Privacy Protection Research Articles

An efficient method for privacy protection in big data analytics using oppositional fruit fly algorithm

This work employs anonymization techniques to safeguard privacy. Data plays a vital role in corporate decision-making in the current information-centric landscape. Various sectors, like banking and healthcare, gather confidential information on a daily basis. This information is disseminated by multiple sources through numerous methods. Securing sensitive data is of paramount importance for any data mining application. This study safeguarded confidential information using an anonymization technique. Several machine learning methodologies have a deficiency in accuracy. The study seeks to generate superior and more precise results compared to alternative methodologies. For large datasets, numerous solutions exhibit increased time complexity and memory use. For huge datasets, numerous solutions require more time and memory. The enhanced fuzzy C-means (FCM) algorithm surpasses existing approaches in terms of both accuracy and information preservation. This study provides a comprehensive analysis of data anonymization utilizing the oppositional fruit fly approach, a technique that enhances privacy. The clustering method being presented utilizes an enhanced version of the FCM algorithm. The secrecy of the recommended oppositional fruit fly algorithm is effective. The comparison demonstrated that the proposed research enhanced both accuracy and privacy in comparison to two existing methods. The existing strategy outperforms data anonymization-based privacy preservation by 82.17%, while the suggested method surpasses it by 94.17%.

Federated learning meets Bayesian neural network: Robust and uncertainty-aware distributed variational inference.

Federated Learning (FL) is a popular framework for data privacy protection in distributed machine learning. However, current FL faces some several problems and challenges, including the limited amount of client data and data heterogeneity. These lead to models trained on clients prone to drifting and overfitting, such that we just obtain suboptimal performance of the aggregated model. To tackle the aforementioned problems, we introduce a novel approach explicitly integrating Bayesian neural networks (BNNs) into the FL framework. The proposed approach is able to enhance the robustness. We refer to this approach as FedUAB, standing for FL with uncertainty-aware BNNs. In the FedUAB algorithm, each FL client independently trains a BNN using the Bayes by backprop algorithm. The weights of approximating model are modeled as Gaussian distributions, which mitigates the overfitting issue and also ensures better data privacy. Besides, we apply novel methods to overcome other key challenges in the fusion of BNNs and FL, such as selecting an optimal prior distribution, aggregating weights characterized by Gaussian forms across multiple clients, and rigorously managing weights variances. In the simulation of a FL environment, FedUAB demonstrated superior performance with both its server-side global model and client-side personalized models, outperforming traditional FL and other Bayesian FL methods. Moreover, it possesses the capability to quantify and decompose uncertainties. We have open-sourced our project at https://github.com/lpf111222/FedUAB/.

What are the Thoughts of Women Whose Husbands' Frozen Testicular Sperm Is Thawed for in vitro Fertilization on the Day of Oocyte Retrieval? A Qualitative Study.

The psychological experiences will be analyzed to understand the needs and burdens of women on the day of oocyte retrieval when the thawed testicular sperm of their husbands is used for in vitro fertilization, in order to provide a basis for the subsequent formulation of relevant nursing measures. This study utilized a descriptive phenomenological research approach. A cohort of 13 women undergoing oocyte retrieval on the day when thawed testicular sperm from their husbands is used for in vitro fertilization at the Reproductive Medicine Center of Sir Run Run Shaw Hospital, Zhejiang University School of Medicine, between August and October 2024, were chosen as participants for this study. Semi-structured, in-depth interviews were conducted to gather qualitative data. The Colaizzi 7-step analysis method was subsequently utilized to analyze the interview transcripts and identify emerging themes. The analysis yielded 3 themes: (1) facing challenges and requiring assistance (a. demonstrate a strong desire to acquire professional knowledge; b. demonstrate a strong desire for recognition); (2) overwhelming psychological stress (a. concerns about potential privacy breaches; b. concerns about security issues; c. concerns about the therapeutic efficacy); (3) improving family atmosphere and marital relationships (a. communication between couples has increased.; b. shared expectations for children help couples establish a common goal). It is recommended that healthcare practitioners implement strategies to enhance patient education and awareness surrounding assisted reproductive technology. Additionally, establishing and strengthening social support, reinforcing privacy protections and safety measures, promoting effective communication between partners, and managing treatment outcome expectations are crucial steps forward. Women prioritize treatment efficacy, privacy preservation, and the necessity of acquiring comprehensive professional knowledge when employing their partner's testicular sperm for in vitro fertilization. The study is instrumental in shaping policy dialogues on reproductive health, highlighting the critical significance of women's psychological health in the realm of in vitro fertilization.

