Pub Date : 2025-03-11DOI: 10.1016/j.eswa.2025.127093
Fuguo Zhang, Yunhe Liu, Shaoying Feng
The uncertainty in the prediction of a single rating by recommender systems can vary significantly owing to the diverse historical interaction data between users and items under a given recommendation model. In group recommendations, high uncertainty in individual rating predictions may lead to erroneous group decisions. However, previous studies have often overlooked the impact of the uncertainty of individual rating predictions in the group recommendation process. To address this, this study proposes a measurement method for individual prediction certainty that employs the validity of bipartite graph voting. In addition, a group recommendation algorithm named consideration member reliability group recommendation (CMRGR), which integrates the individual prediction uncertainty of each group member, is presented. The results of experiments on the MovieLens-1M, Netflix, and MovieTweetings datasets show that the CMRGR algorithm improved the recommendation accuracy by at least 10% compared with the baseline. Moreover, incorporating the prediction uncertainty into recommendations was found to have approximately twice the impact on group recommendation accuracy compared with individual recommendation accuracy.
{"title":"Enhancing group recommendation performance by integrating individual prediction uncertainty","authors":"Fuguo Zhang, Yunhe Liu, Shaoying Feng","doi":"10.1016/j.eswa.2025.127093","DOIUrl":"10.1016/j.eswa.2025.127093","url":null,"abstract":"<div><div>The uncertainty in the prediction of a single rating by recommender systems can vary significantly owing to the diverse historical interaction data between users and items under a given recommendation model. In group recommendations, high uncertainty in individual rating predictions may lead to erroneous group decisions. However, previous studies have often overlooked the impact of the uncertainty of individual rating predictions in the group recommendation process. To address this, this study proposes a measurement method for individual prediction certainty that employs the validity of bipartite graph voting. In addition, a group recommendation algorithm named consideration member reliability group recommendation (CMRGR), which integrates the individual prediction uncertainty of each group member, is presented. The results of experiments on the MovieLens-1M, Netflix, and MovieTweetings datasets show that the CMRGR algorithm improved the recommendation accuracy by at least 10% compared with the baseline. Moreover, incorporating the prediction uncertainty into recommendations was found to have approximately twice the impact on group recommendation accuracy compared with individual recommendation accuracy.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127093"},"PeriodicalIF":7.5,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591951","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}
Pub Date : 2025-03-11DOI: 10.1016/j.eswa.2025.127092
Luyao You, Xiaodi Li
This paper studies a novel observer-based sliding mode control problem for impulsive systems involving unknown mismatched disturbances. Different from the existing sliding mode control coupling the complete information of known disturbances, our proposed observer-based sliding mode control depends on the disturbance estimation and incorporates the bound of disturbance estimation error into the switch gain design. It is shown that the proposed observer-based sliding mode control not only implicitly restrains the negative effects of unknown mismatched disturbances, but also ensures the reachability of the designed sliding surface in a finite time. Some reachability criteria of impulsive systems are established, where a potential relationship between mismatched disturbances, impulse actions, and sliding function is presented. This relationship fully estimates the effects of discrete dynamics and avoids the resulting sliding mode dynamics jumping out of the designed sliding surface at impulse instants. Following that, some linear matrix inequalities-based conditions are obtained to stabilize the resulting sliding mode dynamics. Finally, two examples are provided to verify our results, including the one focusing on the mass–spring-damper systems.
