또한, 많은 연구가 필요한 본 분야에 연구자들이 쉽게 접근할 수 있도록 다양한 형태로 주어지는 학습 말뭉치 및 챌린지들도 정리한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권1호
• /
• pp.51-73
• /
• 2023

Social media play a significant role in communicating information across the globe, connecting with loved ones, getting the news, communicating ideas, etc. However, a group of people uses social media to spread fake information, which has a bad impact on society. Therefore, minimizing fake news and its detection are the two primary challenges that need to be addressed. This paper presents a multi-modal deep learning technique to address the above challenges. The proposed modal can use and process visual and textual features. Therefore, it has the ability to detect fake information from visual and textual data. We used EfficientNetB0 and a sentence transformer, respectively, for detecting counterfeit images and for textural learning. Feature embedding is performed at individual channels, whilst fusion is done at the last classification layer. The late fusion is applied intentionally to mitigate the noisy data that are generated by multi-modalities. Extensive experiments are conducted, and performance is evaluated against state-of-the-art methods. Three real-world benchmark datasets, such as MediaEval (Twitter), Weibo, and Fakeddit, are used for experimentation. Result reveals that the proposed modal outperformed the state-of-the-art methods and achieved an accuracy of 86.48%, 82.50%, and 88.80%, respectively, for MediaEval (Twitter), Weibo, and Fakeddit datasets.

• Journal of Contemporary Eastern Asia
• /
• 제18권2호
• /
• pp.33-48
• /
• 2019

The current Moon Jai-in administration in South Korea is facing serious challenges as a result of a scandal involving the manipulation of news online. Staff in Moon's camp are suspected of manipulating public opinion by creating millions of fake news comments online, contributing to Moon being elected president. This South Korean political scandal raises a number of theoretical issues with regard to new platform technologies and media manipulation. First, the incident exposes the technological limits of blocking manipulation of the news, partly because of the nature of social media and partly because of the nature of contemporary technology. Contemporary social media is often monopolistic in nature; with the majority of people are using the same platforms, and hence it is likely that they will be subject to forms of media manipulation. Second, the Korean case of news manipulation demonstrates a unique cultural aspect of Korean society. News comments and readers' replies have become a major channel of alternative news in Korea. This phenomenon is often designated as "reply journalism," since people are interested in reading the news replies of ordinary readers equally to reading news reports themselves. News replies are considered indicators of public opinion and are seen as affecting trias politica in Korean society. Third, the Korean incident of news manipulation implicates a new form of populism in the 21st century and the nature of democratic participation. This article aims to explicate key issues in media manipulation by including wider technological, cultural, and political aspects in the South Korean news media context.

• International journal of advanced smart convergence
• /
• 제10권3호
• /
• pp.131-142
• /
• 2021

Recently, there has been an active service that provides customized news to news subscribers. In this study, we intend to design a customized news service system through Deep Learning-based Social Network Service (SNS) activity analysis, applying real news and avoiding fake news. In other words, the core of this study is the study of delivery methods and delivery devices to provide customized news services based on analysis of users, SNS activities. First of all, this research method consists of a total of five steps. In the first stage, social network service site access records are received from user terminals, and in the second stage, SNS sites are searched based on SNS site access records received to obtain user profile information and user SNS activity information. In step 3, the user's propensity is analyzed based on user profile information and SNS activity information, and in step 4, user-tailored news is selected through news search based on user propensity analysis results. Finally, in step 5, custom news is sent to the user terminal. This study will be of great help to news service providers to increase the number of news subscribers.

• International Journal of Internet, Broadcasting and Communication
• /
• 제11권2호
• /
• pp.59-66
• /
• 2019

The emergence of new IT technologies and convergence industries, such as artificial intelligence, bigdata and the Internet of Things, is another chance for South Korea, which has established itself as one of the world's top IT powerhouses. On the other hand, however, privacy concerns that may arise in the process of using such technologies raise the task of harmonizing the development of new industries and the protection of personal information at the same time. In response, the government clearly presented the criteria for deidentifiable measures of personal information and the scope of use of deidentifiable information needed to ensure that bigdata can be safely utilized within the framework of the current Personal Information Protection Act. It strives to promote corporate investment and industrial development by removing them and to ensure that the protection of the people's personal information and human rights is not neglected. This study discusses the strategy of deidentifying personal information protection based on the analysis of fake news. Using the strategies derived from this study, it is assumed that deidentification information that is appropriate for deidentification measures is not personal information and can therefore be used for analysis of big data. By doing so, deidentification information can be safely utilized and managed through administrative and technical safeguards to prevent re-identification, considering the possibility of re-identification due to technology development and data growth.