Analyzing the TotalSegmentator for facial feature removal in head CT scans.

Facial recognition technology in medical imaging, particularly with head scans, poses privacy risks due to identifiable facial features. This study evaluates the use of facial recognition software in identifying facial features from head CT scans and explores a defacing pipeline using TotalSegmentator to reduce re-identification risks while preserving data integrity for research. 1404 high-quality renderings from the UCLH EIT Stroke dataset, both with and without defacing were analysed. The performance of defacing with the face mask created by TotalSegmentator was compared to a state-of-the-art CT defacing algorithm. Face detection was performed using deep learning models. The cosine similarity between facial embeddings for intra- and inter-patient images was compared. A Support Vector Machine was trained on cosine similarity values to assess defacing performance, determining if two renderings came from the same patient. This analysis was conducted on defaced and non-defaced images using 5-fold cross-validation. Faces were detected in 76.5% of non-defaced images. Intra-patient images exhibited a median cosine similarity of 0.65 (IQR: 0.47-0.80), compared to 0.50 (IQR: 0.39-0.62) for inter-patient images. A binary classifier performed moderately on non-defaced images, achieving a ROC-AUC of 0.69 (SD=0.01) and an accuracy of 0.65 (SD=0.01) in distinguishing whether a scan belonged to the same or a different individual. Following defacing, performance declined markedly. Defacing with the TotalSegmentator decreased the ROC-AUC to 0.55 (SD=0.02) and the accuracy to 0.56 (SD=0.01), whereas the CTA-DEFACE algorithm brought the performance down to a ROC-AUC of 0.60 (SD=0.02) and an accuracy of 0.59 (SD=0.01). These results demonstrate the effectiveness of defacing algorithms in mitigating re-identification risks, with the TotalSegmentator providing slightly superior privacy protection. Facial recognition software can identify facial features from partial and complete head CT scan renderings. However, using the TotalSegmentator to deface images reduces re-identification risks to a near-chance level. We offer code to implement this privacy-preserving pipeline. Utilizing the TotalSegmentator framework, the proposed pipeline efficiently removes facial features from CT images, making it ideal for multi-site research and data sharing. It is a useful tool for radiographers and radiologists who must comply with medico-legal requirements necessitating the removal of facial features.

Data Privacy Protection Diagnostic Algorithm for Industrial Robot Joint Harmonic Reducers Based on Swarm Learning

Data Privacy Protection Diagnostic Algorithm for Industrial Robot Joint Harmonic Reducers Based on Swarm Learning

IPA기법을 이용한 예비교사의 에듀테크 역량에 대한 인식 연구

This study aimed to identify and propose key areas and factors of EdTech Competencies required for teachers through a literature review and analysis. It also investigated prospective teachers’ perceptions of each factor, providing statistical insights and implications. The results are as follows. First, the literature review and analysis identified 7 areas and 22 factors for EdTech Competencies. Among these, a survey was conducted with 19 questions covering 6 areas that prospective teachers found appropriate. Second, prospective teachers recognized skills in digital citizenship instruction, digital equipment usage, digital copyright and intellectual property rights compliance, data management skills, and learning and staying updated on the latest EdTech trends as important factors. Third, prospective teachers reported that among various factors, data security and privacy protection, software and hardware problem-solving skills, AI-based learning support skills, and creative content production skills were relatively lacking. Fourth, according to the IPA results, the main EdTech Competencies required by prospective teachers include the ability to solve software or hardware problems, data management skills, staying updated on the latest EdTech trends, peer learning and community participation, creative content creation skills, and AI-based learning support. Fifth, effective utilization of EdTech requires guidelines on balancing digital and non-digital classes while managing students' digital dependency. This study offers useful insights into designing and developing educational programs focused on EdTech competency diagnostics and enhancement.

