• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4345-4363
• /
• 2021

Deep Learning as a Service (DLaaS), utilizing the cloud-based deep neural network models to provide customer prediction services, has been widely deployed on mobile cloud computing (MCC). Such services raise privacy concerns since customers need to send private data to untrusted service providers. In this paper, we devote ourselves to building an efficient protocol to classify users' images using the convolutional neural network (CNN) model trained and held by the server, while keeping both parties' data secure. Most previous solutions commonly employ homomorphic encryption schemes based on Ring Learning with Errors (RLWE) hardness or two-party secure computation protocols to achieve it. However, they have limitations on large communication overheads and costs in MCC. To address this issue, we present LeHE4SCNN, a scalable privacy-preserving and communication-efficient framework for CNN-based DLaaS. Firstly, we design a novel low-expansion rate homomorphic encryption scheme with packing and unpacking methods (LeHE). It supports fast homomorphic operations such as vector-matrix multiplication and addition. Then we propose a secure prediction framework for CNN. It employs the LeHE scheme to compute linear layers while exploiting the data shuffling technique to perform non-linear operations. Finally, we implement and evaluate LeHE4SCNN with various CNN models on a real-world dataset. Experimental results demonstrate the effectiveness and superiority of the LeHE4SCNN framework in terms of response time, usage cost, and communication overhead compared to the state-of-the-art methods in the mobile cloud computing environment.

• Information Systems Review
• /
• v.24 no.3
• /
• pp.69-91
• /
• 2022

The video recommendation services help to save the user's information search time in the overflowing online information, and algorithms for more efficient and accurate recommendation are continuously developed. In particular, TikTok has the largest number of users in the short video industry due to its unique recommendation algorithms. In this study, by applying a privacy calculation model, the research tried to compare users' responses to each type of TikTok's recommendation service. Users are well aware of the privacy concerns and benefits of TikTok's recommendation service. Although there is a risk, it was found that users continue to use TikTok's recommendation service because the benefits are greater.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.4
• /
• 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.

• The Journal of Society for e-Business Studies
• /
• v.12 no.4
• /
• pp.37-51
• /
• 2007

As the highly improved Internet and information technology has led to the diversification of users' demands, personalization service attract lots of attention as a means to meet highly diversified demand of users. However, personalization service costs a lot. Also concerns over a possible violation of privacy have been raised since the service uses technology to find out the users' profiles. This research studies the advantages individuals acquire from personalization service and how privacy concern influences service acceptance. Research on related documents and information gathering from e-commerce sites derived six representative types of service. Questionnaires were utilized to research privacy concern according to services, service usefulness, and service acceptance. As expected, privacy concern has a negative relation to acceptance while service usefulness has a positive relation to it, thereby resulting in an offset between two variables. Moreover, they play a different role depending on what kinds of service or in formation should be provided. The results derived from this paper will help the e-commerce sites provide personalization service by collecting personal information while protecting users' privacy.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.19 no.6
• /
• pp.121-134
• /
• 2009

Consumer-centric marketing business is surely one of the most successful emerging business but it poses a threat to personal privacy. Between the service provider and the user there are many contrary issues to each other. The enterprise asserts that to abuse the privacy data which is anonymous there is not a problem. The individual only will not be able to willingly submit the problem which is latent. Web traffic analysis technology itself doesn't create issues, but this technology when used on data of personal nature might cause concerns. The most criticized ethical issue involving web traffic analysis is the invasion of privacy. So we need to inspect how many and what kind of personal informations being used and if there is any illegal treatment of personal information. In this paper, we inspect the operation of consumer-centric marketing tools such as web log analysis solutions and data gathering services with web browser toolbar. Also we inspect Microsoft explorer-based toolbar application which records and analyzes personal web browsing pattern through reverse engineering technology. Finally, this identified and explored security and privacy requirement issues to develop more reliable solutions. This study is very important for the balanced development with personal privacy protection and web traffic analysis industry.

