Malware detection for mobile computing using secure and privacy-preserving machine learning approaches: A comprehensive survey

Computers & Electrical Engineering, Год журнала: 2024, Номер 117, С. 109233 - 109233

Опубликована: Апрель 11, 2024

Mobile devices have become an essential element in our day-to-day lives. The chances of mobile attacks are rapidly increasing with the growing use devices. Exploiting vulnerabilities from as well stealing personal information, principal targets attackers. Researchers also developing various techniques for detecting and analyzing malware to overcome these issues. As new gets introduced frequently by developers, it is very challenging come up comprehensive algorithms detect this malware. There many machine-learning deep-learning been developed researchers. accuracy models largely depends on size quality training dataset. Training model a diversified dataset necessary predict accurately. However, process may raise issue privacy loss due disclosure sensitive information users. proposed mitigate issue, such differential privacy, homomorphic encryption, federated learning. This survey paper explores significance applying learning operating systems, contrasting traditional machine deep approaches detection. We delve into unique challenges opportunities architecture in-built systems their implications user security. Moreover, we assess risks associated real-life applications recommend strategies secure framework domain

Язык: Английский

---

Blockchain-Enabled Federated Reinforcement Learning (B-FRL) Model for Privacy Preservation Service in IoT Systems

Wireless Personal Communications, Год журнала: 2024, Номер 136(4), С. 2545 - 2571

Опубликована: Июнь 1, 2024

Язык: Английский

---

Tech Business Analytics in Quaternary Industry Sector

International Journal of Case Studies in Business IT and Education, Год журнала: 2024, Номер unknown, С. 69 - 159

Опубликована: Апрель 19, 2024

Purpose: The knowledge-based segment of the economy is referred to as "quaternary sector," which comprises businesses like information technology, telecommunications, research and development, other professional services. Businesses in this industry may find that technology-driven business analytics greatly aids helping them make data-driven decisions, optimize workflows, enhance overall performance. Utilizing technology analyse can significantly improve market trends, consumer behaviour, an organization's operational Through analysis data, companies more informed decisions support expansion competitiveness. Analytics tools assist identifying inefficiencies their processes operations so they changes reduce expenses, boost output, ultimately revenue. Customer loyalty satisfaction rise a result this. Information regarding emerging technologies integration with data science prediction trends could present chances for growth innovation. Methodology: There are particular potential challenges Quaternary sector because its emphasis on activities, innovation, cutting-edge technology. Here, we methodical strategy using industry, allowing obtain useful long-term planning calculations. This approach gives framework utilizing analytics. helps competitive advantages increasingly environment by access important insights spur Findings/Result: study looks at how digital have been used control from birth present. Originality/Value: An explanation tech differs traditional within industry. It also includes general design be technical purposes, it examines thirty recently submitted recommendations related Tech Business industries. Paper Type: Exploratory research.

Язык: Английский

---

Privacy-Preserving Federated Learning for Intrusion Detection in IoT Environments: A Survey

IEEE Access, Год журнала: 2024, Номер 12, С. 127018 - 127050

Опубликована: Янв. 1, 2024

Язык: Английский

---

VeriProd: A Privacy‐Preserving and Verifiable FL Framework for Secure Aggregation and Dropout Resilience

Security and Privacy, Год журнала: 2025, Номер 8(4)

Опубликована: Май 20, 2025

ABSTRACT Federated learning (FL) is a decentralized machine approach where multiple devices collaboratively train global model without directly sharing their raw data. This method utilizes local computational resources while depending on central server for coordination. Although FL enhances efficiency in edge computing, it susceptible to adversarial attacks. A compromised aggregator can degrade performance by introducing data or poisoning. To address these risks, must safeguard the confidentiality and integrity of updates ensuring authenticity before aggregation. work introduces VeriProd , verifiable, privacy‐preserving framework designed practical applications, addressing challenge implementing secure aggregation at scale. The preserves user privacy verifiability gradients. proposed encryption securely masks users' gradients, allowing aggregate them revealing individual Simultaneously, users verify correctness aggregated results. Additionally, incorporates group management mechanism handle dropouts, latter seamlessly rejoin future rounds disruption. Through comprehensive analysis experimental evaluation, we demonstrate framework's security, robustness, efficiency. results indicate that achieves accuracy comparable FedAvg maintaining strong communication computation costs, outperforming well‐established verifiable schemes.