특수교육 분야의 자문화기술지 국내 연구 동향 분석

Objectives The purpose of this study is to examine the trends in autoethnographic research in special education, provide necessary information to researchers who intend to conduct autoethnographic research in the future, and explore the directions and applications of autoethnographic research in special education. Methods 54 autoethnography studies in the field of special education in Korea from 2000 to February 2024 were analyzed. The analysis criteria were established based on previous studies, and trends were analyzed according to ‘research type and publication year, researcher, research topic, data collection method, data analysis method, writing method, research validity, and research ethics’. Results As a result of the analysis, The most common type of research was journal articles, with 40 out of 54 (74.1%), followed by master's theses with 11(20.4%) and doctoral dissertations with 4(5.5%). The number of publications increased from 1(1.8%) in 2011 to 12(16.6%) after 2023, indicating a growing trend in autoethnography research. The most researcher type was special education teachers, accounting for 42(63.6%). The most research topic was instruction for students with disabilities, appearing in 23(42.6%) studies. Self-recall materials and cultural artifacts were primarily used for data collection. Holistic-category analysis, in vivo coding, emotional coding were the main data analysis methods. The most common writing method was confessional-emotional writing(35 studies, 33.7%), with imaginative-creative writing appearing in 2(2.9%) studies. Peer check was the most commonly used method for research validity(38 studies, 25.2%), with more than two analysis criteria employed. In terms of research ethics, privacy protection was the most common consideration(32 studies, 42.7%), while 16 studies(21.3%) did not mention research ethics. Conclusions Based on these results, implications for the direction and application of autoethnography research in the field of special education in Korea were discussed.

FSPPCFs: a privacy-preserving collaborative filtering recommendation scheme based on fuzzy C-means and Shapley value

Collaborative filtering recommendation systems generate personalized recommendation results by analyzing and collaboratively processing a large numerous of user ratings or behavior data. The widespread use of recommendation systems in daily decision-making also brings potential risks of privacy leakage. Recent literature predominantly employs differential privacy to achieve privacy protection, however, many schemes struggle to balance user privacy and recommendation performance effectively. In this work, we present a practical privacy-preserving scheme for user-based collaborative filtering recommendation that utilizes fuzzy C-means clustering and Shapley value, FSPPCFs, aiming to enhance the recommendation performance while ensuring privacy protection. Specifically, (i) we have modified the traditional recommendation scheme by introducing a similarity balance factor integrated into the Pearson similarity algorithm, enhancing recommendation system performance; (ii) FSPPCFs first clusters the dataset through fuzzy C-means clustering and Shapley value, grouping users with similar interests and attributes into the same cluster, thereby providing more accurate data support for recommendations. Then, differential privacy is used to achieve the user’s personal privacy protection when selecting the neighbor set from the target cluster. Finally, it is theoretically proved that our scheme satisfies differential privacy. Experimental results illustrate that our scheme significantly outperforms existing methods.

