Computer Standards & Interfaces最新文献_第3页

ARMOR: A multi-layered adaptive defense framework for robust deep learning systems against evolving adversarial threats ARMOR：一个多层自适应防御框架，用于鲁棒深度学习系统抵御不断发展的敌对威胁

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-17 DOI: 10.1016/j.csi.2025.104117

Mahmoud Mohamed, Fayaz AlJuaid

Introduction:

Adversarial attacks represent a major challenge to deep learning models deployed in critical fields such as healthcare diagnostics and financial fraud detection. This paper addresses the limitations of single-strategy defenses by introducing ARMOR (Adaptive Resilient Multi-layer Orchestrated Response), a novel multi-layered architecture that seamlessly integrates multiple defense mechanisms.

Methodology:

We evaluate ARMOR against seven state-of-the-art defense methods through extensive experiments across multiple datasets and five attack methodologies. Our approach combines adversarial detection, input transformation, model hardening, and adaptive response layers that operate with intentional dependencies and feedback mechanisms.

Results:

Quantitative results demonstrate that ARMOR significantly outperforms individual defense methods, achieving a 91.7% attack mitigation rate (18.3% improvement over ensemble averaging), 87.5% clean accuracy preservation (8.9% improvement over adversarial training alone), and 76.4% robustness against adaptive attacks (23.2% increase over the strongest baseline).

Discussion:

The modular framework design enables flexibility against emerging threats while requiring only 1.42

\times

computational overhead compared to unprotected models, making it suitable for resource-constrained environments. Our findings demonstrate that activating and integrating complementary defense mechanisms represents a significant advance in adversarial resilience.

导读：对抗性攻击对部署在医疗诊断和金融欺诈检测等关键领域的深度学习模型构成了重大挑战。本文通过引入自适应弹性多层协调响应（ARMOR）来解决单策略防御的局限性，这是一种无缝集成多种防御机制的新型多层体系结构。方法：我们通过跨多个数据集和五种攻击方法的广泛实验，对七种最先进的防御方法进行评估。我们的方法结合了对抗检测、输入转换、模型强化和自适应响应层，这些层与有意的依赖关系和反馈机制一起运作。结果：定量结果表明，ARMOR显著优于单个防御方法，实现了91.7%的攻击缓解率（比集合平均提高18.3%），87.5%的干净准确性保持（比单独的对抗性训练提高8.9%），以及76.4%的自适应攻击鲁棒性（比最强基线提高23.2%）。讨论：模块化框架设计能够灵活地应对新出现的威胁，而与未受保护的模型相比，只需要1.42倍的计算开销，使其适合资源受限的环境。我们的研究结果表明，激活和整合互补防御机制代表了对抗弹性的重大进步。

{"title":"ARMOR: A multi-layered adaptive defense framework for robust deep learning systems against evolving adversarial threats","authors":"Mahmoud Mohamed, Fayaz AlJuaid","doi":"10.1016/j.csi.2025.104117","DOIUrl":"10.1016/j.csi.2025.104117","url":null,"abstract":"<div><h3>Introduction:</h3><div>Adversarial attacks represent a major challenge to deep learning models deployed in critical fields such as healthcare diagnostics and financial fraud detection. This paper addresses the limitations of single-strategy defenses by introducing ARMOR (Adaptive Resilient Multi-layer Orchestrated Response), a novel multi-layered architecture that seamlessly integrates multiple defense mechanisms.</div></div><div><h3>Methodology:</h3><div>We evaluate ARMOR against seven state-of-the-art defense methods through extensive experiments across multiple datasets and five attack methodologies. Our approach combines adversarial detection, input transformation, model hardening, and adaptive response layers that operate with intentional dependencies and feedback mechanisms.</div></div><div><h3>Results:</h3><div>Quantitative results demonstrate that ARMOR significantly outperforms individual defense methods, achieving a 91.7% attack mitigation rate (18.3% improvement over ensemble averaging), 87.5% clean accuracy preservation (8.9% improvement over adversarial training alone), and 76.4% robustness against adaptive attacks (23.2% increase over the strongest baseline).</div></div><div><h3>Discussion:</h3><div>The modular framework design enables flexibility against emerging threats while requiring only 1.42<span><math><mo>×</mo></math></span> computational overhead compared to unprotected models, making it suitable for resource-constrained environments. Our findings demonstrate that activating and integrating complementary defense mechanisms represents a significant advance in adversarial resilience.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104117"},"PeriodicalIF":3.1,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Chaos experiments in microservice architectures: A systematic literature review 微服务架构中的混沌实验：系统的文献综述

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-15 DOI: 10.1016/j.csi.2025.104116

Emrah Esen , Akhan Akbulut , Cagatay Catal

This study analyzes the implementation of Chaos Engineering in modern microservice systems. It identifies key methods, tools, and practices used to effectively enhance the resilience of software systems in production environments. In this context, our Systematic Literature Review (SLR) of 31 research articles has uncovered 38 tools crucial for carrying out fault injection methods, including several tools such as Chaos Toolkit, Gremlin, and Chaos Machine. The study also explores the platforms used for chaos experiments and how centralized management of chaos engineering can facilitate the coordination of these experiments across complex systems. The evaluated literature reveals the efficacy of chaos engineering in improving fault tolerance and robustness of software systems, particularly those based on microservice architectures. The paper underlines the importance of careful planning and execution in implementing chaos engineering and encourages further research in this field to uncover more effective practices for the resilience improvement of microservice systems.

