Applied Intelligence最新文献

{"title":"Enhancing adversarial robustness in power quality classification systems: an attention-based defense framework","authors":"Mubarak Alanazi,&nbsp;Nasser S. Alkhaldi","doi":"10.1007/s10489-025-06865-9","DOIUrl":"10.1007/s10489-025-06865-9","url":null,"abstract":"<div><p>Power quality monitoring is essential for ensuring the reliability, stability, and security of modern electrical networks. While deep learning models have demonstrated exceptional performance in classifying power quality disturbances, they remain critically vulnerable to adversarial perturbations—posing significant risks to smart grid cybersecurity. This paper introduces three novel contributions to the field of power quality cybersecurity: (1) Signal-Agnostic Adversarial (SAA) attacks—a perturbation method tailored specifically for power quality signals; (2) an attention-based convolutional neural network (CNN) architecture that consistently achieves 5–7% points higher robustness under attack compared to conventional models; and (3) comprehensive vulnerability fingerprinting, which exposes architecture-specific adversarial attack patterns and provides insights into structural weaknesses. We conduct a systematic analysis of CNN-based power quality classification models subjected to adversarial manipulations and propose effective defense strategies. Three attack methodologies are introduced and evaluated: the Fast Gradient Sign Method (FGSM), Signal-Specific Adversarial (SSA) attacks, and the proposed SAA attacks. Experimental results reveal catastrophic degradation in model performance, with accuracy reductions of up to 80–90% points under attack. To mitigate these vulnerabilities, our attention-based CNN model demonstrates significantly improved resilience, and adversarial training further enhances robustness—achieving up to 58.47% accuracy against SSA, the most potent attack vector. The findings underscore critical security implications of deep learning in power systems and offer practical mitigation strategies for enhancing robustness in real-world smart grid deployments.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145210386","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}

{"title":"CLDM-Palm: A controllable latent diffusion model for high-fidelity palmprint generation based on Bézier curves","authors":"Yuanpan Zhu,&nbsp;Donghuai Jia,&nbsp;Kevin Chu,&nbsp;Wenshuang Zhi,&nbsp;Weide Li,&nbsp;Shukai Chen","doi":"10.1007/s10489-025-06923-2","DOIUrl":"10.1007/s10489-025-06923-2","url":null,"abstract":"<div><p>Using limited real data to synthesize realistic palmprints and expand training samples for recognition models has become a promising direction in palmprint recognition. However, the pseudo-palmprints generated by existing models still exhibit significant discrepancies from real ones, particularly in crease structures and fine-grained details. In this paper, we first introduce Latent Diffusion Models (LDM) as the backbone to improve the quality of palmprint generation. Secondly, to incorporate Bézier curves as control conditions into the model, we propose the Palm-to-Bézier Module (P2BM), which maps real palmprints to their corresponding Bézier-style pseudo-Bézier curves. These curves establish the connection between real palmprints and Bézier curves, which are used as conditional inputs during diffusion model training. At inference time, Bézier curves can be provided as conditions to generate high-resolution, fine-grained, and highly realistic synthetic palmprints. Thirdly, to enable Bézier curves to better model palmprint creases, we propose 12 Bézier curves templates based on real crease distribution priors. With only 10-step Denoising Diffusion Implicit Models (DDIM) sampling, our method achieves a significantly lower Fréchet Inception Distance (FID) compared to existing palmprint generation approaches. Moreover, the recognition models trained on the synthetic palmprints generated by our model achieve new state-of-the-art results in both Fisher Discriminant Ratio (FDR) and <i>TAR@FAR</i> metrics. Under a 1:1 train-test fine-tuning setting, our model improves average <i>TAR@FAR</i>=<span>(10^{-6})</span> performance by over <span>(10%)</span> compared to prior methods. We name our model CLDM-Palm (Controllable Latent Diffusion Model-Palm).</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145210805","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}

