• Journal of Information Technology Applications and Management
• /
• 제15권4호
• /
• pp.37-59
• /
• 2008

This study focuses on the antecedents to the privacy concerns and their influence on trust and online transaction intention. Based on previous exploratory works and the literature review of privacy concerns, four antecedents are identified-Internet literacy, social awareness, perceived vulnerability, and perceived ability to information control. Incorporating these antecedents, privacy concerns, trust and online transaction intention, a conceptual model is developed and seven research hypotheses are proposed for empirical testing. The proposed model is examined through structural equation analysis. The results show that Internet literacy, social awareness, and perceived vulnerability have statistically significant effect on the privacy concerns of users and the privacy concerns has a positive influence on the trust. Finally, the trust has a positive effect on the online transaction intention. Implications of these findings are discussed for both researchers and practitioners and future research issues are raised as well.

• 한국정보시스템학회지:정보시스템연구
• /
• 제25권4호
• /
• pp.1-21
• /
• 2016

Purpose The study analyzed the factors affecting continuance use and self-disclosure in the SNS(social networking service) context based on a dual-factor model. As SNS users have concerned privacy for a long time, privacy concern affects continuous use and self-disclosure. In details, concern over privacy may have a stronger effect on self-disclosure than on continuance use as users' personal information can be more exposed during posting their dailies and photos. Design/Methodology/Approach SNS benefits, trust in SNS providers, and social influence are served as the key enablers and privacy concern as the inhibitor. Moreover, the relative impacts of SNS benefits and privacy concern on continuance use and self-disclosure were analysed in this study. From the data of 327 Facebook users, the researchers tested proposed theoretical model by using PLS. Findings Users' continuance intention and self-disclosure behavior are differently affected by different antecedents. Trust in SNS provider had a significant effect on self-disclosure intention, while it has no significant effect on continuance intention. Concern over privacy was negatively related to self-disclosure intention, while it was positively associated with continuance intention.

• 한국정보과학회논문지:데이타베이스
• /
• 제37권4호
• /
• pp.209-219
• /
• 2010

최근 빠르게 확산되고 있는 온라인 소셜 네트워크 서비스는 수많은 데이터를 저장하고 이를 분석하여 여러 연구 분야에 활용하고 있다. 정보의 효율성을 높이기 위해 기업이나 공공기관은 자신들이 가진 데이터를 배포하고, 배포된 데이터를 이용하여 여러 목적에 사용한다. 그러나 배포되는 소셜 네트워크에는 개인과 관련된 정보가 포함되어 있으므로 개인 프라이버시가 노출될 수 있는 문제가 있다. 배포되는 소셜 네트워크에서 단순히 이름 등의 식별자를 지우는 것으로는 개인 프라이버시 보호에 충분하지 않으며, 소셜 네트워크가 가진 구조적 정보에 의해서도 개인 프라이버시가 노출될 수 있다. 본 논문에서는 내용 정보를 포함하고 있는 소셜 네트워크 배포 시 개인 프라이버시 노출에 이용되는 복합된 공격법을 제시하고 이를 방지할 수 있는 새로운 모델인 $\ell$-차수 다양성($\ell$-degree diversity)을 제안한다. $\ell$-차수 다양성은 소셜 네트워크 데이터 배포에서 $\ell$-다양성을 최초로 적용한 모델이며 높은 정보 보존율을 가짐을 실험을 통해 볼 수 있다.

• 디지털융복합연구
• /
• 제11권1호
• /
• pp.237-242
• /
• 2013

최근 페이스북(Facebook)과 트위터(Twitter) 등의 폭발적인 성장에 따라 SNS는 사회적 학문적인 관심의 대상으로 부상하고 있다. 그러나, SNS는 이용자의 신상 정보와 사적인 의견 교환을 근간으로 사용자의 프라이버시가 노출될 수 있는 문제가 존재한다. 본 논문에서는 현재 SNS에서 이용자의 개인 프라이버시를 보호하기 위해 사용되고 있는 블록킹 대신 데이터 분리와 데이터 허위 정보를 이용한 SNS 사용자 프라이버시 보호 모델을 제안한다. 제안 모델은 사용자의 내용 정보를 분리하여 분리된 내용 정보에 허위 정보를 추가함으로써 제3자가 사용자의 내용 정보를 수집하여도 정확한 정보를 추출하지 못하도록 하고 있다. 또한, 제3자가 이용자의 정보를 불법적으로 악용하지 않도록 SNS 서비스 제공자가 이용자의 정보를 활용할 경우 이용자에게 사전에 동의를 구한다.

• 정보보호학회논문지
• /
• 제20권4호
• /
• pp.135-144
• /
• 2010

프라이버시 보호는 기업의 온/오프-라인 데이터 처리 시스템 안에서 프라이버시 정책들을 수행할 수 있을 때에 달성될 수 있다. 프라이버시 정책 모델 중에는 P-RBAC과 목적모델, 의무모델이 있다. 그러나 이들 각각의 모델들만으로는 급변하는 기업환경에 능동적으로 대처하기 어렵다. 동일한 역할에 속해있는 사용자 중 최적의 조건을 만족하는 자만을 선발하여 일정기간 새로운 임무를 부여할 수 있어야하고, 풍부한 접근제약조건 표현을 허용하여 프라이버시 보호를 강화할 수 있어야 한다. 이를 위해 우리는 목적모델과 P-RBAC 모델, 의무모델을 통합시킨 통합형 프라이버시 보호 모델을 제안한다. 그리고 우리의 모델이 구현플랫폼과 응용에 종속적이지 않고 자동화될 수 있도록 XML 기반 정책언어모델을 정의한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제18권4호
• /
• pp.826-842
• /
• 2024

