Proceedings of the 28th Asia and South Pacific Design Automation Conference, Journal Year: 2025, Volume and Issue: unknown, P. 1302 - 1307
Published: Jan. 20, 2025
Language: Английский
Information Fusion, Journal Year: 2023, Volume and Issue: 99, P. 101896 - 101896
Published: June 24, 2023
Trustworthy Artificial Intelligence (AI) is based on seven technical requirements sustained over three main pillars that should be met throughout the system's entire life cycle: it (1) lawful, (2) ethical, and (3) robust, both from a social perspective. However, attaining truly trustworthy AI concerns wider vision comprises trustworthiness of all processes actors are part cycle, considers previous aspects different lenses. A more holistic contemplates four essential axes: global principles for ethical use development AI-based systems, philosophical take ethics, risk-based approach to regulation, mentioned requirements. The (human agency oversight; robustness safety; privacy data governance; transparency; diversity, non-discrimination fairness; societal environmental wellbeing; accountability) analyzed triple perspective: What each requirement is, Why needed, How can implemented in practice. On other hand, practical implement systems allows defining concept responsibility facing law, through given auditing process. Therefore, responsible system resulting notion we introduce this work, utmost necessity realized processes, subject challenges posed by regulatory sandboxes. Our multidisciplinary culminates debate diverging views published lately about future AI. reflections matter conclude regulation key reaching consensus among these views, will crucial present our society.
Language: Английский
Citations
316
Future Internet, Journal Year: 2025, Volume and Issue: 17(1), P. 30 - 30
Published: Jan. 11, 2025
The proliferation of the Internet Things (IoT) has transformed digital landscape, enabling a vast array interconnected devices to communicate and share data seamlessly. However, rapid expansion IoT networks also introduced significant cybersecurity challenges. This paper presents comprehensive survey in ecosystem, examining current state research, identifying critical security vulnerabilities, exploring advanced strategies for mitigating threats. covers various facets security, including device authentication, integrity, privacy, network emerging role artificial intelligence (AI) bolstering defenses. By synthesizing existing research highlighting ongoing challenges, this aims provide holistic understanding guide future endeavors.
Language: Английский
Citations
5
Sensors, Journal Year: 2023, Volume and Issue: 23(7), P. 3566 - 3566
Published: March 29, 2023
The advancement of biometric technology has facilitated wide applications biometrics in law enforcement, border control, healthcare and financial identification verification. Given the peculiarity features (e.g., unchangeability, permanence uniqueness), security data is a key area research. Security privacy are vital to enacting integrity, reliability availability biometric-related applications. Homomorphic encryption (HE) concerned with manipulation cryptographic domain, thus addressing issues faced by biometrics. This survey provides comprehensive review state-of-the-art HE research context Detailed analyses discussions conducted on various approaches according categories different traits. Moreover, this presents perspective integrating other emerging technologies machine/deep learning blockchain) for security. Finally, based latest development biometrics, challenges future directions put forward.
Language: Английский
Citations
33
IEEE Communications Surveys & Tutorials, Journal Year: 2023, Volume and Issue: 25(4), P. 2654 - 2713
Published: Jan. 1, 2023
The deployment of the fifth-generation (5G) wireless networks in Internet Everything (IoE) applications and future (e.g., sixth-generation (6G) networks) has raised a number operational challenges limitations, for example terms security privacy. Edge learning is an emerging approach to training models across distributed clients while ensuring data Such when integrated network infrastructures 6G) can potentially solve challenging problems such as resource management behavior prediction. However, edge (including deep learning) are known be susceptible tampering manipulation. This survey article provides holistic review extant literature focusing on learning-related vulnerabilities defenses 6G-enabled Things (IoT) systems. Existing machine approaches 6G–IoT learning-associated threats broadly categorized based modes, namely: centralized, federated, distributed. Then, we provide overview enabling technologies intelligence. We also existing research attacks against classify threat into eight categories, backdoor attacks, adversarial examples, combined poisoning Sybil byzantine inference dropping attacks. In addition, comprehensive detailed taxonomy comparative summary state-of-the-art defense methods vulnerabilities. Finally, new realized, overall prospects IoT discussed.