{"title":"Three-way clustering ensemble based on shadowed sets with five approximation regions","authors":"Huangjian Yi,&nbsp;Dongkai Guo,&nbsp;Qinran Zhang,&nbsp;Xiaowei He,&nbsp;Ruisi Ren","doi":"10.1007/s10489-025-06726-5","DOIUrl":"10.1007/s10489-025-06726-5","url":null,"abstract":"<div><p>Clustering ensemble is a powerful technique for aggregating multiple clustering results. In order to address the challenge in clustering analysis which was brought by the uncertainty information in the datasets, this work presents a novel three-way clustering ensemble method based on shadowed sets with five approximation regions (3WCE-S5). Firstly, a set of clustering members are generated by fuzzy c-means clustering (FCM). A new shadowed sets is approximated by five regions, named as shadowed sets with five approximation regions (S5). Then, all objects are initially partitioned into five regions according to their membership degrees, which are provided by FCM. Secondly, according to multi-granularity rough sets, objects are further assigned into six approximated regions, namely a core region and five fringe regions. There is a partial order relationship between these six different approximate regions. Finally, the above six regions are processed by the new shadowed sets again to generate the output of three-way clustering. Ten University of California Irvine (UCI) data sets are employed to test the performance of this approach and five comparative methods. Accuracy (ACC), adjusted rand index (ARI), normalized mutual information (NMI) and time cost are utilized to quantify the clustering results.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145169697","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}

{"title":"Semantic interaction-enhanced encoding network for math word problem solving","authors":"Lingsheng Xiao,&nbsp;Yuzhong Chen,&nbsp;Zhanghui Liu,&nbsp;Jiayuan Zhong,&nbsp;Yu Dong","doi":"10.1007/s10489-025-06850-2","DOIUrl":"10.1007/s10489-025-06850-2","url":null,"abstract":"<div><p>Solving math word problems (MWPs) requires machines to understand not only the literal meaning of text but also the abstract logic and mathematical reasoning embedded within it. However, existing models often lack explicit reasoning capabilities for semantic information, particularly when dealing with complex math word problem texts. Additionally, these models tend to embed all kinds of information without fine-grained selection, which may introduce unexpected noise for mathematical expression generation. To address these challenges, we propose a Semantic Interaction-Enhanced Encoding Network (SIEN) for math expression generation is proposed in this paper. Firstly, SIEN constructs a semantic role interaction graph for each problem and employs a graph attention neural network to learn interaction and semantic information, offering a more structured and enriched view of the math word problem text. Secondly, SIEN introduces a multi-channel adapter module that simultaneously learns comprehensive contextual information from numeric information channel, hierarchical semantic information channel, and interaction information channel. Furthermore, SIEN introduces a dynamic weighting mechanism that adjusts the information weight from each channel to prioritize relevant information and reduce noise. Experimental results on three public benchmark datasets demonstrate that SIEN achieves significant performance improvement over other state-of-the-art baseline models.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145169174","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}

{"title":"Rapid federated unlearning with tuning parameters based on fisher information matrix","authors":"Fengda Zhao,&nbsp;Qianyi Xu,&nbsp;Hao Wang,&nbsp;Dingding Guo","doi":"10.1007/s10489-025-06593-0","DOIUrl":"10.1007/s10489-025-06593-0","url":null,"abstract":"<div><p>Federated learning is a distributed machine learning approach widely applied in privacy-sensitive scenarios. With the emergence of the “right to be forgotten” and the pursuit of data accuracy, there is an increasing demand to quickly and accurately delete targeted information from models while ensuring model performance. Therefore, federated unlearning has been introduced. Although current federated unlearning methods achieve effective unlearning, they often involve lengthy processes and require servers to store extensive historical update information. We propose a novel rapid federated unlearning method named FedTune. This method leverages the Fisher information matrix computed on the client side to assess the correlation between model parameters and the target data, identifying key parameters for adjustment. Based on the importance of these parameters and the frequency of client participation, FedTune determines appropriate adjustment ratios to increase the classification loss on the target data, thereby reducing the model’s accuracy and achieving effective data unlearning. Finally, the server collaborates with the remaining clients for a few rounds of retraining to restore the overall classification performance rapidly. We evaluated the FedTune method on the MNIST, CIFAR-10, and PURCHASE datasets, considering both fixed and dynamic client selection scenarios in privacy-sensitive and contamination settings. Experimental results show that FedTune reduces the time consumed by the unlearning process and server storage costs of the unlearning algorithm while ensuring model classification accuracy and effective unlearning compared to other unlearning algorithms.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145169699","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}