GDPR vs. Weakly Protected Parties in Other Countries

The General Data Protection Regulation (GDPR) establishes strong privacy rights for EU residents. However, many countries, including the United States, Asia, and Africa, lack similar protections. This creates challenges for multinational corporations and individuals navigating fragmented global data protection frameworks. This study examines the evolution, principles, and extraterritorial effects of GDPR, comparing it with weaker regulatory frameworks globally. Using qualitative and doctrinal research and document analysis, it identifies enforcement gaps, compliance costs, and operational difficulties in cross-border data transfers. Findings emphasize the need for consistent global standards to address the digital divide and ensure fair data practices. Recommendations highlight multilateral cooperation to develop an interoperable and ethical digital ecosystem. While GDPR offers a foundational model, achieving global harmonization requires collective international commitment. The study concludes with actionable suggestions to strengthen global privacy frameworks and protect vulnerable parties in weakly regulated regions.

A Secure Data Sharing Model Utilizing Attribute-Based Signcryption in Blockchain Technology.

With the rapid development of the Internet of Things (IoT), the scope of personal data sharing has significantly increased, enhancing convenience in daily life and optimizing resource management. However, this also poses challenges related to data privacy breaches and holdership threats. Typically, blockchain technology and cloud storage provide effective solutions. Nevertheless, the centralized storage architecture of traditional cloud servers is susceptible to single points of failure, potentially leading to system outages. To achieve secure data sharing, access control, and verification auditing, we propose a data security sharing scheme based on blockchain technology and attribute-based encryption, applied within the InterPlanetary File System (IPFS). This scheme employs multi-agent systems and attribute-based signcryption algorithms to process data, thereby enhancing privacy protection and verifying data holdership. The encrypted data are then stored in the distributed IPFS, with the returned hash values and access control policies uploaded to smart contracts, facilitating automated fine-grained access control services. Finally, blockchain data auditing is performed to ensure data integrity and accuracy. The results indicate that this scheme is practical and effective compared to existing solutions.

The Revolution of Blockchain in Digital Payment Systems: Legal Implications and Regulatory Challenges

This study aims to analyze the revolution of blockchain technology in digital payment systems, as well as explore the legal implications and regulatory challenges faced in its adoption. The research method used is a qualitative approach with normative analysis, utilizing data sources from literature, applicable regulations, and case studies from various countries. The results of the study show that blockchain has changed the digital payment landscape by providing solutions to improve efficiency, security, and transparency, this technology also poses legal challenges related to privacy, data security, and consumer protection. Differences in regulatory approaches between countries, especially between developed and developing countries, result in legal uncertainty that can hinder further adoption. These findings emphasize the importance of developing a uniform and adaptive regulatory framework to support blockchain innovation while maintaining stability and protection in the financial sector. This study provides important insights for policymakers and practitioners in the financial sector in dealing with complex regulatory challenges related to the adoption of blockchain technology.

Flexible Local Differential Privacy Mechanism for Collecting Location Data

To address issues in existing location data collection methods, such as poor data utility and large deviations between perturbed and true locations, this paper proposes a personalized local differential privacy (LDP) mechanism for location data collection. Users can select their privacy protection range based on their needs, limiting the perturbed output to this range, thereby improving data utility. To address the issue of large deviations between perturbed and true locations, we introduce a new strategy where the location is perturbed with a probability that is higher the closer it is to the true location. By analyzing the mutual information upper bound of the true and estimated location distributions, the optimal perturbation probability range is determined. Finally, a probability transition matrix is generated from the location transfer probabilities, and the true location distribution is estimated from the perturbed location distribution.

A secure and efficient blockchain enabled federated Q-learning model for vehicular Ad-hoc networks