本研究分析了混沌工程在现代微服务系统中的应用。它确定了用于有效增强生产环境中软件系统弹性的关键方法、工具和实践。在此背景下，我们对31篇研究文章进行了系统文献综述（SLR），发现了38个工具对于执行故障注入方法至关重要，包括一些工具，如混沌工具箱、Gremlin和混沌机器。该研究还探讨了用于混沌实验的平台，以及混沌工程的集中管理如何促进跨复杂系统的这些实验的协调。评估的文献揭示了混沌工程在提高软件系统容错性和鲁棒性方面的有效性，特别是基于微服务架构的软件系统。本文强调了在实施混沌工程时仔细规划和执行的重要性，并鼓励在该领域进一步研究，以发现更有效的微服务系统弹性改进实践。

引用次数: 0

Post-quantum PAKE over lattices revised: Smaug-T.PAKE for mobile devices 后量子PAKE在晶格上的修正：smaugt。移动设备的PAKE

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-15 DOI: 10.1016/j.csi.2025.104118

Kübra Seyhan , Sedat Akleylek , Ahmet Faruk Dursun

In this paper, an efficient post-quantum secure password-authenticated key exchange (PAKE) scheme from a well-structured lattice-based key encapsulation mechanism (KEM) is proposed. The generic KEM to PAKE idea, OCAKE, is modified by considering hybrid module learning with errors (MLWE) + module learning with rounding (MLWR) assumptions to obtain explicit password-based authentication from SMAUG-T.KEM procedures. As a KEM primitive, SMAUG-T.KEM is chosen due to its performance against the National Institute of Standards and Technology (NIST) standard Crystals-Kyber (Kyber) to obtain an efficient and post-quantum secure PAKE scheme. Firstly, the anonymity and fuzziness properties of SMAUG-T.KEM are proven to fit the OCAKE approach in constructing the PAKE version of Smaug.KEM. Then, the post-quantum security of the proposed SMAUG-T.PAKE is analyzed in the universal composability (UC) model based on the hybrid security assumptions and proved properties. The reference C and JAVA codes are written to evaluate whether the targeted efficiency is achieved in different platforms. Based on the central processing unit (CPU) and memory usage, run time, and energy consumption metrics, the proposed solution is compared with current PAKE proposals. The performance results showed that SMAUG-T.PAKE, with two optional encryption modes, Advanced Encryption Standard (AES) or Ascon, presents better performance than the other module-based PAKE solutions from lattices in terms of both reference and mobile results.

本文提出了一种基于结构良好的格子密钥封装机制的后量子安全密码认证密钥交换方案。通过考虑混合误差模块学习(MLWE) +舍入模块学习（MLWR）假设，对通用的KEM到PAKE思想OCAKE进行了改进，从而从smaugt获得显式的基于密码的身份验证。克姆程序。作为KEM原语smaugt。选择KEM是由于它与美国国家标准与技术研究所（NIST）标准晶体-Kyber （Kyber）的性能相比较，以获得有效的后量子安全PAKE方案。首先，SMAUG-T的匿名性和模糊性。在构建Smaug.KEM的PAKE版本时，KEM已被证明适合OCAKE方法。然后，讨论了SMAUG-T的后量子安全性。基于混合安全假设和证明的性质，在通用可组合性（UC）模型下对PAKE进行了分析。编写了参考C和JAVA代码来评估在不同的平台上是否实现了目标效率。基于中央处理器（CPU）和内存使用、运行时间和能耗指标，将所提出的解决方案与当前的PAKE方案进行比较。性能结果表明，smaugt。PAKE具有高级加密标准（AES）或Ascon两种可选加密模式，在参考和移动结果方面都比其他基于模块的PAKE解决方案具有更好的性能。

{"title":"Post-quantum PAKE over lattices revised: Smaug-T.PAKE for mobile devices","authors":"Kübra Seyhan , Sedat Akleylek , Ahmet Faruk Dursun","doi":"10.1016/j.csi.2025.104118","DOIUrl":"10.1016/j.csi.2025.104118","url":null,"abstract":"<div><div>In this paper, an efficient post-quantum secure password-authenticated key exchange (PAKE) scheme from a well-structured lattice-based key encapsulation mechanism (KEM) is proposed. The generic KEM to PAKE idea, OCAKE, is modified by considering hybrid module learning with errors (MLWE) + module learning with rounding (MLWR) assumptions to obtain explicit password-based authentication from SMAUG-T.KEM procedures. As a KEM primitive, SMAUG-T.KEM is chosen due to its performance against the National Institute of Standards and Technology (NIST) standard Crystals-Kyber (Kyber) to obtain an efficient and post-quantum secure PAKE scheme. Firstly, the anonymity and fuzziness properties of SMAUG-T.KEM are proven to fit the OCAKE approach in constructing the PAKE version of Smaug.KEM. Then, the post-quantum security of the proposed SMAUG-T.PAKE is analyzed in the universal composability (UC) model based on the hybrid security assumptions and proved properties. The reference C and JAVA codes are written to evaluate whether the targeted efficiency is achieved in different platforms. Based on the central processing unit (CPU) and memory usage, run time, and energy consumption metrics, the proposed solution is compared with current PAKE proposals. The performance results showed that SMAUG-T.PAKE, with two optional encryption modes, Advanced Encryption Standard (AES) or Ascon, presents better performance than the other module-based PAKE solutions from lattices in terms of both reference and mobile results.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104118"},"PeriodicalIF":3.1,"publicationDate":"2025-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SiamIDS: A novel cloud-centric Siamese Bi-LSTM framework for interpretable intrusion detection in large-scale IoT networks SiamIDS：一种新的以云为中心的Siamese Bi-LSTM框架，用于大规模物联网网络中的可解释入侵检测