• Journal of Contemporary Eastern Asia
• /
• 제18권2호
• /
• pp.68-89
• /
• 2019

When comparing media freedom in Malaysia, Myanmar, and Thailand, so-called "fake news" appears as threats to a deliberative (online) public sphere in these three diverse contexts. However, "racist propaganda", "information operations" and "negative campaigning" might be more accurate terms that explain these forms of systematic manipulative political communication. The three cases show forms of disinformation in under-researched contexts and thereby expand the often Western focused discourses on hate speech and fake news. Additionally, the analysis shows that harmful disinformation disseminated online originates from differing contextual trajectories and is not an "online phenomenon". Drawing on an analysis of connotative context factors, this explorative comparative study enables an understanding of different forms of harmful disinformation in Malaysia, Myanmar, and Thailand. The connotative context factors were inductively inferred from 32 expert interviews providing explanations for the formation of political communication (control) mechanisms.

• Journal of Information Processing Systems
• /
• 제15권5호
• /
• pp.1029-1035
• /
• 2019

In recent years, artificial intelligence (AI) services have become one of the most essential parts to extend human capabilities in various fields such as face recognition for security, weather prediction, and so on. Various learning algorithms for existing AI services are utilized, such as classification, regression, and deep learning, to increase accuracy and efficiency for humans. Nonetheless, these services face many challenges such as fake news spread on social media, stock selection, and volatility delay in stock prediction systems and inaccurate movie-based recommendation systems. In this paper, various algorithms are presented to mitigate these issues in different systems and services. Convolutional neural network algorithms are used for detecting fake news in Korean language with a Word-Embedded model. It is based on k-clique and data mining and increased accuracy in personalized recommendation-based services stock selection and volatility delay in stock prediction. Other algorithms like multi-level fusion processing address problems of lack of real-time database.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2020년도 제62차 하계학술대회논문집 28권2호
• /
• pp.199-202
• /
• 2020

본 논문에서는 가짜뉴스 탐지 모델에 워드 임베딩 기법을 접목하여 성능을 향상시키는 방법을 제안한다. 기존의 한국어 가짜뉴스 탐지 연구는 희소 표현인 빈도-역문서 빈도(TF-IDF)를 활용한 탐지 모델들이 주를 이루었다. 하지만 이는 가짜뉴스 탐지의 관점에서 뉴스의 언어적 특성을 파악하는 데 한계가 존재하는데, 특히 문맥에서 드러나는 언어적 특성을 구조적으로 반영하지 못한다. 이에 밀집 표현 기반의 워드 임베딩 기법인 Word2vec을 활용한 텍스트 전처리를 통해 문맥 정보까지 반영한 가짜뉴스 탐지 모델을 본 연구의 제안 모델로 생성한 후 TF-IDF 기반의 가짜뉴스 탐지 모델을 비교 모델로 생성하여 두 모델 간의 비교를 통한 성능 검증을 수행하였다. 그 결과 Word2vec 기반의 제안모형이 더욱 우수하였음을 확인하였다.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2020년도 제61차 동계학술대회논문집 28권1호
• /
• pp.37-39
• /
• 2020

인터넷 상에서의 허위정보 생산과 유통은 주로 가짜 뉴스를 통하여 이루어진다. 과거에는 신문이나 공중파 TV등 뉴스 기사의 생산과 유통이 매우 제한적이었지만 지금은 인터넷의 발달로 누구나 쉽게 뉴스를 생산하고 유통할 수 있다. 뉴스 생산의 용이성은 정보 공유의 즉각성과 수월성이라는 장점을 제공하지만 반대로 불확실한 뉴스 남발로 인한 정보의 신뢰성 하락과 선량한 피해자를 양산하는 단점 또한 존재한다. 이는 가짜 뉴스가 사회적 문제로 대두되고 있는 이유이다. 에이전트나 스파이더 등의 소프트웨어를 통해 인터넷으로 급속도로 전파되는 가짜 뉴스를 전통 방식인 소수의 전문가가 수동으로 잡아내는 것은 불가능하다. 이에 기술발달로 잡아내기 힘들어진 가짜뉴스에 대해, 역으로 발달된 기술을 활용하여 잡아내려는 시도가 늘어나고 있다. 본 논문에서는 가짜뉴스를 판별하는 다양한 기법들을 탐색하고 해결방안을 제시하고자 한다.