As 5G and AI continue to develop, there has been a significant surge in the healthcare industry. The COVID-19 pandemic has posed immense challenges to the global health system. This study proposes an FL-supported edge computing model based on federated learning (FL) for predicting clinical outcomes of COVID-19 patients during hospitalization. The model aims to address the challenges posed by the pandemic, such as the need for sophisticated predictive models, privacy concerns, and the non-IID nature of COVID-19 data. The model utilizes the FATE framework, known for its privacy-preserving technologies, to enhance predictive precision while ensuring data privacy and effectively managing data heterogeneity. The model's ability to generalize across diverse datasets and its adaptability in real-world clinical settings are highlighted by the use of SHAP values, which streamline the training process by identifying influential features, thus reducing computational overhead without compromising predictive precision. The study demonstrates that the proposed model achieves comparable precision to specific machine learning models when dataset sizes are identical and surpasses traditional models when larger training data volumes are employed. The model's performance is further improved when trained on datasets from diverse nodes, leading to superior generalization and overall performance, especially in scenarios with insufficient node features. The integration of FL with edge computing contributes significantly to the reliable prediction of COVID-19 patient outcomes with greater privacy. The research contributes to healthcare technology by providing a practical solution for early intervention and personalized treatment plans, leading to improved patient outcomes and efficient resource allocation during public health crises.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2024년도 춘계학술발표대회
• /
• pp.711-714
• /
• 2024

In the dynamic landscape of modern machine learning, Federated Learning (FL) has emerged as a compelling paradigm designed to enhance privacy by enabling participants to collaboratively train models without sharing their private data. Specifically, Distillation-based Federated Learning, like Federated Learning with Model Distillation (FedMD), Federated Gradient Encryption and Model Sharing (FedGEMS), and Differentially Secure Federated Learning (DS-FL), has arisen as a novel approach aimed at addressing Non-IID data challenges by leveraging Federated Learning. These methods refine the standard FL framework by distilling insights from public dataset predictions, securing data transmissions through gradient encryption, and applying differential privacy to mask individual contributions. Despite these innovations, our survey identifies persistent vulnerabilities, particularly concerning the susceptibility to logit inversion attacks where malicious actors could reconstruct private data from shared public predictions. This exploration reveals that even advanced Distillation-based Federated Learning systems harbor significant privacy risks, challenging the prevailing assumptions about their security and underscoring the need for continued advancements in secure Federated Learning methodologies.

• 지식경영연구
• /
• 제13권4호
• /
• pp.83-100
• /
• 2012

Given the prevalence of mobile social network services (SNS) such as Facebook and Kakaotalk, it has become important to understand user's continuance behavior in a mobile SNS environment. Although trust and privacy concerns play a key role in SNS users' decision-making processes, most studies on SNS have shed little light on the effects of trust and privacy concerns on SNS continuance intention. In this regard, this paper developed an integrated model to deeply understand the key antecedents of user's continuance intention to use mobile SNS by incorporating trust and privacy concerns into extended expectation-confirmation model. The proposed research model was tested by using survey data collected from 170 users who have experience with Kakaotalk. The findings of this study found that the proposed theoretical framework provides a statistically significant explanation of the variance in continuance intention of mobile SNS. The analysis results indicate that trust serves as the salient antecedent of continuance intention to use mobile SNS. However, it was found that privacy concerns negatively influence trust, whereas it is not significantly related to continuance intention of mobile SNS. The theoretical and practical implications of the findings were described.

• 융합보안논문지
• /
• 제12권1호
• /
• pp.71-77
• /
• 2012

정보통신망의 발전과 함께 인터넷 사용 인구와 다양한 개방적 구조의 서비스 이용률이 지속적으로 증가하고 있다. 하지만 서비스 이용자들의 보안의식은 크게 달라지지 않아 서비스 이용자들의 직접적인 입력으로 인터넷상에 노출되는 개인정보가 늘어나고 있는 실정이며 이로 인한 이차적인 침해로 인하여 개인에게 정신적인 피해와 금전적 손괴 심지어는 신체적인 위험을 주는 각종 범죄가 행해지고 있다. 서비스 이용자의 직접적인 입력으로 인한 개인정보 노출을 예방하기 위해서는 게시물을 등록하는 서비스 이용자에게 게시물에 포함된 개인정보와 개인정보 노출의 위험성을 인지시켜 이차적 침해의 위험성이 높은 개인정보는 게시하지 않도록 해야 한다. 본 논문에서는 게시물의 노출 수준과 개인정보자산 가치를 이용하여 각 개인정보의 위험수준을 산정하는 모델과 게시판과 같은 서비스에서 개인정보노출정도에 따른 위험도를 사용하는 방법을 보였다.

• 지식경영연구
• /
• 제19권1호
• /
• pp.19-39
• /
• 2018

With the development of information technology, social network services (SNS) such as Facebook and Twitter became popular and many users disclose their personal and sensitive information like private story, photographs and location information through posting and sharing. Despite the privacy concerns in SNSs, individuals continue to disclose their identity online. This phenomenon is called 'privacy paradox'. The purpose of this study is to examine the role of collective efficacy on self-disclosure in SNS context and to explain privacy paradox phenomenon. Drawing upon the communication privacy management theory, research model was developed and empirically tested with cross-sectional data from 306 individuals. Results revealed that collective efficacy has a direct positive effect on self-disclosure while privacy risk is negatively related to self-disclosure. However, privacy concern is not directly related to self-disclosure. The relationship between privacy concern and self-disclosure was moderated by collective efficacy.