Язык: Английский

---

Vulnerable JavaScript functions detection using stacking of convolutional neural networks

PeerJ Computer Science, Год журнала: 2024, Номер 10, С. e1838 - e1838

Опубликована: Фев. 29, 2024

System security for web-based applications is paramount, and the avoidance of possible cyberattacks it important to detect vulnerable JavaScript functions. Developers analysts have long relied upon static analysis investigate vulnerabilities faults within programs. Static tools are used analyzing a program’s source code identifying sections that need be further examined by human analyst. This article suggests new approach in programs using ensemble convolutional neural networks (CNNs) models. These models use information features related code. For different functions, an has been tested which involves stacking CNNs with misbalancing, random under sampler, over sampler. Our uses these improve current techniques’ limitations. Previous research introduced several approaches identify programs, but often their own limitations such as low accuracy rates high false-positive or false-negative results. addresses this power proven highly effective detection functions could cybercriminals. The stacked CNN approximately 98% accuracy, proving its robustness usability real-world scenarios. To evaluate efficacy, proposed method trained publicly available blocks, results assessed various performance metrics. offers valuable insight into better ways protect systems from potential threats, leading safer online environment all.

Язык: Английский

---

IPFS-blockchain-based delegation model for internet of medical robotics things telesurgery system

Connection Science, Год журнала: 2024, Номер 36(1)

Опубликована: Июнь 17, 2024

The concept of the Internet Medical Robotics Things (IoMRT) is where intelligent robots assess surrounding events, combine information from their sensors, use both local and dispersed intelligence to determine best course action, move or command objects. Telesurgery one application IoMRT (TS). With 5G-enabled Tactile (TI) enabling telesurgery (TS), there ample opportunity provide exceptional, accurate, ultra-responsive, real-time virtual surgical procedures. potential for accurate diagnosis involving exchange patient electronic medical records (EMR) with several doctors using an assistant robot (AR) could be greatly useful in field. As a part this, permission delegation has emerged as novel approach data sharing TI. Robust control access guidelines combined configurable scheme promise secure EMR exchange. present research proposes multi-hop strategy based on blockchain technology depth. Furthermore, original EMRs are stored interplanetary file system (IPFS). Permission uses smart contracts proxy re-encryption technology. Attribute-based encryption, which offers fine-grained management access, used guarantee security. Blockchain also utilized accomplish immutability traceability. Delegators may regulate depth by contracts. suggested satisfies intended aims, according analysis protocol. Lastly, Ethereum test chain put method into practice. outcomes conducted experiments demonstrate that protocol operates better than competitors.

Язык: Английский

---

A privacy-preserving federated learning protocol with a secure data aggregation for the Internet of Everything

Computer Communications, Год журнала: 2024, Номер 223, С. 1 - 14

Опубликована: Май 10, 2024

Язык: Английский

---

Assessing the Effect of Model Poisoning Attacks on Federated Learning in Android Malware Detection

Опубликована: Май 25, 2024

Android devices are central to our daily lives, which leads an increase in mobile security threats. Attackers try exploit vulnerabilities and steal personal information from the installed applications on these devices. Because of their widespread usage, prime targets cyber attacks. To get rid this, malware detection has become increasingly significant. Federated learning, is a decentralized machine learning approach, been utilized improve privacy sensitive user data. However, integration federated also introduces vulnerability model poisoning attacks, where adversaries deliberately bias process impair performance metrics. This paper presents comprehensive assessment effect attacks systems deployed for detection. We explain exhaustive feature selection methodology that employs both static dynamic features created novel dataset. focus incorporating recent samples while creating dataset make robust adaptable new malware. Furthermore, we quantify degradation accuracy reliability following attack scenario through series experiments. Additionally, explore defense mechanisms mitigate based studies.

Язык: Английский

---

RAX-ClaMal: Dynamic android malware classification based on RAX register values

Internet of Things, Год журнала: 2024, Номер unknown, С. 101482 - 101482

Опубликована: Дек. 1, 2024

Язык: Английский

---