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-15 DOI: 10.1016/j.csi.2025.104119

Prabu Kaliyaperumal , Palani Latha , Selvaraj Palanisamy , Sridhar Pushpanathan , Anand Nayyar , Balamurugan Balusamy , Ahmad Alkhayyat

The rapid proliferation of Internet of Things (IoT) devices has heightened the need for scalable and interpretable intrusion detection systems (IDS) capable of operating efficiently in cloud-centric environments. Existing IDS approaches often struggle with real-time processing, zero-day attack detection, and model transparency. To address these challenges, this paper proposes SiamIDS, a novel cloud-native framework that integrates contrastive Siamese Bi-directional LSTM (Bi-LSTM) modeling, autoencoder-based dimensionality reduction, SHapley Additive exPlanations (SHAP) for interpretability, and Ordering Points To Identify the Clustering Structure (OPTICS) clustering for unsupervised threat categorization. The framework aims to enhance the detection of both known and previously unseen threats in large-scale IoT networks by learning behavioral similarity across network flows. Trained on the CIC IoT-DIAD 2024 dataset, SiamIDS achieves superior detection performance with an F1-score of 99.45%, recall of 98.96%, and precision of 99.94%. Post-detection OPTICS clustering yields a Silhouette Score of 0.901, DBI of 0.092, and ARI of 0.889, supporting accurate threat grouping. The system processes over 220,000 samples/sec with a RAM usage under 1.5 GB, demonstrating real-time readiness. Compared to state-of-the-art methods, SiamIDS improves F1-score by 2.8% and reduces resource overhead by up to 25%, establishing itself as an accurate, efficient, and explainable IDS for next-generation IoT ecosystems.

物联网（IoT）设备的快速扩散，提高了对可扩展和可解释的入侵检测系统（IDS）的需求，这些系统能够在以云为中心的环境中高效运行。现有的IDS方法经常与实时处理、零日攻击检测和模型透明性作斗争。为了解决这些挑战，本文提出了SiamIDS，这是一种新的云原生框架，它集成了对比Siamese双向LSTM （Bi-LSTM）建模、基于自编码器的降维、SHapley加性解释（SHAP）的可解释性，以及用于无监督威胁分类的排序点识别聚类结构（OPTICS）聚类。该框架旨在通过学习跨网络流的行为相似性来增强对大规模物联网网络中已知和以前未见过的威胁的检测。在CIC IoT-DIAD 2024数据集上训练后，SiamIDS达到了优异的检测性能，f1得分为99.45%，召回率为98.96%，准确率为99.94%。检测后的OPTICS聚类得到的Silhouette Score为0.901，DBI为0.092，ARI为0.889，支持准确的威胁分组。系统每秒处理超过220,000个样本，RAM使用率低于1.5 GB，显示了实时准备。与最先进的方法相比，SiamIDS将f1分数提高了2.8%，并将资源开销降低了25%，使其成为下一代物联网生态系统中准确、高效、可解释的IDS。

{"title":"SiamIDS: A novel cloud-centric Siamese Bi-LSTM framework for interpretable intrusion detection in large-scale IoT networks","authors":"Prabu Kaliyaperumal , Palani Latha , Selvaraj Palanisamy , Sridhar Pushpanathan , Anand Nayyar , Balamurugan Balusamy , Ahmad Alkhayyat","doi":"10.1016/j.csi.2025.104119","DOIUrl":"10.1016/j.csi.2025.104119","url":null,"abstract":"<div><div>The rapid proliferation of Internet of Things (IoT) devices has heightened the need for scalable and interpretable intrusion detection systems (IDS) capable of operating efficiently in cloud-centric environments. Existing IDS approaches often struggle with real-time processing, zero-day attack detection, and model transparency. To address these challenges, this paper proposes SiamIDS, a novel cloud-native framework that integrates contrastive Siamese Bi-directional LSTM (Bi-LSTM) modeling, autoencoder-based dimensionality reduction, SHapley Additive exPlanations (SHAP) for interpretability, and Ordering Points To Identify the Clustering Structure (OPTICS) clustering for unsupervised threat categorization. The framework aims to enhance the detection of both known and previously unseen threats in large-scale IoT networks by learning behavioral similarity across network flows. Trained on the CIC IoT-DIAD 2024 dataset, SiamIDS achieves superior detection performance with an F1-score of 99.45%, recall of 98.96%, and precision of 99.94%. Post-detection OPTICS clustering yields a Silhouette Score of 0.901, DBI of 0.092, and ARI of 0.889, supporting accurate threat grouping. The system processes over 220,000 samples/sec with a RAM usage under 1.5 GB, demonstrating real-time readiness. Compared to state-of-the-art methods, SiamIDS improves F1-score by 2.8% and reduces resource overhead by up to 25%, establishing itself as an accurate, efficient, and explainable IDS for next-generation IoT ecosystems.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104119"},"PeriodicalIF":3.1,"publicationDate":"2025-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Refining decision boundaries via dynamic label adversarial training for robust traffic classification 基于动态标签对抗训练的鲁棒流量分类决策边界优化