Language: Английский
Citations
24
Egyptian Informatics Journal, Journal Year: 2024, Volume and Issue: 27, P. 100519 - 100519
Published: Aug. 23, 2024
Cloud computing security and data protection are becoming increasingly critical as its use increases. The research we present demonstrates how undercover sharing techniques homomorphic encryption can be combined to protect private information in cloud scenarios. We create a reliable, private, confidential computation platform by utilizing this dual approach. Our strategy involves protecting while dividing it among multiple servers. By using distribution, the system is less likely suffer from single points of failure has higher level. To ensure privacy security, restricts access authorized individuals only. As an additional feature, employ enable operations on encrypted without direct originals. sensitive protected disclosure or misuse being processed. Therefore, original confidentiality preserved when shares. Several performance tests were conducted prove our strategy's practicality effectiveness. considerations extended beyond decryption time processing overhead. In research, demonstrate that method strikes right balance between computational efficiency.
Language: Английский
Citations
7
Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(3)
Published: Jan. 6, 2025
Machine Learning (ML) is rapidly advancing, enabling various applications that improve people's work and daily lives. However, this technical progress brings privacy concerns, leading to the emergence of Privacy-Preserving (PPML) as a popular research topic. In work, we investigate protection topic in ML, showcase advantages Homomorphic Encryption (HE) among different privacy-preserving techniques. Additionally, presents an introduction approximate HE, emphasizing its providing detail some representative schemes. Moreover, systematically review related works about HE based PPML schemes from four three advanced applications, along with their application scenarios, models datasets. Finally, suggest potential future directions guide readers extending PPML.
Language: Английский
Citations
1
Lecture notes in computer science, Journal Year: 2023, Volume and Issue: unknown, P. 370 - 394
Published: Jan. 1, 2023
Language: Английский
Citations
15
Sensors, Journal Year: 2023, Volume and Issue: 23(10), P. 4640 - 4640
Published: May 10, 2023
To ensure the success of energy transition and achieve target reducing carbon footprint systems, management systems needs to be decentralized. Public blockchains offer favorable features support sector democratization reinforce citizens' trust, such as tamper-proof data registration sharing, decentralization, transparency, for peer-to-peer (P2P) trading. However, in blockchain-based P2P markets, transactional are public accessible, which raises privacy concerns related prosumers' profiles while lacking scalability featuring high costs. In this paper, we employ secure multi-party computation (MPC) assure on a flexibility market implementation Ethereum by combining orders storing it safely chain. We provide an encoding mechanism obfuscate amount traded creating groups prosumers, splitting from bids offers, group-level orders. The solution wraps around smart contracts-based marketplace, assuring all operations order submission, matching commitment trading settlement. experimental results show that proposed is effective supporting trading, number transactions, gas consumption with limited computational time overhead.
Language: Английский
Citations
14
IEEE Transactions on Industrial Informatics, Journal Year: 2024, Volume and Issue: 20(9), P. 10802 - 10813
Published: May 16, 2024
Language: Английский
Citations
6
Mathematics, Journal Year: 2023, Volume and Issue: 11(13), P. 2948 - 2948
Published: July 1, 2023
The widespread adoption of cloud infrastructures has revolutionized data storage and access. However, it also raised concerns regarding the privacy sensitive data. To address these concerns, encryption techniques have been widely used. traditional schemes limit efficient search retrieval encrypted tackle this challenge, innovative approaches emerged, such as utilization Homomorphic Encryption (HE) in Searchable (SE) schemes. This paper provides a comprehensive analysis advancements HE-based privacy-preserving techniques, focusing on their application SE. main contributions work include identification classification existing SE that utilize HE, types HE used SE, an examination how shapes process structure enables additional functionalities, promising directions for future research findings reveal increasing usage schemes, particularly Partially Encryption. popularity type especially Paillier’s cryptosystem, can be attributed to its simplicity, proven security properties, availability open-source libraries. highlights prevalence index-based using support ranked multi-keyword queries, need further exploration functionalities verifiability ability authorize revoke users. Future exploring other alongside addressing omissions like fuzzy keyword search, leveraging recent Fully
Language: Английский
Citations
13