{"title":"A large-scale group decision-making approach for quality function deployment based on Dempster-Shafer evidence theory and hierarchical clustering algorithm","authors":"Zhengmin Liu,&nbsp;Xuan Feng,&nbsp;Jihao Zhang,&nbsp;Bo Zhang,&nbsp;Wenxin Wang,&nbsp;Peide Liu","doi":"10.1007/s10489-025-06724-7","DOIUrl":"10.1007/s10489-025-06724-7","url":null,"abstract":"<div><p>Quality Function Deployment (QFD) is a classic customer requirements (CRs)-oriented quality management method. However, the increasing complexity and diversity of CRs in the modern society makes it impossible for the traditional QFD approach with a limited number of team members (TMs) to fully satisfy CRs. Therefore, in order to solve the QFD problem in complex environments, this paper proposes an improved QFD method based on Dempster–Shafer evidence theory (D-S theory) and hierarchical clustering algorithm in large-scale group environments. Firstly, utilizing the advantages of D-S theory in information processing and synthesis, the evaluation of quality characteristics (QCs) in the form of probabilistic linguistic term sets (PLTSs) is transformed into basic probability assignments (BPAs) to handle uncertainty more flexibly. Secondly, this paper designs a hierarchical clustering algorithm based on bounded confidence to divide TMs into subgroups, and fully considers the interaction willingness of TMs during the clustering process to ensure the efficiency and accuracy of decision-making. On this basis, the Stepwise Weight Assessment Ratio Analysis (SWARA) method based on distance degree is introduced to calculate the weight of CRs in a more objective way. Then, the Decision-making Trial and Evaluation Laboratory (DEMATEL) method based on D-S theory is used to deeply analyze the mutual influence relationship between QCs to reveal its internal logic. Besides, combined with the psychological expectations of TMs, the disappointment theory is used to prioritize QCs to ensure that products or services are more in line with customer expectations. Finally, this paper applies the proposed method to the development process of mobile health applications (mHealth apps) from the perspective of privacy security, verifying the practicability and superiority of the method. The effectiveness of the method in CRs transformation and product design optimization is further demonstrated through parametric and comparative analyses.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145145202","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}

{"title":"KRMNet: learning core representations for partial discharge pattern recognition via masked autoencoders and mixed position coding","authors":"Yi Deng,&nbsp;Jiawen Chen,&nbsp;Quan Xie,&nbsp;Dapeng Tan,&nbsp;Hai Liu","doi":"10.1007/s10489-025-06899-z","DOIUrl":"10.1007/s10489-025-06899-z","url":null,"abstract":"<div><p>Partial discharge pattern recognition (PDPR) is a crucial cornerstone for condition monitoring and safe operation of electrical devices. It has become an important hotspot in the field of energy systems. However, it faces several challenges, including noise interference, signal complexity, and difficulty in data labeling. This study proposes an efficient multi-scale masked autoencoder (MAE) network (KRMNet) to effectively address these challenges. KRMNet learns common and important multi-scale features and long-range semantic dependencies of partial discharge signals. Furthermore, by using the MAE structure and the transformer as the backbone, the model extracts key distinguishing features from phase-resolved partial discharge (PRPD) signals with background noise, interference, and labeling issues in an efficient and self-supervised manner. In addition, the efficient multi-scale module uses an efficient multi-scale attention mechanism to aggregate key information from multiple feature dimensions. The integration of the efficient multi-scale attention mechanism and contrastive learning methods improves the model’s ability to distinguish key information and resist interference. Experiments on two challenging PRPD datasets show that our proposed KRMNet achieves detection accuracies of 88.5% and 90.2% on noisy (ECPD) and clean (PUPD) datasets, respectively. This finding suggests that the method faces challenges in effectively managing noise interferences and missing labels.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145169698","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}