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-13 DOI: 10.1016/j.csi.2025.104111

Haoyu Tong , Meixia Miao , Yundong Liu , Xiaoyu Zhang , Xiangyang Luo , Willy Susilo

Network traffic classification plays a critical role in securing modern communication systems, as it enables the identification of malicious or abnormal patterns within traffic data. With the growing complexity of network environments, deep learning models have emerged as a compelling solution due to their ability to automatically learn discriminative representations from raw traffic. However, these models are highly vulnerable to adversarial examples, which can significantly degrade their performance by introducing imperceptible perturbations. While adversarial training (AT) has emerged as a primary defense, it often suffers from label noise, particularly when hard labels are forcibly assigned to adversarial examples whose true class may be ambiguous. In this work, we first analyze the detrimental effect of label noise on adversarial training, revealing that forcing hard labels onto adversarial examples can cause excessive shifts of the decision boundary away from the adversarial examples, which in turn degrades the model’s generalization. Motivated by the theoretical analysis, we propose Dynamic Label Adversarial Training (DLAT), a novel AT framework that mitigates label noise via dynamically mixed soft labels. DLAT interpolates the logits of clean and adversarial examples to estimate the labels of boundary-adjacent examples, which are then used as soft labels for adversarial examples. By adaptively aligning the decision boundary toward the vicinity of adversarial examples, the framework constrains unnecessary boundary shifts and alleviates generalization degradation caused by label noise. Extensive evaluations on network traffic classification benchmarks validate the effectiveness of DLAT in outperforming standard adversarial training and its variants in both robustness and generalization.

网络流分类在确保现代通信系统的安全方面起着至关重要的作用，因为它可以识别流量数据中的恶意或异常模式。随着网络环境的日益复杂，深度学习模型已经成为一个令人信服的解决方案，因为它们能够从原始流量中自动学习判别表示。然而，这些模型非常容易受到对抗性示例的影响，对抗性示例可以通过引入难以察觉的扰动来显着降低其性能。虽然对抗性训练（AT）已经成为一种主要的防御手段，但它经常受到标签噪音的影响，特别是当硬标签被强制分配给真实类别可能不明确的对抗性示例时。在这项工作中，我们首先分析了标签噪声对对抗训练的有害影响，揭示了将硬标签强加到对抗样本上可能会导致决策边界过度偏离对抗样本，从而降低模型的泛化能力。在理论分析的启发下，我们提出了动态标签对抗训练（Dynamic Label Adversarial Training， DLAT），这是一种通过动态混合软标签来减轻标签噪声的新型标签对抗训练框架。DLAT插值干净和对抗示例的逻辑来估计边界相邻示例的标签，然后将其用作对抗示例的软标签。通过自适应地将决策边界对齐到对抗样本附近，该框架限制了不必要的边界移动，减轻了由标签噪声引起的泛化退化。对网络流量分类基准的广泛评估验证了DLAT在鲁棒性和泛化方面优于标准对抗性训练及其变体的有效性。

{"title":"Refining decision boundaries via dynamic label adversarial training for robust traffic classification","authors":"Haoyu Tong , Meixia Miao , Yundong Liu , Xiaoyu Zhang , Xiangyang Luo , Willy Susilo","doi":"10.1016/j.csi.2025.104111","DOIUrl":"10.1016/j.csi.2025.104111","url":null,"abstract":"<div><div>Network traffic classification plays a critical role in securing modern communication systems, as it enables the identification of malicious or abnormal patterns within traffic data. With the growing complexity of network environments, deep learning models have emerged as a compelling solution due to their ability to automatically learn discriminative representations from raw traffic. However, these models are highly vulnerable to adversarial examples, which can significantly degrade their performance by introducing imperceptible perturbations. While adversarial training (AT) has emerged as a primary defense, it often suffers from label noise, particularly when hard labels are forcibly assigned to adversarial examples whose true class may be ambiguous. In this work, we first analyze the detrimental effect of label noise on adversarial training, revealing that forcing hard labels onto adversarial examples can cause excessive shifts of the decision boundary away from the adversarial examples, which in turn degrades the model’s generalization. Motivated by the theoretical analysis, we propose Dynamic Label Adversarial Training (DLAT), a novel AT framework that mitigates label noise via dynamically mixed soft labels. DLAT interpolates the logits of clean and adversarial examples to estimate the labels of boundary-adjacent examples, which are then used as soft labels for adversarial examples. By adaptively aligning the decision boundary toward the vicinity of adversarial examples, the framework constrains unnecessary boundary shifts and alleviates generalization degradation caused by label noise. Extensive evaluations on network traffic classification benchmarks validate the effectiveness of DLAT in outperforming standard adversarial training and its variants in both robustness and generalization.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104111"},"PeriodicalIF":3.1,"publicationDate":"2025-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Efficient and secure multi-user kNN queries with dynamic POIs updating 具有动态poi更新的高效安全的多用户kNN查询

