Knowledge-Based Systems最新文献

{"title":"DGFFA: Joint multimodal entity-relation extraction via dual-channel graph fusion and fine-grained alignment","authors":"Wenjie Liu ,&nbsp;Xingwen Li ,&nbsp;Zhijie Ren","doi":"10.1016/j.knosys.2026.115470","DOIUrl":"10.1016/j.knosys.2026.115470","url":null,"abstract":"<div><div>Joint multimodal entity and relation extraction (JMERE) is a key task in multimodal knowledge graph completion (MKGC), aimed at integrating textual and visual information for better knowledge representation and semantic reasoning. However, existing paradigms often struggle with suboptimal cross-modal alignment and typically neglect the intrinsic correlations between entities and relations within word-pair structures. To tackle these challenges, we propose a JMERE framework via Dual-Channel Graph Fusion and Fine-Grained Alignment, namely DGFFA. Specifically, a fine-grained cross-modal alignment module is designed, which leverages token-patch similarity priors from a pre-trained vision-language model to guide optimal-transport matching, which suppresses noisy visual regions and yields more precise multimodal correspondences. To fully leverage the connections between entities and relationships, a dual-channel graph architecture was designed to jointly optimize the representations of nodes and edges in a unified prediction space, thereby effectively modeling bidirectional dependencies. Extensive experiments demonstrate that our model consistently outperforms state-of-the-art methods such as EEGA and TESGA, achieving average improvements of 2.4%, 3.2%, and 1.6% in Precision, Recall, and F1 on JMERE tasks. Our approach not only offers a new paradigm for multimodal entity-relation extraction, but also contributes novel insights into multimodal knowledge graph construction and unified multimodal reasoning.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115470"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174779","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}

{"title":"FedPDM: Representation enhanced federated learning with privacy preserving diffusion models","authors":"Wei Guo ,&nbsp;Fuzhen Zhuang ,&nbsp;Yiqi Tong ,&nbsp;Xiao Zhang ,&nbsp;Zhaojun Hu ,&nbsp;Jiejie Zhao ,&nbsp;Jin Dong","doi":"10.1016/j.knosys.2026.115452","DOIUrl":"10.1016/j.knosys.2026.115452","url":null,"abstract":"<div><div>Most existing semi-parameter-sharing federated learning (FL) frameworks utilize generative models to achieve partial parameter sharing with the server, which effectively enhances the data privacy of each client. However, these generative models often suffer from model utility degradation due to <em>poor representation robustness</em>. Meanwhile, <em>representation inconsistency</em> between local and global models exacerbates the client drift problem under non-IID scenarios. Furthermore, existing semi-parameter-sharing FL frameworks overlook <em>representation leakage</em> risks associated with generator sharing, while failing to balance privacy and utility. To alleviate these challenges, we propose FedPDM, a semi-parameter-sharing FL framework built upon a privacy-preserving diffusion model (PDM). Specifically, our proposed PDM enables model alignment with features from the privacy extractor without requiring direct exposure of this extractor, effectively mitigating utility degradation caused by poor representation robustness. Moreover, a feature-level penalty term is introduced into the optimization objective of PDM to avoid representation leakage. We further design a two-stage aggregation strategy that addresses representation inconsistency through initialization correction with a Gaussian constraint for knowledge distillation. Finally, we provide the first theoretical convergence analysis for semi-parameter-sharing FL, demonstrating that our framework converges at a rate of <span><math><mrow><mi>O</mi><mo>(</mo><mn>1</mn><mo>/</mo><mi>T</mi><mo>)</mo></mrow></math></span>. Extensive experiments on four datasets show that FedPDM achieves average accuracy improvements of 1.78% to 5.56% compared with various state-of-the-art baselines.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115452"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174781","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}

{"title":"FasterGCN: Accelerating and enhancing graph convolutional network for recommendation","authors":"Jiaxin Wu ,&nbsp;Chenglong Pang ,&nbsp;Guangxiong Chen ,&nbsp;Jie Zhao ,&nbsp;Jihong Wan","doi":"10.1016/j.knosys.2026.115533","DOIUrl":"10.1016/j.knosys.2026.115533","url":null,"abstract":"<div><div>In recommendation systems, graph convolutional networks (GCNs) are widely used to capture high-order user–item interactions. However, deeper GCNs often suffer from over-smoothing, where node representations become indistinguishable. Conversely, excessive efforts to avoid over-smoothing can lead to under-smoothing, resulting in prolonged training and insufficient aggregation of neighborhood information. To address both issues, we propose FasterGCN, a linear GCN specifically designed for recommendation tasks. By distilling potential interaction information (PII) from high-order connectivity, FasterGCN achieves optimal smoothing rapidly and maintains stable performance even with deeper architectures. Moreover, its concise design eliminates complex parameters, enhancing scalability and extensibility. Extensive experiments on six real-world datasets demonstrate that FasterGCN not only consistently outperforms state-of-the-art GCNs but also improves training efficiency by up to 96%, highlighting its potential as a strong backbone for recommender systems.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115533"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175151","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}