Vehicular Ad-hoc Networks (VANETs) are growing into more desirable targets for malicious individuals due to the quick rise in the number of automated vehicles around the roadside. Secure data transfer is necessary for VANETs to preserve the integrity of the entire network. Federated learning (FL) is often suggested as a safe technique for exchanging data among VANETs, however, its capacity to protect private information is constrained. This research proposes an extra level of security to Federated Q-learning by merging Blockchain technology with VANETs. Initially, traffic data is encrypted utilizing the Extended Elliptic Curve Cryptography (EX-ECC) technique to enhance the security of data. Then, the Federated Q-learning model trains the data and ensures higher privacy protection. Moreover, interplanetary file system (IPFS) technology allows Blockchain storage to improve the security of VANETs information. Additionally, the validation process of the proposed Blockchain framework is performed by utilizing a Delegated Practical Byzantine Fault Tolerance (DPBFT) based consensus algorithm. The proposed approach to federated Q-learning offered by Blockchain technology has the potential to develop VANET safety and performance. Comprehensive simulation tests are performed with several assessment criteria considered for number of vehicles 100, Throughput (102465.8 KB/s), Communication overhead (360.57 Mb), Average Latency (864.425 ms), Communication Time (19.51 s), Encryption time (0.98 ms), Decryption time (1.97 ms), Consensus delay (50 ms) and Validation delay (1.68 ms), respectively. As a result, the proposed approach performs significantly better than the existing approaches.

Legal Privacy Protection Machine Learning Method Based on Word2Vec Algorithm

This study uses Word2Vec's word vector representation technology to finely capture the semantic relationships of vocabulary in legal texts through the Skip-gram model. By introducing Hierarchical Softmax optimization, a legal privacy protection model based on Word2Vec algorithm is ultimately designed. The results showed that the model performed better than other comparative algorithms in both the macro classification performance (Fl_macro) and the micro classification performance (Fl_micro). In practical legal sensitive word recognition tasks, the accuracy, recall rate, and F1 score of the model reached 92.56%, 88.78%, and 90.62%, respectively. Therefore, the proposed model effectively improved the accuracy of identifying sensitive legal privacy words and providing new methods for the personal information security protection system.

Transfer learning for linear regression with differential privacy

Transfer learning, as a machine learning approach to enhance model generalization, has found widespread applications across various domains. However, the risk of privacy leakage during the transfer process remains a crucial consideration. Differential privacy, with its rigorous mathematical foundations, has been proven to offer consistent and robust privacy protection. This study delves into the problem of linear regression transfer learning under differential privacy and, on this basis, proposes a novel strategy incorporating prior information as a constraint to further enhance model performance and stability. In scenarios where the transferable source is known, a two-step transfer learning algorithm incorporating prior information is proposed. This approach leverages prior knowledge to effectively constrain the model parameters, ensuring that the solution space remains reasonable throughout the transfer process. For cases where transferable sources are unknown, a non-algorithmic, cross-validation-based method for transferable source detection is introduced to mitigate adverse impacts stemming from non-informative sources. The effectiveness of the proposed algorithms is validated through simulations and real-world data experiments.

Verifiable attribute-based multi-keyword ranked search scheme in blockchain

Abstract In recent years, cloud storage has gradually become a promising tool for providing data sharing and data storage services with privacy protection. However, to enable users to search over encrypted data and to enable data owners to perform fine-grained search authorization on their encrypted files is still a great challenge. Although attribute-based keyword search (ABKS) is a well-received solution to the challenge, the direct adoption of the traditional ABKS schemes in cloud storage suffers two issues. The first problem is the lack of integrity verification of search results. That is, the users cannot verify whether the server has actually performed a search without compromise. The second issue is that the retrieval modes of most ABKS schemes are not flexible enough. Aiming at the above two issues, a verifiable ABKS scheme is proposed herein. Besides, we proposed security definitions for two types of adversaries and proved that the proposed scheme is verifiable and able of resisting outside keyword guessing attack and chosen keyword attack. Finally, we carried out all-round simulations with actual data set, showing that the proposed scheme has advantage in efficiency over other similar schemes.