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-11 DOI: 10.1016/j.csi.2025.104112

Yining Jia , Yali Liu , Congai Zeng , Xujie Ding , Jianting Ning

The

k

-nearest neighbors (

k

NN) query is a key operation in spatial and multimedia databases, which is widely applied in fields such as electronic healthcare and Location-Based Services (LBS). With the rapid development of cloud computing, uploading private data of Data Owner (DO) to Cloud Servers (CS) has become a trend. However, existing

k

NN queries schemes are not designed for multi-user environments, cannot timely update the points of interest (POIs) stored in CS, and suffer from low query efficiency. Therefore, this paper proposes efficient and secure multi-user

k

NN queries with dynamic POIs updating, named DESM

k

NN, which achieves secure multi-user

k

NN queries. To improve query efficiency, DESM

k

NN adopts a two-stage search framework, which consists of an initial filtering stage based on hierarchical clustering to effectively constrain the search range, followed by a more efficient precise search stage. Based on this framework, DESM

k

NN designs a set of security protocols for efficient query processing and enables dynamic POIs updates. Meanwhile, DESM

k

NN not only utilizes Distributed Two Trapdoors Public-Key Cryptosystem (DT-PKC) to enable multi-user queries but also ensures data privacy, query privacy, result privacy and access pattern privacy. Moreover, DESM

k

NN can verify the correctness and completeness of queries results. Finally, security analysis proves that DESM

k

NN meets the formal security definition of multiparty computation, and experimental evaluation shows that DESM

k

NN improves query efficiency by up to 45.5% compared with existing

k

NN queries scheme.

kNN查询是空间和多媒体数据库中的一项关键操作，广泛应用于电子医疗保健和基于位置的服务（LBS）等领域。随着云计算的快速发展，将数据所有者（data Owner， DO）的私有数据上传到云服务器（cloud Servers， CS）已成为一种趋势。然而，现有的kNN查询方案不是针对多用户环境设计的，不能及时更新存储在CS中的兴趣点（poi），并且查询效率较低。为此，本文提出了一种高效、安全且具有动态poi更新的多用户kNN查询方法DESMkNN，实现了安全的多用户kNN查询。为了提高查询效率，DESMkNN采用了两阶段搜索框架，即基于分层聚类的初始过滤阶段有效约束搜索范围，然后是更高效的精确搜索阶段。基于这个框架，DESMkNN设计了一组安全协议，用于高效的查询处理，并支持动态poi更新。同时，DESMkNN不仅利用Distributed Two Trapdoors Public-Key Cryptosystem （DT-PKC）实现多用户查询，还保证了数据隐私、查询隐私、结果隐私和访问模式隐私。此外，DESMkNN可以验证查询结果的正确性和完整性。最后，安全性分析证明DESMkNN符合多方计算的正式安全定义，实验评估表明，与现有的kNN查询方案相比，DESMkNN查询效率提高了45.5%。

{"title":"Efficient and secure multi-user kNN queries with dynamic POIs updating","authors":"Yining Jia , Yali Liu , Congai Zeng , Xujie Ding , Jianting Ning","doi":"10.1016/j.csi.2025.104112","DOIUrl":"10.1016/j.csi.2025.104112","url":null,"abstract":"<div><div>The <span><math><mi>k</mi></math></span>-nearest neighbors (<span><math><mi>k</mi></math></span>NN) query is a key operation in spatial and multimedia databases, which is widely applied in fields such as electronic healthcare and Location-Based Services (LBS). With the rapid development of cloud computing, uploading private data of Data Owner (DO) to Cloud Servers (CS) has become a trend. However, existing <span><math><mi>k</mi></math></span>NN queries schemes are not designed for multi-user environments, cannot timely update the points of interest (POIs) stored in CS, and suffer from low query efficiency. Therefore, this paper proposes efficient and secure multi-user <span><math><mi>k</mi></math></span>NN queries with dynamic POIs updating, named DESM<span><math><mi>k</mi></math></span>NN, which achieves secure multi-user <span><math><mi>k</mi></math></span>NN queries. To improve query efficiency, DESM<span><math><mi>k</mi></math></span>NN adopts a two-stage search framework, which consists of an initial filtering stage based on hierarchical clustering to effectively constrain the search range, followed by a more efficient precise search stage. Based on this framework, DESM<span><math><mi>k</mi></math></span>NN designs a set of security protocols for efficient query processing and enables dynamic POIs updates. Meanwhile, DESM<span><math><mi>k</mi></math></span>NN not only utilizes Distributed Two Trapdoors Public-Key Cryptosystem (DT-PKC) to enable multi-user queries but also ensures data privacy, query privacy, result privacy and access pattern privacy. Moreover, DESM<span><math><mi>k</mi></math></span>NN can verify the correctness and completeness of queries results. Finally, security analysis proves that DESM<span><math><mi>k</mi></math></span>NN meets the formal security definition of multiparty computation, and experimental evaluation shows that DESM<span><math><mi>k</mi></math></span>NN improves query efficiency by up to 45.5% compared with existing <span><math><mi>k</mi></math></span>NN queries scheme.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104112"},"PeriodicalIF":3.1,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790409","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Co-distillation-based defense framework for federated knowledge graph embedding against poisoning attacks 基于协同蒸馏的联邦知识图嵌入中毒攻击防御框架

