Information Sciences最新文献_第7页

Adversarial Infrared Catmull-Rom Spline: A black-box attack on infrared pedestrian detectors in the physical world 对抗性红外Catmull-Rom样条：物理世界中对红外行人探测器的黑盒攻击

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-07 DOI: 10.1016/j.ins.2025.122263

Zhe Jia , Chengyin Hu , Jiarui Zhang , Kalibinuer Tiliwalidi , Ling Tian , Xian Li , Xu Kang

{"title":"Adversarial Infrared Catmull-Rom Spline: A black-box attack on infrared pedestrian detectors in the physical world","authors":"Zhe Jia , Chengyin Hu , Jiarui Zhang , Kalibinuer Tiliwalidi , Ling Tian , Xian Li , Xu Kang","doi":"10.1016/j.ins.2025.122263","DOIUrl":"10.1016/j.ins.2025.122263","url":null,"abstract":"<div><div>The security of deep neural networks (DNNs) is increasingly threatened by adversarial attacks, yet research in the infrared domain remains limited. Existing white-box attacks, such as light bulb panels and QR code garments, lack stealth and real-world applicability. Additionally, black-box attacks using hot and cold patches exhibit poor robustness. To address these challenges, we introduce AdvICRS, a novel black-box attack method that combines Catmull-Rom Spline curves with Evolutionary Strategy (ES) and the Expectation Over Transformation (EOT) framework to optimize adversarial perturbations. This method generates stealthy and efficient adversarial samples, enabling successful physical attacks using cold patches. Experimental results demonstrate that AdvICRS achieves a 94.9% success rate in digital attacks and 95.2% in physical attacks, outperforming current methods. Stealth analysis confirms that perturbations blend seamlessly with their surroundings, enhancing real-world applicability. Robustness tests show an average success rate of 87.2% across various object detectors, highlighting its adaptability. Ablation studies, generalization evaluations, transfer attacks, and adversarial defense tests further validate AdvICRS's superior performance. These findings not only expose vulnerabilities in infrared detection but also advance adversarial attack strategies in the infrared domain, providing a foundation for future research.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122263"},"PeriodicalIF":8.1,"publicationDate":"2025-05-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143936727","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

HINT: Hypernetwork approach to training weight interval regions in continual learning 提示：连续学习中训练权值区间区域的超网络方法

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-07 DOI: 10.1016/j.ins.2025.122261

Patryk Krukowski , Anna Bielawska , Kamil Książek , Paweł Wawrzyński , Paweł Batorski , Przemysław Spurek

{"title":"HINT: Hypernetwork approach to training weight interval regions in continual learning","authors":"Patryk Krukowski , Anna Bielawska , Kamil Książek , Paweł Wawrzyński , Paweł Batorski , Przemysław Spurek","doi":"10.1016/j.ins.2025.122261","DOIUrl":"10.1016/j.ins.2025.122261","url":null,"abstract":"<div><div>Recently, a new Continual Learning (CL) paradigm was presented to control catastrophic forgetting, called Interval Continual Learning (InterContiNet), which relies on enforcing interval constraints on the neural network parameter space. Unfortunately, InterContiNet training is challenging due to the high dimensionality of the weight space, making intervals difficult to manage. To address this issue, we introduce HINT, a technique that employs interval arithmetic within the embedding space and utilizes a hypernetwork to map these intervals to the parameter space of the target network. We train interval embeddings for consecutive tasks and train a hypernetwork to transform these embeddings into weights of the target network. An embedding for a given task is trained along with the hypernetwork, preserving the response of the target network for the previous task embeddings. Interval arithmetic works with lower-dimensional embedding space rather than directly preparing intervals in a high-dimensional weight space. Furthermore, HINT maintains the guarantee of not forgetting. At the end of the training, we can choose one universal embedding to produce a single network dedicated to all tasks. In such a framework, we can utilize one set of weights. HINT obtains significantly better results than InterContiNet and gives SOTA results on several benchmarks.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122261"},"PeriodicalIF":8.1,"publicationDate":"2025-05-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143936731","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 mirrored echo state network with application to time series prediction 一种用于时间序列预测的镜像回声状态网络

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-06 DOI: 10.1016/j.ins.2025.122260