{"title":"Observer-based sliding mode control of impulsive systems with unknown mismatched disturbances","authors":"Luyao You, Xiaodi Li","doi":"10.1016/j.eswa.2025.127092","DOIUrl":"10.1016/j.eswa.2025.127092","url":null,"abstract":"<div><div>This paper studies a novel observer-based sliding mode control problem for impulsive systems involving unknown mismatched disturbances. Different from the existing sliding mode control coupling the complete information of known disturbances, our proposed observer-based sliding mode control depends on the disturbance estimation and incorporates the bound of disturbance estimation error into the switch gain design. It is shown that the proposed observer-based sliding mode control not only implicitly restrains the negative effects of unknown mismatched disturbances, but also ensures the reachability of the designed sliding surface in a finite time. Some reachability criteria of impulsive systems are established, where a potential relationship between mismatched disturbances, impulse actions, and sliding function is presented. This relationship fully estimates the effects of discrete dynamics and avoids the resulting sliding mode dynamics jumping out of the designed sliding surface at impulse instants. Following that, some linear matrix inequalities-based conditions are obtained to stabilize the resulting sliding mode dynamics. Finally, two examples are provided to verify our results, including the one focusing on the mass–spring-damper systems.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127092"},"PeriodicalIF":7.5,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591953","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}
Pub Date : 2025-03-11DOI: 10.1016/j.eswa.2025.127090
Dexuan Zhao , Fan Yang , Taizhang Hu , Xing Wei , Chong Zhao , Yang Lu
Unsupervised domain adaptation (UDA) methods for image classification have rapidly advanced in recent years. This advancement is attributed to their ability to address performance degradation caused by distributional differences between source and target domain images. However, previous methods often focus solely on sample-level relationships, neglecting the problem of redundancy within samples due to the same feature information being expressed repeatedly. This oversight can smooth critical discriminative features, ultimately reducing the model’s accuracy in image classification. To address this problem, we propose Dual-Level Redundancy Elimination for Unsupervised Domain Adaptation (DLRE), which aims to enhance UDA by reducing redundancy at both the sample feature and cluster levels. Specifically, we use feature mutual information to evaluate and reduce the feature-level redundancy of different dimensions, thereby ensuring the amount of discriminative information contained in the feature vector, and maximize the dimensional mutual information between pairs of positive samples to obtain a more unbiased feature representation. In addition, we propose a new clustering framework based on dictionary storage and mutual information weighting to reduce cluster-level redundancy and help improve the performance of classification tasks. We conduct extensive experiments on five widely used UDA vision benchmarks, and the results show that DLRE has good adaptability and effectiveness, significantly outperforms current domain adaptation methods.
{"title":"Dual-Level Redundancy Elimination for Unsupervised Domain Adaptation","authors":"Dexuan Zhao , Fan Yang , Taizhang Hu , Xing Wei , Chong Zhao , Yang Lu","doi":"10.1016/j.eswa.2025.127090","DOIUrl":"10.1016/j.eswa.2025.127090","url":null,"abstract":"<div><div>Unsupervised domain adaptation (UDA) methods for image classification have rapidly advanced in recent years. This advancement is attributed to their ability to address performance degradation caused by distributional differences between source and target domain images. However, previous methods often focus solely on sample-level relationships, neglecting the problem of redundancy within samples due to the same feature information being expressed repeatedly. This oversight can smooth critical discriminative features, ultimately reducing the model’s accuracy in image classification. To address this problem, we propose Dual-Level Redundancy Elimination for Unsupervised Domain Adaptation (DLRE), which aims to enhance UDA by reducing redundancy at both the sample feature and cluster levels. Specifically, we use feature mutual information to evaluate and reduce the feature-level redundancy of different dimensions, thereby ensuring the amount of discriminative information contained in the feature vector, and maximize the dimensional mutual information between pairs of positive samples to obtain a more unbiased feature representation. In addition, we propose a new clustering framework based on dictionary storage and mutual information weighting to reduce cluster-level redundancy and help improve the performance of classification tasks. We conduct extensive experiments on five widely used UDA vision benchmarks, and the results show that DLRE has good adaptability and effectiveness, significantly outperforms current domain adaptation methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127090"},"PeriodicalIF":7.5,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591956","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}
Pub Date : 2025-03-11DOI: 10.1016/j.eswa.2025.127159
Meiwen Li , Xinyue Long , Muhua Liu , Jing Guo , Xuhui Zhao , Lin Wang , Qingtao Wu
To solve min–max optimization problems, decentralized adaptive methods have been presented over multi-agent networks. In the non-convex non-concave structure, however, existing decentralized adaptive min–max methods may be divergence due to the inconsistency in the adaptive learning rate. To address this issue, we propose a novel decentralized adaptive algorithm named DADAMC, where the consensus protocol is introduced to synchronize the adaptive learning rates of all agents. Furthermore, we rigorously analyze that DADAMC converges to an -stochastic first-order stationary point with complexity. In addition, we also conduct experiments to verify the performance of DADAMC for solving a robust regression problem. The experimental results show that DADAMC outperforms state-of-the-art decentralized min–max algorithms.