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-09 DOI: 10.1016/j.csi.2025.104113

Yiqin Lu, Jiarui Chen, Jiancheng Qin

Federated knowledge graph embedding (FKGE) enables collaborative knowledge sharing without data exchange, but it also introduces risks of poisoning attacks that degrade model accuracy or force incorrect outputs. Protecting FKGE from poisoning attacks becomes a critical research problem. This paper reveals the malicious strategy of untargeted FKGE poisoning attacks and proposes CoDFKGE, a co-distillation-based FKGE framework for defending against poisoning attacks. CoDFKGE deploys two collaborative knowledge graph embedding models on clients, decoupling prediction parameters from shared parameters as a model-agnostic solution. By designing distinct distillation loss functions, CoDFKGE transfers clean knowledge from potentially poisoned shared parameters while compressing dimensions to reduce communication overhead. Experiments show CoDFKGE preserves link prediction performance with lower communication costs, eliminates malicious manipulations under targeted poisoning attacks, and significantly mitigates accuracy degradation under untargeted poisoning attacks.

联邦知识图嵌入（FKGE）可以在没有数据交换的情况下实现协作知识共享，但它也引入了中毒攻击的风险，这种攻击会降低模型的准确性或强制输出错误。保护FKGE免受投毒攻击成为关键的研究问题。本文揭示了非目标FKGE投毒攻击的恶意策略，提出了基于共蒸馏的FKGE框架CoDFKGE来防御投毒攻击。CoDFKGE在客户端部署了两个协作知识图嵌入模型，将预测参数与共享参数解耦，作为模型不可知的解决方案。通过设计不同的蒸馏损失函数，CoDFKGE从潜在的有毒共享参数中转移干净的知识，同时压缩维度以减少通信开销。实验表明，CoDFKGE在保持链路预测性能的同时降低了通信成本，消除了针对性投毒攻击下的恶意操作，并显著减轻了非针对性投毒攻击下的精度下降。

引用次数: 0

Robust zero-watermarking method for multi-medical images based on Chebyshev–Fourier moments and Contourlet-FFT 基于chebyhev - fourier矩和Contourlet-FFT的多医学图像鲁棒零水印方法

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-08 DOI: 10.1016/j.csi.2025.104115

Xinhui Lu , Guangyun Yang , Yu Lu , Xiangguang Xiong

Classical robust watermarking methods embed secret data into a cover image designed to protect its copyright. However, they suffer from the problem of balancing imperceptibility and robustness. To address this issue, the impact of conventional attacks on the stability of feature vectors extracted from the cover image is examined. Accordingly, we proposed a zero-watermarking method with high attack resistance for multi-medical images by employing Contourlet transform (CT), Chebyshev–Fourier moments (CHFMs), and fast Fourier transform (FFT). First, each medical image is normalized separately, and the normalized images are fused using a dual-tree complex wavelet transform-based method. Second, the effective region is extracted and subjected to the CT. The CHFMs of the low-frequency sub-bands are calculated, and the FFT is performed on the generated amplitude sequence to construct a feature matrix. A feature image is generated by combining the magnitude of each feature value with the overall mean. Finally, the copyrighted image is encrypted using the Lorenz chaotic system and Fibonacci Q-matrix, after which an exclusive-OR operation is applied between the generated feature image and the encrypted copyrighted image to produce a zero-watermarking signal. The results show that the proposed method exhibits excellent resistance to attack with a normalized correlation coefficient of up to 0.994 between the extracted image and the original copyrighted one. Furthermore, the average anti-attack performance of our proposed method is approximately 2% higher compared to similar existing methods, indicating that our proposed method is highly resistant to conventional, geometric, and combinatorial attacks.

经典的鲁棒水印方法将秘密数据嵌入封面图像中，以保护其版权。然而，它们面临着平衡不可感知性和鲁棒性的问题。为了解决这个问题，研究了传统攻击对从封面图像中提取的特征向量稳定性的影响。据此，我们提出了一种基于Contourlet变换（CT）、chebyshef - Fourier矩（CHFMs）和快速傅里叶变换（FFT）的抗攻击多医学图像零水印方法。首先，对每张医学图像分别进行归一化处理，并采用基于双树复小波变换的方法对归一化后的图像进行融合。其次，提取有效区域并进行CT处理；计算低频子带的CHFMs，对生成的幅值序列进行FFT，构造特征矩阵。将每个特征值的大小与整体均值相结合，生成特征图像。最后，利用Lorenz混沌系统和Fibonacci q矩阵对版权图像进行加密，然后在生成的特征图像与加密后的版权图像之间进行异或运算，产生零水印信号。结果表明，该方法具有良好的抗攻击性能，提取的图像与原始版权图像的归一化相关系数高达0.994。此外，与现有的类似方法相比，我们提出的方法的平均抗攻击性能提高了约2%，这表明我们提出的方法对传统攻击、几何攻击和组合攻击具有很高的抵抗力。