Review of the Problems and Legal Protection for Parties in the Implementation of Authentic Deeds Made Based on the Cyber Notary Concept in Indonesia

In the ever-evolving digital era, information technology has had a significant impact on various aspects of life, including in the fields of law and notary. One of the important innovations that has emerged is the concept of cyber notary, which uses digital technology in making authentic deeds, replacing conventional methods. It is hoped that cyber notary can increase efficiency, transparency, and accessibility in notary services, as well as reduce bureaucracy that often takes time and costs. However, the implementation of the cyber notary concept in Indonesia faces various complex problems. One of the main issues is the legal validity of deeds made digitally, including the recognition of electronic signatures and the process of authenticating the identities of the parties. In addition, data security and privacy protection are challenges in ensuring that information stored and transmitted digitally is not misused or hacked. On the other hand, the current regulations do not fully accommodate cyber notary practices. Legal gaps and unclear regulations create uncertainty for notaries and the parties who use their services. This reinforces the urgency to examine more deeply the need for comprehensive and adaptive legal updates to technological developments. This study focuses on the analysis of the problems in the implementation of authentic deeds based on cyber notary and legal protection for the parties involved. With a normative legal approach, this study aims to identify the main challenges and provide recommendations for policy makers and legal practitioners to improve the existing legal framework. It is hoped that the solutions found can support digital transformation in the notary sector in Indonesia, creating a more efficient, secure, and reliable system.

Cross-Platform Advertising Integration Strategies from a Brand Communication Perspective Research

With the rapid development of digitalization and multi-media environment, cross-platform advertising integration has gradually become an important way of brand communication. From the perspective of brand communication and based on the significance of cross-platform advertising integration, this study puts forward the strategies of improving laws and regulations and regulatory mechanisms, strengthening R&D investment in data privacy and security protection, and cultivating and introducing composite talents in response to the problems of unsound laws, regulations and regulatory mechanisms, outstanding data privacy and security issues, and the lack of composite talents that exist in the cross-platform advertising integration, which provides a new perspective on the construction of the theoretical system of cross-platform advertising integration, and also provides a reference for brand communication practice in the digital era. It provides a new perspective for the construction of the theoretical system of cross-platform advertising integration, and also provides a reference for the communication practice of brands in the digital era.

The Influence of Information Exposure on Teenagers’ Privacy Protection Behaviour under the Intelligent Recommendation System

This study focuses on the influence of information exposure under an intelligent recommendation system on teenagers’ privacy protection behaviour. By employing the Communication Privacy Management (CPM) theory, we examine how information exposure through intelligent recommendation systems influences teenagers’ attitudes and actions towards privacy. The results of this study emphasize the need for the CPM advocated by Petronio [1] and the role of various factors in influencing privacy protection behaviour. The study also explores teenagers’ privacy protection behaviours on social media. Through quantitative research, this study provides a perspective on the interaction among adolescents’ digital literacy, privacy issues and behavioural responses in the context of intelligent recommendation systems.

Current Situation and Countermeasures of Teaching Classroom in the Context of Artificial Intelligence

The purpose of this paper is to discuss the status quo of the teaching classroom under the background of artificial intelligence as well as the existing problems, and put forward corresponding countermeasures and suggestions. First of all, this paper outlines the application of artificial intelligence in education, including intelligent teaching system, personalized recommended learning resources, etc., and introduces the changes in teaching mode, such as flipped classroom and online education. Meanwhile, the change of teachers' roles and the change of students' learning styles are also important features of the teaching classroom in the context of artificial intelligence. However, there are some problems in the teaching classroom in the context of artificial intelligence. Technical problems are one of them, including the stability of the system and the accuracy of data. The issue of educational equity is also not to be ignored, as the gap in technological resources leads to inequity in learning opportunities among students. In addition, teacher training issues and student privacy protection issues are also important issues that need to be addressed. In order to address these issues, this paper presents some countermeasures and recommendations. Technological improvements are necessary, including improving the reliability of the system and protecting the security of data. Policy support is also key, and the government should increase investment in AI education and promote the balanced distribution of educational resources. In addition, teacher training and development is also important. Teachers should improve their educational technology skills and adapt to the needs of teaching in the context of AI. At the same time, the protection of students' rights and interests is also necessary, and relevant laws and regulations should be formulated to protect students' privacy and rights and interests.