{"title":"A decentralized adaptive method with consensus step for non-convex non-concave min-max optimization problems","authors":"Meiwen Li , Xinyue Long , Muhua Liu , Jing Guo , Xuhui Zhao , Lin Wang , Qingtao Wu","doi":"10.1016/j.eswa.2025.127159","DOIUrl":"10.1016/j.eswa.2025.127159","url":null,"abstract":"<div><div>To solve min–max optimization problems, decentralized adaptive methods have been presented over multi-agent networks. In the non-convex non-concave structure, however, existing decentralized adaptive min–max methods may be divergence due to the inconsistency in the adaptive learning rate. To address this issue, we propose a novel decentralized adaptive algorithm named DADAMC, where the consensus protocol is introduced to synchronize the adaptive learning rates of all agents. Furthermore, we rigorously analyze that DADAMC converges to an <span><math><mi>ϵ</mi></math></span>-stochastic first-order stationary point with <span><math><mrow><mi>O</mi><mrow><mo>(</mo><msup><mrow><mi>ϵ</mi></mrow><mrow><mo>−</mo><mn>4</mn></mrow></msup><mo>)</mo></mrow></mrow></math></span> complexity. In addition, we also conduct experiments to verify the performance of DADAMC for solving a robust regression problem. The experimental results show that DADAMC outperforms state-of-the-art decentralized min–max algorithms.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127159"},"PeriodicalIF":7.5,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601670","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}
The alternating direction method of multipliers (ADMM) is a widely employed first-order method due to its efficiency and simplicity. Nonetheless, like other splitting methods, ADMM’s performance degrades substantially as the scale of the optimization problems it addresses increases. This work is devoted to studying an accelerated stochastic generalized ADMM framework with a class of variance-reduced gradient estimators for solving large-scale nonconvex nonsmooth optimization problems with linear constraints, in which we combine inertial technique and Bregman distance. Under the assumption that the objective functions are semi-algebraic which satisfies the Kurdyka–Łojasiewicz (KL) property, we establish the global convergence and convergence rate of the sequence generated by our proposed algorithm. Finally, numerical experiments on conducting a graph-guided fused lasso illustrates the efficiency of the proposed method.