Xiufang Chen , Liangming Chen , Shuai Li , Long Jin

{"title":"A mirrored echo state network with application to time series prediction","authors":"Xiufang Chen , Liangming Chen , Shuai Li , Long Jin","doi":"10.1016/j.ins.2025.122260","DOIUrl":"10.1016/j.ins.2025.122260","url":null,"abstract":"<div><div>In recent years, the echo state network (ESN) has been increasingly developed and investigated. In this paper, for the first time, a mirrored algorithm is proposed to optimize input weights, and then a mirrored echo state network (MESN) is constructed, where the order of determining weights is exchanged, forming a mirror symmetry with the traditional ESN. Combining the mirrored algorithm and the traditional ESN training method, a novel weight determination scheme is proposed for the MESN, where multiple pseudoinverse processes are involved and utilized, and then the optimal input weights and retrained output weights are acquired. To meet the echo state property, the reservoir connection weights are determined with the assistance of the singular value decomposition. Moreover, the stepwise incremental method and the achievements of predecessors are combined and used, based on which the structure of the reservoir is determined. Finally, experiments on the Mackey-Glass system (MGS), as well as two real-world datasets, along with comparisons with existing works, are conducted, and the results demonstrate the superiority and stability of the proposed MESN in predicting MGS with large chaotic factors and more complex real-world problems.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"716 ","pages":"Article 122260"},"PeriodicalIF":8.1,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143923256","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 hybrid framework for spatio-temporal traffic flow prediction with multi-scale feature extraction 基于多尺度特征提取的交通流时空预测混合框架

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-06 DOI: 10.1016/j.ins.2025.122259

Ang Ji , Zhuo Liu , Lingyun Su , Zhe Dai

{"title":"A hybrid framework for spatio-temporal traffic flow prediction with multi-scale feature extraction","authors":"Ang Ji , Zhuo Liu , Lingyun Su , Zhe Dai","doi":"10.1016/j.ins.2025.122259","DOIUrl":"10.1016/j.ins.2025.122259","url":null,"abstract":"<div><div>Efficient and accurate traffic flow prediction has become increasingly crucial with the advancement of intelligent transportation systems. This paper proposes a hybrid framework that combines depthwise separable convolutions and Transformer modules to learn spatio-temporal dependencies in traffic flow data. First, multi-scale features are extracted by depthwise separable convolutions, which decompose the convolution operation into independent spatial and temporal dimensions. This approach aims to reduce computational costs and effectively capture complex local spatio-temporal flow patterns in road networks. By adopting hierarchical processing, the model can learn dynamics across various scenarios and adapt to diverse traffic flow conditions. Then, we integrate a Transformer module into the model, leveraging its self-attention mechanism to capture the global patterns within traffic data. The integrated Transformer learns long-range dependencies across different road sections, which is particularly beneficial in road networks with complex interaction effects. Experiments on multiple real-world traffic datasets demonstrate that the proposed model outperforms traditional methods in both prediction accuracy and computational efficiency. The integration of depthwise separable convolutions and Transformer-based modeling exhibits superior performance in traffic flow prediction, providing a sufficient tool for urban traffic management.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"716 ","pages":"Article 122259"},"PeriodicalIF":8.1,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143923255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Observer-based event-triggered impulsive synchronization control of heterogeneous complex networks with bit-rate constraints 基于观测器的具有比特率约束的异构复杂网络事件触发脉冲同步控制

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-05 DOI: 10.1016/j.ins.2025.122268

Jing Guo , Ling Huang , Peng Shi

{"title":"Observer-based event-triggered impulsive synchronization control of heterogeneous complex networks with bit-rate constraints","authors":"Jing Guo , Ling Huang , Peng Shi","doi":"10.1016/j.ins.2025.122268","DOIUrl":"10.1016/j.ins.2025.122268","url":null,"abstract":"<div><div>This paper proposes an observer-based event-triggered impulsive control synchronization for heterogeneous complex networks under bit-rate constraints. An isolated target node and observer systems are introduced to derive synchronization and observation error dynamics. A dynamic quantization-based encoding–decoding strategy is designed for wireless communication with bit-rate constraints, and the relationship between encoded error and bit rate is established. To conserve resources, a new event-triggered mechanism based on the sampling period is proposed. This mechanism reduces unnecessary triggers while ensuring effective impulsive control through a maximum triggering interval. Identical decoder models on both sensor and controller sides eliminate additional communication for control signal transmission. An impulsive controller is designed using decoded signals at triggering moments, and a new framework for closed-loop augmented error systems is constructed. Sufficient conditions for bounded synchronization are derived using hybrid systems theory and Lyapunov stability theory. The relationship among heterogeneity, bit rate, and synchronization error is demonstrated. An optimization algorithm is proposed to minimize a weighted sum of the upper bound of the augmented errors and the maximum triggering interval. Numerical simulations verify the effectiveness of the proposed approaches.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122268"},"PeriodicalIF":8.1,"publicationDate":"2025-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143949051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Distributed state estimation for cyber-physical systems against sensor attacks based on Isolation Forest 基于隔离森林的网络物理系统抗传感器攻击的分布式状态估计