• Journal of Intelligence and Information Systems
• /
• v.29 no.4
• /
• pp.165-184
• /
• 2023

Recently, Deep learning analysis of unstructured text data using language models, such as Google's BERT and OpenAI's GPT has shown remarkable results in various applications. Most language models are used to learn generalized linguistic information from pre-training data and then update their weights for downstream tasks through a fine-tuning process. However, some concerns have been raised that privacy may be violated in the process of using these language models, i.e., data privacy may be violated when data owner provides large amounts of data to the model owner to perform fine-tuning of the language model. Conversely, when the model owner discloses the entire model to the data owner, the structure and weights of the model are disclosed, which may violate the privacy of the model. The concept of offsite tuning has been recently proposed to perform fine-tuning of language models while protecting privacy in such situations. But the study has a limitation that it does not provide a concrete way to apply the proposed methodology to text classification models. In this study, we propose a concrete method to apply offsite tuning with an additional classifier to protect the privacy of the model and data when performing multi-classification fine-tuning on Korean documents. To evaluate the performance of the proposed methodology, we conducted experiments on about 200,000 Korean documents from five major fields, ICT, electrical, electronic, mechanical, and medical, provided by AIHub, and found that the proposed plug-in model outperforms the zero-shot model and the offsite model in terms of classification accuracy.

• Knowledge Management Research
• /
• v.25 no.2
• /
• pp.49-68
• /
• 2024

Increasing concerns regarding personal data privacy have complicated the acquisition of customer data through online marketing. This study investigates factors influencing customers' willingness to disclose information via a generative AI-based chatbot. Drawing on privacy calculus theory and regulatory focus theory, we explore how customer regulatory focus and familiarity with the generative AI-based chatbot shape disclosure intentions. Our study, involving 473 participants, reveals that low familiarity with the chatbot leads individuals with a prevention focus to perceive higher privacy risks and lower perceived usefulness compared to those with a promotion focus. However, with high familiarity, these differences diminish. Moreover, individuals with a promotion focus show a greater inclination to disclose information when familiarity with the generative AI-based chatbot is low, whereas this regulatory focus does not significantly impact disclosure intentions when familiarity is high. Perceived privacy risks mediate these relationships, underscoring the importance of understanding familiarity with the generative AI-based chatbot in facilitating personal information disclosure.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.8
• /
• pp.103-109
• /
• 2013

Mobile ad hoc networks (MANETs) are infrastructure-less and stand-alone wireless networks with dynamic topologies. To support user's safety in MANETs, fundamental and various security services should be supported. Especially in mobile commercial market, one of the major concerns regarding security is user privacy. Recently, researches about security system to protect user privacy in cluster-based MANETs have been introduced. This paper propose a cluster reconfiguration protocol under anonymous cluster-based MANETs to enhance the network stability. The improved anonymous cluster-based MANETs can recover the network structure against abnormal states of clutserheads.

• Journal of Fashion Business
• /
• v.9 no.3
• /
• pp.134-149
• /
• 2005

Although Internet retailing is becoming a viable channel for apparel retailing, consumers are still reluctant to use Internet for apparel purchasing because at their concerns at Internet security and the difficulties at virtual shopping in unfamiliar shopping environment. The purpose at this study is to examine the nature at perceived risk associated with Internet apparel shopping and risk-reduction strategies used by Internet apparel shoppers. The data were collected via an online survey by a online research company. A total at 4,254 Internet users participated in this survey. Among these Internet users, 1,146 respondents had previous shopping experience in Internet shopping. Within this group, 195 were Internet apparel information seekers, and 589 were Internet apparel purchasers. Descriptive statistics, analysis of variance, and t-test were used to analyze the data. The perceived risks and risk-reduction strategies used by Internet apparel no-interest shoppers, Internet apparel information seekers (browsers), moderate Internet apparel purchasers, heavy Internet apparel purchasers were examined and compared. The results indicated that these tour groups were significantly different in apparel related risk, performance risk, and privacy risk. Internet purchasers tend to perceive more apparel-related, performance, and privacy risks than others. The results also indicated that these tour groups were significantly different in their opinions of risk-reduction strategies.

• Informatization Policy
• /
• v.23 no.4
• /
• pp.38-58
• /
• 2016

The ever-evolving Internet environment along with changes in the mass media has been creating a new way of communicating in the virtual cyber world. The Internet users have more services at their disposal to communicate with ease. Such a new way of communication styles, however, makes them vulnerable to personal information leakage, increasing the concerns of cyber security. A thorny issue is how we can control the disclosure of personal information. Lately, the Korean government implemented privacy policies to resolve and prevent personal information leakage incidents that incur social problems. Here, we seek to identify problems in the privacy policies for better solutions.