{"title":"An inertial stochastic Bregman generalized alternating direction method of multipliers for nonconvex and nonsmooth optimization","authors":"Longhui Liu, Congying Han, Tiande Guo, Shichen Liao","doi":"10.1016/j.eswa.2025.126939","DOIUrl":"10.1016/j.eswa.2025.126939","url":null,"abstract":"<div><div>The alternating direction method of multipliers (ADMM) is a widely employed first-order method due to its efficiency and simplicity. Nonetheless, like other splitting methods, ADMM’s performance degrades substantially as the scale of the optimization problems it addresses increases. This work is devoted to studying an accelerated stochastic generalized ADMM framework with a class of variance-reduced gradient estimators for solving large-scale nonconvex nonsmooth optimization problems with linear constraints, in which we combine inertial technique and Bregman distance. Under the assumption that the objective functions are semi-algebraic which satisfies the Kurdyka–Łojasiewicz (KL) property, we establish the global convergence and convergence rate of the sequence generated by our proposed algorithm. Finally, numerical experiments on conducting a graph-guided fused lasso illustrates the efficiency of the proposed method.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 126939"},"PeriodicalIF":7.5,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601673","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}
Pub Date : 2025-03-10DOI: 10.1016/j.eswa.2025.127131
Marcin Czajkowski, Krzysztof Jurczuk, Marek Kretowski
This paper presents a novel classification algorithm for multi-omics data, called Evolutionary Multi-Test Tree with Relative Expression (EMTTree+RX). The innovation lies in the model’s design, which integrates multi-test decision nodes with Relative Expression Analysis (RXA). Each decision node combines traditional univariate tests and top-scoring pair (TSP) comparisons, allowing the algorithm to capture complex relationships between features without relying solely on absolute values. This approach enables the proposed method to detect subtle patterns across various omics layers while maintaining a high level of interpretability, a feature crucial for clinical and bioinformatics applications. The tree structure is induced through Evolutionary Algorithms (EA), optimizing both the global architecture and local multi-test nodes to balance classification accuracy, test diversity, and feature cost. Applied to large-scale multi-omics datasets, where conventional decision tree methods often struggle with underfitting or overfitting, the proposed method consistently outperforms traditional models in terms of accuracy and transparency. This makes it a valuable tool for precision medicine and multi-modal data integration.
{"title":"Enhancing transparency of omics data analysis with the Evolutionary Multi-Test Tree and Relative Expression","authors":"Marcin Czajkowski, Krzysztof Jurczuk, Marek Kretowski","doi":"10.1016/j.eswa.2025.127131","DOIUrl":"10.1016/j.eswa.2025.127131","url":null,"abstract":"<div><div>This paper presents a novel classification algorithm for multi-omics data, called Evolutionary Multi-Test Tree with Relative Expression (EMTTree+RX). The innovation lies in the model’s design, which integrates multi-test decision nodes with Relative Expression Analysis (RXA). Each decision node combines traditional univariate tests and top-scoring pair (TSP) comparisons, allowing the algorithm to capture complex relationships between features without relying solely on absolute values. This approach enables the proposed method to detect subtle patterns across various omics layers while maintaining a high level of interpretability, a feature crucial for clinical and bioinformatics applications. The tree structure is induced through Evolutionary Algorithms (EA), optimizing both the global architecture and local multi-test nodes to balance classification accuracy, test diversity, and feature cost. Applied to large-scale multi-omics datasets, where conventional decision tree methods often struggle with underfitting or overfitting, the proposed method consistently outperforms traditional models in terms of accuracy and transparency. This makes it a valuable tool for precision medicine and multi-modal data integration.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127131"},"PeriodicalIF":7.5,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143611236","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}
Pub Date : 2025-03-10DOI: 10.1016/j.eswa.2025.126861
Debasish Pradhan, Ranjit Kumar Upadhyay