{"title":"HiSURF: Hierarchical semantic-guided unified radiance field for generalizing across unseen scenes","authors":"Qiang Liu ,&nbsp;Teng Wang ,&nbsp;Zhiguo Zhang ,&nbsp;Jun Nie ,&nbsp;Xiao Lu ,&nbsp;Chunyang Sheng ,&nbsp;Shibin Song ,&nbsp;Qiaoqiao Sun ,&nbsp;Haixia Wang","doi":"10.1016/j.knosys.2026.115530","DOIUrl":"10.1016/j.knosys.2026.115530","url":null,"abstract":"<div><div>Recent advancements in neural field representations have significantly improved novel view synthesis for seen scenes. However, generalizing seen representations to unseen scenes remains challenging. Addressing this problem, we propose the Hierarchical Semantic-guided Unified Radiance Field (HiSURF) to leverage hierarchical semantic attributes from seen scenes as prior knowledge. The synthesis of scene representations for unseen environments can be enabled by establishing an interpretable mapping between semantic attributes and visual features. Specifically, HiSURF consists of a local semantic embedding module, a global semantic mapping module, and a composite rendering module. For a scene with multiple objects, the local module disentangles attributes of objects to generate fine object-level triplanes, which preserve structural and surface details for objects. At the same time, the global module utilizes attributes of the holistic scene to construct a coarse scene-level triplane, which ensures layout consistency and contextual coherence for the scene. Then, the composite rendering module integrates features from both object-level and scene-level triplanes for high-quality novel view synthesis. Experimental results on the ClevrTex and Kubric datasets demonstrate that our HiSURF not only outperforms existing approaches in novel view synthesis but also exhibits superior generalization capability to unseen scenes.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115530"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175149","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}

{"title":"Face and gait based authentication using similarity-optimized bidirectional recurrent neural transformer model","authors":"Sugantha Priyadharshini P ,&nbsp;Grace Selvarani A","doi":"10.1016/j.knosys.2026.115445","DOIUrl":"10.1016/j.knosys.2026.115445","url":null,"abstract":"<div><div>Biometric recognition is a necessary task in security control systems, yet unimodal techniques often suffer from missing modalities, noise and limited robustness. Thus, the proposed research introduced a multi-modal biometric identification system by integrating both face and gait based images. The key frames are selected by using an enhanced agglomerative nesting clustering algorithm (EAg-NCA) to preserve various information with minimal redundancy in the input video. Noise in the selected key frames is removed by using the Trimmed Pixel density based median filter (TPDMF). Then, from the pre-processed images, faces are detected using a Dung beetle optimization tuned YOLO-V9, while gait silhouettes are extracted by Mask Region based Convolutional Neural Network (Mask-RCNN). The features are extracted from the face, and the gait image is accomplished through a pooled convolutional dense net model (PoC-Den). The extracted features are fused, and a decision has been made regarding the authentication of a person by matching the current features with the features in the database by using a novel similarity based optimized hybrid bidirectional recurrent neural pooling transformer encoder block (Sim-OpPTr). Finally, get the classified result as the person is authorized or unauthorized. The results are evaluated by using various performance metrics, proposed methodology obtained an accuracy of 99.08%. The proposed hybrid strategy improves multi-modal fusion, robustness to noise, and authentication accuracy, making it suitable for real-world surveillance applications.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115445"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175147","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}