{"title":"Robust zero-watermarking method for multi-medical images based on Chebyshev–Fourier moments and Contourlet-FFT","authors":"Xinhui Lu , Guangyun Yang , Yu Lu , Xiangguang Xiong","doi":"10.1016/j.csi.2025.104115","DOIUrl":"10.1016/j.csi.2025.104115","url":null,"abstract":"<div><div>Classical robust watermarking methods embed secret data into a cover image designed to protect its copyright. However, they suffer from the problem of balancing imperceptibility and robustness. To address this issue, the impact of conventional attacks on the stability of feature vectors extracted from the cover image is examined. Accordingly, we proposed a zero-watermarking method with high attack resistance for multi-medical images by employing Contourlet transform (CT), Chebyshev–Fourier moments (CHFMs), and fast Fourier transform (FFT). First, each medical image is normalized separately, and the normalized images are fused using a dual-tree complex wavelet transform-based method. Second, the effective region is extracted and subjected to the CT. The CHFMs of the low-frequency sub-bands are calculated, and the FFT is performed on the generated amplitude sequence to construct a feature matrix. A feature image is generated by combining the magnitude of each feature value with the overall mean. Finally, the copyrighted image is encrypted using the Lorenz chaotic system and Fibonacci Q-matrix, after which an exclusive-OR operation is applied between the generated feature image and the encrypted copyrighted image to produce a zero-watermarking signal. The results show that the proposed method exhibits excellent resistance to attack with a normalized correlation coefficient of up to 0.994 between the extracted image and the original copyrighted one. Furthermore, the average anti-attack performance of our proposed method is approximately 2% higher compared to similar existing methods, indicating that our proposed method is highly resistant to conventional, geometric, and combinatorial attacks.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104115"},"PeriodicalIF":3.1,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737266","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A novel hybrid WOA–GWO algorithm for multi-objective optimization of energy efficiency and reliability in heterogeneous computing 一种新的混合WOA-GWO算法用于异构计算中能效和可靠性的多目标优化

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-07 DOI: 10.1016/j.csi.2025.104106

Karishma, Harendra Kumar

Heterogeneous computing systems are widely adopted for their capacity to optimize performance and energy efficiency across diverse computational environments. However, most existing task scheduling techniques address either energy reduction or reliability enhancement, rarely achieving both simultaneously. This study proposes a novel hybrid whale optimization algorithm–grey wolf optimizer (WOA–GWO) integrated with dynamic voltage and frequency scaling (DVFS) and an insert-reversed block operation to overcome this dual challenge. The proposed Hybrid WOA–GWO (HWWO) framework enhances task prioritization using the dynamic variant rank heterogeneous earliest-finish-time (DVR-HEFT) approach to ensure efficient processor allocation and reduced computation time. The algorithm’s performance was evaluated on real-world constrained optimization problems from CEC 2020, as well as Fast Fourier Transform (FFT) and Gaussian Elimination (GE) applications. Experimental results demonstrate that HWWO achieves substantial gains in both energy efficiency and reliability, reducing total energy consumption by 55% (from 170.52 to 75.67 units) while increasing system reliability from 0.8804 to 0.9785 compared to state-of-the-art methods such as SASS, EnMODE, sCMAgES, and COLSHADE. The experimental results, implemented on varying tasks and processor counts, further demonstrate that the proposed algorithmic approach outperforms existing state-of-the-art and metaheuristic algorithms by delivering superior energy efficiency, maximizing reliability, minimizing computation time, reducing schedule length ratio (SLR), optimizing the communication-to-computation ratio (CCR), enhancing resource utilization, and minimizing sensitivity analysis. These findings confirm that the proposed model effectively bridges the existing research gap by providing a robust, energy-aware, and reliability-optimized scheduling framework for heterogeneous computing environments.

异构计算系统因其在不同计算环境中优化性能和能源效率的能力而被广泛采用。然而，大多数现有的任务调度技术要么解决能耗降低问题，要么解决可靠性提高问题，很少同时实现这两个目标。为了克服这一双重挑战，本研究提出了一种新的混合鲸鱼优化算法-灰狼优化器（WOA-GWO），该算法集成了动态电压和频率缩放（DVFS）和插入反转块操作。提出的混合WOA-GWO （HWWO）框架利用动态可变等级异构最早完成时间（DVR-HEFT）方法增强任务优先级，以确保高效的处理器分配和减少计算时间。该算法的性能在CEC 2020的实际约束优化问题以及快速傅里叶变换（FFT）和高斯消去（GE）应用中进行了评估。实验结果表明，与SASS、EnMODE、sCMAgES和COLSHADE等最先进的方法相比，HWWO在能源效率和可靠性方面都取得了巨大的进步，将总能耗降低了55%（从170.52单位降低到75.67单位），同时将系统可靠性从0.8804提高到0.9785。在不同任务和处理器数量上的实验结果进一步表明，该算法通过提供卓越的能效、最大限度的可靠性、最小的计算时间、降低调度长度比（SLR）、优化通信与计算比（CCR）、提高资源利用率和最小化灵敏度分析，优于现有的最先进的和元启发式算法。这些发现证实了所提出的模型通过为异构计算环境提供鲁棒性、能量感知和可靠性优化的调度框架，有效地弥补了现有的研究差距。