<div><div>Exposure of astrocytes to amyloid beta (<span><math><msub><mrow><mi>A</mi></mrow><mrow><mi>β</mi></mrow></msub></math></span>) is believed to trigger the dysregulation of intracellular calcium (<span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span>) oscillations. This study explores a fractional-order model of <span><math><msub><mrow><mi>A</mi></mrow><mrow><mi>β</mi></mrow></msub></math></span>-directed astrocytic <span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span> dynamics, focusing on multi-pathway flux contributions, stability, and solution properties. The complex behaviors of the model system are explained under three parameter regimes by incorporating an additional external current <span><math><mrow><mo>(</mo><msub><mrow><mi>I</mi></mrow><mrow><mi>e</mi><mi>x</mi><mi>t</mi></mrow></msub><mo>)</mo></mrow></math></span> in voltage-gated calcium channels. These behaviors include quiescent states, periodic spiking, and mixed-mode oscillations (MMOs), describing the memory effect of aberrant calcium signaling. Additionally, the bifurcation analysis of <span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span> with <span><math><msub><mrow><mi>A</mi></mrow><mrow><mi>β</mi></mrow></msub></math></span> level and <span><math><msub><mrow><mi>I</mi></mrow><mrow><mi>e</mi><mi>x</mi><mi>t</mi></mrow></msub></math></span> reveals that amyloid beta substantially induces the calcium oscillation and provides insight into why spontaneous astrocytic <span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span> oscillations appear and disappear. Further, a coupled model of astrocytes with fractional order is introduced through linear and nonlinear gap junctions in intercellular diffusion of IP3 and then extended to a network model to study the synchronized firing activities. We demonstrate that for a coupled system with differing fractional-order coefficients <span><math><mrow><mo>(</mo><mi>α</mi><mo>≠</mo><mi>β</mi><mo>)</mo></mrow></math></span>, as the coupling strength <span><math><mi>F</mi></math></span> increases, the system with linear coupling can achieve complete synchronization for a higher value of <span><math><mi>F</mi></math></span>. In contrast, nonlinear coupling fails to synchronize at the same strength, indicating the complexity of the underlying system dynamics. For a network model with linear coupling, we observe that at <span><math><mrow><mi>F</mi><mo>=</mo><mn>20</mn></mrow></math></span> and 110, the system displays evidence of synchronized behavior. These results highlight that linear gap junctions with weak coupling could potentially explain the intricate intracellular oscillations observed during wave propagation in astrocyte networks, which has implications for neurodegenerative dise
{"title":"Deciphering the calcium dynamics of a fractional order Alzheimer’s disease model in astrocytes and their networks","authors":"Debasish Pradhan, Ranjit Kumar Upadhyay","doi":"10.1016/j.eswa.2025.126861","DOIUrl":"10.1016/j.eswa.2025.126861","url":null,"abstract":"<div><div>Exposure of astrocytes to amyloid beta (<span><math><msub><mrow><mi>A</mi></mrow><mrow><mi>β</mi></mrow></msub></math></span>) is believed to trigger the dysregulation of intracellular calcium (<span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span>) oscillations. This study explores a fractional-order model of <span><math><msub><mrow><mi>A</mi></mrow><mrow><mi>β</mi></mrow></msub></math></span>-directed astrocytic <span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span> dynamics, focusing on multi-pathway flux contributions, stability, and solution properties. The complex behaviors of the model system are explained under three parameter regimes by incorporating an additional external current <span><math><mrow><mo>(</mo><msub><mrow><mi>I</mi></mrow><mrow><mi>e</mi><mi>x</mi><mi>t</mi></mrow></msub><mo>)</mo></mrow></math></span> in voltage-gated calcium channels. These behaviors include quiescent states, periodic spiking, and mixed-mode oscillations (MMOs), describing the memory effect of aberrant calcium signaling. Additionally, the bifurcation analysis of <span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span> with <span><math><msub><mrow><mi>A</mi></mrow><mrow><mi>β</mi></mrow></msub></math></span> level and <span><math><msub><mrow><mi>I</mi></mrow><mrow><mi>e</mi><mi>x</mi><mi>t</mi></mrow></msub></math></span> reveals that amyloid beta substantially induces the calcium oscillation and provides insight into why spontaneous astrocytic <span><math><msub><mrow><mrow><mo>[</mo><mi>C</mi><mo>]</mo></mrow></mrow><mrow><mi>i</mi></mrow></msub></math></span> oscillations appear and disappear. Further, a coupled model of astrocytes with fractional order is introduced through linear and nonlinear gap junctions in intercellular diffusion of IP3 and then extended to a network model to study the synchronized firing activities. We demonstrate that for a coupled system with differing fractional-order coefficients <span><math><mrow><mo>(</mo><mi>α</mi><mo>≠</mo><mi>β</mi><mo>)</mo></mrow></math></span>, as the coupling strength <span><math><mi>F</mi></math></span> increases, the system with linear