{"title":"PLeFF-Net: parallel Le-Net forward fractional network for sarcoma cancer detection using histopathological image","authors":"Balashanmuga Vadivu Palanivel ,&nbsp;Gopalsamy Venkadakrishnan Sriramakrishnan ,&nbsp;Vadamodula Prasad ,&nbsp;Subbiah Vairamuthu ,&nbsp;Deena Gnanasekaran ,&nbsp;Ilavarasan Sargunan","doi":"10.1016/j.knosys.2026.115476","DOIUrl":"10.1016/j.knosys.2026.115476","url":null,"abstract":"<div><div>A sarcoma is considered a rare kind of tumor that generally occurs in various connective tissues that surround and support different organs and bones in the human body. Sarcoma cancer mainly affects the connective tissues, like nerves, blood vessels, muscles, fat, joints, and bones. The primary symptoms of sarcoma cancer are based on the location and size of tumors. Various imaging techniques have been recently applied to identify sarcoma cancer. However, these approaches do not adequately detect the cell response of individuals. Thus, a novel Deep Learning (DL) model, Parallel Le-Net Forward Fractional Network (PLeFF-Net), is proposed for the detection of sarcoma cancer from histopathological images. The histopathological images are primarily preprocessed utilizing homomorphic filtering, and then a Fully Convolutional Neural Network (FCNN) is exploited to segment the cells. Later, image-level features, shape-based features, color-based features, and network-level features are extracted from the segmented areas. For sarcoma cancer detection, the PLeFF-Net model receives the extracted features as its input. Using the k-fold cross-validation for k-value 8 on dataset 2, the proposed PLeFF-Net revealed superior performance with maximum accuracy of 93.789%, True Positive Rate (TNR) of 94.667%, True Negative Rate (TNR) of 92.100%, Precision of 93.357%, and F1-score of 94.007%.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115476"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175180","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}

{"title":"Towards heterogeneity-aware federated self-supervised learning via knowledge anchoring","authors":"Hongpu Jiang ,&nbsp;Jinxin Zuo ,&nbsp;Yueming Lu ,&nbsp;Haonan Li","doi":"10.1016/j.knosys.2026.115446","DOIUrl":"10.1016/j.knosys.2026.115446","url":null,"abstract":"<div><div>Federated Self-Supervised Learning (FSSL) is a promising paradigm for extracting robust representations from decentralized unlabeled data. However, its effectiveness is often hindered by non-IID data distributions and label scarcity, which cause model divergence and limit generalization. In this paper, we propose Federated Self-Supervised and Global-Personalized Collaborative Learning (FedGP), a novel framework designed to bridge the gap between global knowledge integration and local client adaptation. The core of FedGP is the Collaborative Knowledge Anchoring (CKA) mechanism, which utilizes adaptive regularization to anchor shared global knowledge while enabling personalized refinement on local data. By dynamically balancing collaborative risks and local empirical losses via learnable coefficients, FedGP ensures stable convergence in heterogeneous environments. Extensive evaluations on multiple benchmarks, including a real-world private Flora dataset, demonstrate that FedGP consistently outperforms state-of-the-art FSSL methods. Our results confirm that FedGP achieves high-quality representation learning with significantly reduced communication overhead and annotation dependency, providing a scalable solution for privacy-preserving decentralized systems.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115446"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175319","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}

{"title":"Infrared and visible image fusion based on multi-modal and multi-scale cross-compensation","authors":"Meitian Li,&nbsp;Jing Sun,&nbsp;Heng Ma,&nbsp;Fasheng Wang,&nbsp;Fuming Sun","doi":"10.1016/j.knosys.2026.115441","DOIUrl":"10.1016/j.knosys.2026.115441","url":null,"abstract":"&lt;div&gt;&lt;div&gt;In the task of infrared and visible image fusion, fully preserving the complementary information from different modalities while avoiding detail loss and redundant information superposition has been a core challenge in recent research. Most existing methods primarily focus on feature processing at a single level or for a single modality, leading to insufficient cross-level information interaction and inadequate cross-modal feature fusion. This deficiency typically results in two types of issues: firstly, the lack of effective compensation between adjacent-level features prevents the synergistic utilization of low-level details and high-level semantics; secondly, the differences between features from different modalities are not explicitly modeled, where direct concatenation or weighted summation often introduces redundancy or even artifacts, thereby compromising the overall quality of the fused image. To address these challenges, this paper proposes a novel infrared and visible image fusion network based on a Multi-modal and Multi-scale Cross-compensation referred to as MMCFusion. The proposed network incorporates an Upper-Lower-level Cross-Compensation (ULCC) module that integrates features from adjacent levels to enhance the richness and diversity of feature representations. Additionally, we introduce a Feature-Difference Cross-Compensation (FDCC) module to facilitate cross-compensation of upper-lower-level information through a differential approach. This design enhances the complementarity between features and effectively mitigates the problem of detail information loss prevalent in conventional methods. To further augment the model’s ability to detect and represent objects across various scales, we also devise the Multi-Scale Fusion Module (MSFM) that effectively integrates feature information from multiple scales, thereby improving the model’s adaptability to diverse objects. Furthermore, we design a Texture Enhancement Module (TEM) to capture and retain local structures and texture information in the image, thereby providing richer detail representation after processing. Finally, to comprehensively capture multi-modal information and perform remote modeling, we employ Pyramid Vision Transformer (PVTv2) to construct a dual-stream Transformer encoder, which can capture valuable information at multiple scales and provide robust global modeling capabilities, thereby improving the fusion results. The efficacy of the proposed method is rigorously evaluated on several datasets, including infrared and visible datasets such as MSRS, TNO, and RoadScene, as well as medical imaging datasets, such as PET-MRI. Experimental results demonstrate that MMCFusion significantly outperforms current state-of-the-art methods in terms of both visual quality and quantitative metrics, while also exhibiting strong generalization capability across different datasets, thereby validating its effectiveness and robustness in practical applications. The source co","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115441"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175324","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}