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-05 DOI: 10.1016/j.ins.2025.122258

Yongbao Xiao , Xiaojian Li

引用次数: 0

A robust temporal multivariable fault detection method for blast furnace: Robust temporal convolution detection network 高炉多变量故障鲁棒时域检测方法：鲁棒时域卷积检测网络

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-04 DOI: 10.1016/j.ins.2025.122269

Xiongzhuo Zhu, Chunjie Yang, Siwei Lou, Yuelin Yang

{"title":"A robust temporal multivariable fault detection method for blast furnace: Robust temporal convolution detection network","authors":"Xiongzhuo Zhu, Chunjie Yang, Siwei Lou, Yuelin Yang","doi":"10.1016/j.ins.2025.122269","DOIUrl":"10.1016/j.ins.2025.122269","url":null,"abstract":"<div><div>As a complex industrial process, the blast furnace ironmaking process (BFIP) often suffers from faults due to changes in raw materials, variables’ setpoints, reaction conditions, etc. However, because of the nonlinearity, dynamics, and mixture of normal and abnormal outliers in BFIP, the existing methods always meet difficulties in practical application. Therefore, this paper proposes a robust temporal convolution detection network (RTCDN) for BFIP fault detection. The basic TCDN network is constructed by stacking the residual blocks of the temporal convolution network (TCN). The residual block consists of the 1-D dilated causal convolution, which is performed on the time scale of variables, giving the network the ability to extract time-scale information. Since TCDN consists of convolutional networks, it balances the extraction of temporal information and speed compared to RNN-type methods. Furthermore, a robust solution is proposed to overcome the interference of abnormal outliers. The robust solution exploits that TCDN can’t reconstruct abnormal outliers and decomposes the training dataset into the clean part and abnormal outliers. In this work, the proximal approach to optimizing the sparse outlier matrix is also improved to achieve complete separation of abnormal outliers. Finally, TCDN is applied to three BF fault datasets and performs better than the conventional temporal detection methods. RTCDN has also been proven to have good outlier separation capability, maintaining satisfactory detection performance when the proportion of abnormal outliers is less than 20%.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"716 ","pages":"Article 122269"},"PeriodicalIF":8.1,"publicationDate":"2025-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143923254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Resilient information and inference networks under mixed-trust sensing 混合信任感知下的弹性信息与推理网络

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-02 DOI: 10.1016/j.ins.2025.122256

Humberto E. Garcia , Dimitrios Pylorof , Wen-Chiao Lin

{"title":"Resilient information and inference networks under mixed-trust sensing","authors":"Humberto E. Garcia , Dimitrios Pylorof , Wen-Chiao Lin","doi":"10.1016/j.ins.2025.122256","DOIUrl":"10.1016/j.ins.2025.122256","url":null,"abstract":"<div><div>With ubiquitous digitization, sensing, and computational intelligence deployed in increasingly more and broader domains, including critical infrastructure, potentially misleading and destabilizing effects of multimodal anomalies and adversarial behavior are growing in importance. We develop randomized and reinforcement learning-based strategies for strategically recruiting and utilizing deployed (and, thus, vulnerable and potentially faulty and/or compromised) nodes from information and inference networks, while defending against adversaries that attempt to misguide assessments of inferred variables. Recognizing that, besides communication and other costs, sampling from any observable node can either provide true data or dangerously expose our inference to misinformation (without being easily distinguishable what actually happens), the proposed strategies proceed by progressively recruiting nodes and cautiously scaling their information contribution based on assumed, or, in our reinforcement learning approach, intelligently weighed trustworthiness, with the learning approach also considering network-wide, threat-inclusive risk/value tradeoffs. While avoiding the hardware, communication, analytical and computational burden of explicit redundancy, the proposed defensive schemes enable on-the-fly assessments of underlying processes, and system-wide situational awareness with demonstrable resilience against adversarial activities.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"716 ","pages":"Article 122256"},"PeriodicalIF":8.1,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143905841","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