{"title":"External information-augmented contrastive learning framework for fake news detection","authors":"Xiaochang Fang,&nbsp;Huaxiang Zhang,&nbsp;Hongchen Wu,&nbsp;Li Liu,&nbsp;Hongzhu Yu,&nbsp;Hongxuan Li,&nbsp;Zhaorong Jing","doi":"10.1007/s10489-025-06807-5","DOIUrl":"10.1007/s10489-025-06807-5","url":null,"abstract":"<div><p>The proliferation of fake news and information overload on social media has led to increased public confusion and poses a serious threat to social stability. Traditional fake news detection methods typically focus solely on the content of the news itself, making them vulnerable to manipulation by disinformation campaigns. This limitation highlights the need for a more comprehensive approach that incorporates external information to improve detection accuracy. In response to this challenge, we propose a novel framework for fake news detection, named External Information-Augmented Contrastive Learning (EACL). The EACL framework consists of three key modules: (1) the External Information Construction Module, which utilizes entity linking, embedding, and retrieval techniques to analyze news from both factual and public opinion perspectives, thus creating an analysis-friendly environment; (2) the Consistency Feature Extraction Module, which employs a distance-aware signed attention mechanism to model the consistency between news content and external information, while filtering out irrelevant data; and (3) the Comparative Learning Enhancement Module, which constructs positive and negative sample pairs to enhance the learning of semantic differences between fake and real news. Extensive qualitative and quantitative experiments conducted on two real-world datasets demonstrate that EACL achieves impressive accuracy rates of 85.2% and 82.9%, significantly outperforming existing baseline methods. The results further illustrate the effectiveness of integrating external information and contrastive learning in combating misinformation.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145168490","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}

{"title":"Multi-stimulus generalized and corrected canonical correlation analysis for enhancing SSVEP detection","authors":"Yanhao Lv,&nbsp;Tian-jian Luo","doi":"10.1007/s10489-025-06859-7","DOIUrl":"10.1007/s10489-025-06859-7","url":null,"abstract":"<div><p>Spatial filter-based calibration-training algorithms play a crucial role in improving the information transfer rate (ITR) of steady-state visual evoked potential based brain-computer interfaces (SSVEP-BCIs). These algorithms optimize spatial filters by suppressing the non-SSVEP related components, thereby enhancing the signal-to-noise ratio (SNR) of electroencephalogram (EEG) signals. However, conventional methods neglect the temporally-varying and spatially-coupled characteristics of EEG signals, leading to inherent ITR bottlenecks in BCIs. To this end, we propose a novel SSVEP detection algorithm, termed as <b>m</b>ulti-<b>s</b>timulus <b>G</b>eneralized and <b>C</b>orrected <b>C</b>anonical <b>C</b>orrelation <b>A</b>nalysis (msGC³A), which is extended and corrected from the generalized canonical correlation analysis algorithm. Specifically, we develop corrected sine-cosine reference templates that enhance the spatial filters’ generalization capability across multiple stimuli. Moreover, we formulate a weighted correlation coefficient that synergistically integrates both generalized and corrected multi-stimulus templates for further enhancement. Empirical experiments have been conducted on two publicly available benchmark SSVEP datasets, and we compared the ensemble version of our msGC³A algorithm with four state-of-the-art algorithms. The results have shown that our algorithm significantly improves SSVEP detection performance while requiring less calibration data. Furthermore, we also conducted ablation experiments to show the adaptive capacity of employing our algorithm for SSVEP-BCIs.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145168489","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}

{"title":"Robust low-rank representation with structured similarity learning for multi-label classification","authors":"Emmanuel Ntaye,&nbsp;Conghua Zhou,&nbsp;Zhifeng Liu,&nbsp;Heping Song,&nbsp;Fadilul-lah Yassaanah Issahaku,&nbsp;Xiang-Jun Shen","doi":"10.1007/s10489-025-06879-3","DOIUrl":"10.1007/s10489-025-06879-3","url":null,"abstract":"<div><p>Handling high-dimensional, noisy data in multi-label classification is challenging, as feature abundance and noise obscure actual data-label relationships. Traditional approaches often model labels and features independently, limiting dependency modeling and noise reduction. To address this, we propose a unified framework combining low-rank representation using nuclear norm regularization with structured similarity learning. This simultaneously projects features and labels into low-rank spaces while preserving key inter-sample and inter-label relationships through structural constraints, further capturing fine-grained correlations via a learned similarity Matrix. Extensive experiments on five benchmark datasets show our model outperforms state-of-the-art methods, achieving a 16% reduction in Hamming Lossl and a 14% improvement in Micro-F1 on high-dimensional, noisy datasets like CAL500 and Corel16k7, with consistent gains in Macro-F1 and Example-F1. These results demonstrate the model’s strong capability for noisy, high-dimensional multi-label classification.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 15","pages":""},"PeriodicalIF":3.5,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145100654","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}