{"title":"Towards robust and high-capacity coverless image steganography","authors":"Bobiao Guo,&nbsp;Ping Ping","doi":"10.1016/j.knosys.2026.115472","DOIUrl":"10.1016/j.knosys.2026.115472","url":null,"abstract":"<div><div>Capacity and robustness are key metrics for Coverless Image Steganography (CIS). Although the theoretical maximum capacity of the CIS has been achieved, the robustness at this level remains insufficient, limiting the practical use of high-capacity settings. Therefore, this paper proposes a novel CIS method that maintains high robustness even at the theoretical maximum capacity supported by a given dataset. Our method has three main parts: constructing a stable feature space using Pseudo-Zernike Moments (PZM), proposing a Stability-Aware Piecewise Quantization Encoding (SA-PQE) to assign a stability score to each image, and introducing stability regularization into the clustering process to build a robust Coverless Index Database (CID). Its robust high-capacity performance derives from two core principles: (1) low-order PZM coefficients remain highly stable under distortion, and the independence among PZM coefficients suppresses distortion propagation across dimensions; and (2) stability regularization penalizes images that are close to the cluster center but exhibit low stability. Extensive experimental results demonstrate that our method achieves high robustness at the theoretical maximum capacity. On the Holidays, VOC, and ImageNet datasets, the average robustness reaches 99.54%, 98.64%, and 97.19% at capacities of 10 bits, 14 bits, and 15 bits, respectively. These results significantly outperform existing advanced methods under the setting without inverse image retrieval.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115472"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175317","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}

{"title":"Detective Behavior Algorithm (DBA): A New Metaheuristic for Design and Engineering Optimization","authors":"Jun Cheng ,&nbsp;Wim De Waele","doi":"10.1016/j.knosys.2026.115434","DOIUrl":"10.1016/j.knosys.2026.115434","url":null,"abstract":"<div><div>Hunting-inspired algorithms have gained widespread attention in the field of optimization because of their simplicity, flexibility, and natural metaphors. However, many suffer from limitations such as slow convergence rates, sensitivity to parameter settings, and a tendency to become trapped in local optima. To address these challenges, this paper proposes the Detective Behavior Algorithm (DBA), a novel meta-heuristic approach that integrates three core search mechanisms: large-area directional exploration, localized exploitation, and direct target-oriented attacks. DBA is designed to balance exploration and exploitation effectively, enabling faster convergence and improved global search capabilities. The performance is validated through comprehensive application on a suite of benchmark functions and real-world engineering problems. A comparative analysis is conducted against eight state-of-the-art optimization algorithms, including recently developed hunting-inspired methods such as the Walrus Optimizer and Sea-Horse Optimizer. Results consistently demonstrate that DBA outperforms these approaches in terms of convergence speed, solution accuracy, and robustness, particularly in complex optimization scenarios. Furthermore, DBA is applied to predict and optimize surface waviness in Wire Arc Additive Manufacturing components. Two predictive models are developed: one employing an Artificial Neural Network (ANN) optimized by DBA, and another using Particle Swarm Optimization (PSO). The DBA-optimized ANN model exhibits superior predictive accuracy and reliability compared to both standard ANN and PSO-optimized ANN models. Leveraging this enhanced prediction capability, DBA is further used to minimize surface waviness, consistently outperforming competing algorithms. These findings underscore the robustness, adaptability, and real-world applicability of DBA in both theoretical and practical contexts. The source codes of DBA are publicly available at (<span><span>https://www.mathworks.com/matlabcentral/fileexchange/183178-detective-behavior-algorithm-dba</span><svg><path></path></svg></span>).</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"338 ","pages":"Article 115434"},"PeriodicalIF":7.6,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146175313","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}