Towards self-interpretable review spammer group detection 面向自解释审查垃圾邮件组检测

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-02 DOI: 10.1016/j.ins.2025.122257

Chenghang Huo , Fuzhi Zhang

{"title":"Towards self-interpretable review spammer group detection","authors":"Chenghang Huo , Fuzhi Zhang","doi":"10.1016/j.ins.2025.122257","DOIUrl":"10.1016/j.ins.2025.122257","url":null,"abstract":"<div><div>Due to data imbalance, existing spammer group detection methods often yield suboptimal performance. Moreover, many of these approaches operate as black boxes, offering little to no interpretability for their detection results. To address these issues, we propose a self-interpretable approach for detecting review spammer groups. Our method begins by constructing a user-product bipartite graph with edge attributes. We integrate user review information with a novel fitness function to develop an adaptive genetic algorithm that effectively identifies high-quality candidate groups. Next, we introduce a hybrid graph neural network enhanced with active learning to generate vector representations of these candidate groups. We then design and construct a prototype layer and a group classification layer to detect spammer groups accurately. To provide interpretability, we incorporate prototype learning to create an interpretation mechanism that explains detection outcomes. Experimental results demonstrate that our method achieves substantial improvements in Precision@k and Recall@k at the top-1000 ranking, outperforming state-of-the-art solutions on the Amazon, YelpChi, YelpNYC, and YelpZip datasets by [11.53 %, 97.36 %, 51.37 %, 32.06 %] and [12.65 %, 54.65 %, 64.19 %, 47.37 %], respectively. Additionally, the Fidelity of our interpretability results under varying Sparsity levels is approximately 4 %, 9 %, 8 %, and 8 % higher than those of existing methods on the same datasets.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"716 ","pages":"Article 122257"},"PeriodicalIF":8.1,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143923251","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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

CCRPS: Customized cross-domain routing with privacy preservation and stable quality-of-experience based on deep reinforcement learning CCRPS：基于深度强化学习的具有隐私保护和稳定体验质量的定制跨域路由

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-02 DOI: 10.1016/j.ins.2025.122255

Jialei Zhang , Zheng Yan , Haiguang Wang , Tieyan Li

{"title":"CCRPS: Customized cross-domain routing with privacy preservation and stable quality-of-experience based on deep reinforcement learning","authors":"Jialei Zhang , Zheng Yan , Haiguang Wang , Tieyan Li","doi":"10.1016/j.ins.2025.122255","DOIUrl":"10.1016/j.ins.2025.122255","url":null,"abstract":"<div><div>Next-generation networks are predominantly heterogeneous, integrating diverse network domains built on various technologies. The transmission of data with specific requirements across multiple network domains necessitates advanced cross-domain routing solutions. However, current approaches fall short in providing cross-domain customized routing that incorporates privacy protection and adapts to dynamic network conditions, often overlooking Quality of Experience (QoE) and its stability. To tackle these challenges, we propose CCRPS, a customized cross-domain routing scheme designed for Integrated Heterogeneous Networks (Inte-HetNet), which enables routing customization, supports dynamic network environments, ensures robust cross-domain privacy protection, and delivers consistent and efficient QoE. Specifically, CCRPS begins by formalizing user customization requirements to facilitate routing customization. Next, it reformulates the cross-domain routing generation problem as a multi-agent Deep Reinforcement Learning (DRL) task and develops a Customized Cross-domain Routing algorithm based on Multi-agent DRL (CCR-MD) to address it, ensuring adaptability to dynamic network conditions. Additionally, CCRPS incorporates privacy protection mechanisms, such as virtual topology construction, node attribute calculation, and random obfuscation, to safeguard privacy during cross-domain routing. Moreover, it introduces a QoE-centric reward function to maintain QoE stability. Extensive experimental evaluations demonstrate the superior performance of CCRPS through comparison with existing related schemes.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"716 ","pages":"Article 122255"},"PeriodicalIF":8.1,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143904498","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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