{"title":"A novel hybrid WOA–GWO algorithm for multi-objective optimization of energy efficiency and reliability in heterogeneous computing","authors":"Karishma, Harendra Kumar","doi":"10.1016/j.csi.2025.104106","DOIUrl":"10.1016/j.csi.2025.104106","url":null,"abstract":"<div><div>Heterogeneous computing systems are widely adopted for their capacity to optimize performance and energy efficiency across diverse computational environments. However, most existing task scheduling techniques address either energy reduction or reliability enhancement, rarely achieving both simultaneously. This study proposes a novel hybrid whale optimization algorithm–grey wolf optimizer (WOA–GWO) integrated with dynamic voltage and frequency scaling (DVFS) and an insert-reversed block operation to overcome this dual challenge. The proposed Hybrid WOA–GWO (HWWO) framework enhances task prioritization using the dynamic variant rank heterogeneous earliest-finish-time (DVR-HEFT) approach to ensure efficient processor allocation and reduced computation time. The algorithm’s performance was evaluated on real-world constrained optimization problems from CEC 2020, as well as Fast Fourier Transform (FFT) and Gaussian Elimination (GE) applications. Experimental results demonstrate that HWWO achieves substantial gains in both energy efficiency and reliability, reducing total energy consumption by 55% (from 170.52 to 75.67 units) while increasing system reliability from 0.8804 to 0.9785 compared to state-of-the-art methods such as SASS, EnMODE, sCMAgES, and COLSHADE. The experimental results, implemented on varying tasks and processor counts, further demonstrate that the proposed algorithmic approach outperforms existing state-of-the-art and metaheuristic algorithms by delivering superior energy efficiency, maximizing reliability, minimizing computation time, reducing schedule length ratio (SLR), optimizing the communication-to-computation ratio (CCR), enhancing resource utilization, and minimizing sensitivity analysis. These findings confirm that the proposed model effectively bridges the existing research gap by providing a robust, energy-aware, and reliability-optimized scheduling framework for heterogeneous computing environments.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104106"},"PeriodicalIF":3.1,"publicationDate":"2025-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Quality assessment for software data validation in automotive industry: A systematic literature review 汽车工业软件数据验证的质量评估：系统的文献综述

IF 3.1 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Computer Standards & Interfaces

Pub Date : 2025-12-04 DOI: 10.1016/j.csi.2025.104110

Gilmar Pagoto , Luiz Eduardo Galvão Martins , Jefferson Seide Molléri

Context

The complexity of automotive systems continues to grow, making software quality assessment crucial for vehicle performance, safety, and cybersecurity.

Objectives

This study explores Quality Assessment (QA) in this context, focusing on its key characteristics, practical implications, and expected deliverables.

Method

We performed a systematic literature review (SLR) by selecting 60 studies from digital libraries.

Results

This SLR highlighted essential QA characteristics that should be incorporated into a software validation phase. Our insights encourage the exploration of advanced techniques, such as Artificial Intelligence (AI), and Machine Learning (ML), to support safety-critical software quality assessments in the automotive domain.

Conclusion

The QA of software data validation requires a holistic approach that combines safety, security, and customer expectations, aligned with industry standards, requirements, and specifications. The relevance of AI and ML in managing complex technologies is evidenced, and the traditional real-world validation dependencies bring risks for safety-critical systems validation.

汽车系统的复杂性持续增长，使得软件质量评估对车辆性能、安全和网络安全至关重要。本研究在此背景下探讨了质量评估（QA），重点关注其关键特征、实际意义和预期可交付成果。方法从数字图书馆中选取60篇文献进行系统文献回顾（SLR）。结果：该单反突出了应纳入软件验证阶段的基本QA特征。我们的见解鼓励探索先进技术，如人工智能（AI）和机器学习（ML），以支持汽车领域安全关键软件质量评估。软件数据验证的QA需要一个整体的方法，将安全、保障和客户期望结合起来，并与行业标准、需求和规范保持一致。人工智能和机器学习在管理复杂技术方面的相关性得到了证明，传统的现实世界验证依赖关系为安全关键系统验证带来了风险。

{"title":"Quality assessment for software data validation in automotive industry: A systematic literature review","authors":"Gilmar Pagoto , Luiz Eduardo Galvão Martins , Jefferson Seide Molléri","doi":"10.1016/j.csi.2025.104110","DOIUrl":"10.1016/j.csi.2025.104110","url":null,"abstract":"<div><h3>Context</h3><div>The complexity of automotive systems continues to grow, making software quality assessment crucial for vehicle performance, safety, and cybersecurity.</div></div><div><h3>Objectives</h3><div>This study explores Quality Assessment (QA) in this context, focusing on its key characteristics, practical implications, and expected deliverables.</div></div><div><h3>Method</h3><div>We performed a systematic literature review (SLR) by selecting 60 studies from digital libraries.</div></div><div><h3>Results</h3><div>This SLR highlighted essential QA characteristics that should be incorporated into a software validation phase. Our insights encourage the exploration of advanced techniques, such as Artificial Intelligence (AI), and Machine Learning (ML), to support safety-critical software quality assessments in the automotive domain.</div></div><div><h3>Conclusion</h3><div>The QA of software data validation requires a holistic approach that combines safety, security, and customer expectations, aligned with industry standards, requirements, and specifications. The relevance of AI and ML in managing complex technologies is evidenced, and the traditional real-world validation dependencies bring risks for safety-critical systems validation.</div></div>","PeriodicalId":50635,"journal":{"name":"Computer Standards & Interfaces","volume":"97 ","pages":"Article 104110"},"PeriodicalIF":3.1,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0