coupling can achieve complete synchronization for a higher value of <span><math><mi>F</mi></math></span>. In contrast, nonlinear coupling fails to synchronize at the same strength, indicating the complexity of the underlying system dynamics. For a network model with linear coupling, we observe that at <span><math><mrow><mi>F</mi><mo>=</mo><mn>20</mn></mrow></math></span> and 110, the system displays evidence of synchronized behavior. These results highlight that linear gap junctions with weak coupling could potentially explain the intricate intracellular oscillations observed during wave propagation in astrocyte networks, which has implications for neurodegenerative dise","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 126861"},"PeriodicalIF":7.5,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143611240","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}
Pub Date : 2025-03-10DOI: 10.1016/j.eswa.2025.127204
Zhigang Sun , Guofu Zhai , Guotao Wang , Qi Liang , Jingting Sun , Hao Chen
The existing research on loose particle localization using machine learning methods takes one segment of loose particle signal as the classification object, and uses the single-channel feature data set creation method to create localization data sets, ignoring the correlation and complementarity between four-channel loose particle signals, thus the trained classifier achieved a limited classification accuracy of 84.08%. In this paper, for the first time, the authors took four-channel loose particle signals as the classification object, and seriously considered the correlation and complementarity between them, thus proposed a feature data set creation method for loose particle localization based on multi-channel characteristic encoding. Specifically, for the four-channel loose particle signals, the three-threshold pulse extraction algorithm was used to extract effective pulses, the barrel-principle-based pulse matching algorithm was newly proposed to match effective pulse groups. The values of the eleven applicable time–frequency-domain features were respectively calculated on effective pulse groups, the mean of the four values corresponding to the same time–frequency-domain feature was calculated, and eleven new values were obtained. Four-bit encoding was newly introduced to clearly quantify the correlation and complementarity between four-channel loose particle signals, thus four values were obtained. On this basis, fused feature vectors were constructed, the encoding localization data set was created. Test results on one type of spaceborne electronic equipment show that, compared with the localization data sets created by existing methods, multiple classifiers trained on the encoding localization data set achieves the highest classification accuracy. The feasibility and superiority of the proposed method are fully demonstrated. The in-depth validation results represented by random forest show that, it achieves the most significant and stable classification effect on the encoding localization data set, with the least time loss. The stability and efficiency of the proposed method are fully demonstrated. Currently, the proposed method achieves the highest classification accuracy in loose particle localization research, at 94.03%, an increase of 9.95% compared to previous.
{"title":"Feature data set creation method for loose particle localization based on multi-channel characteristic encoding","authors":"Zhigang Sun , Guofu Zhai , Guotao Wang , Qi Liang , Jingting Sun , Hao Chen","doi":"10.1016/j.eswa.2025.127204","DOIUrl":"10.1016/j.eswa.2025.127204","url":null,"abstract":"<div><div>The existing research on loose particle localization using machine learning methods takes one segment of loose particle signal as the classification object, and uses the single-channel feature data set creation method to create localization data sets, ignoring the correlation and complementarity between four-channel loose particle signals, thus the trained classifier achieved a limited classification accuracy of 84.08%. In this paper, for the first time, the authors took four-channel loose particle signals as the classification object, and seriously considered the correlation and complementarity between them, thus proposed a feature data set creation method for loose particle localization based on multi-channel characteristic encoding. Specifically, for the four-channel loose particle signals, the three-threshold pulse extraction algorithm was used to extract effective pulses, the barrel-principle-based pulse matching algorithm was newly proposed to match effective pulse groups. The values of the eleven applicable time–frequency-domain features were respectively calculated on effective pulse groups, the mean of the four values corresponding to the same time–frequency-domain feature was calculated, and eleven new values were obtained. Four-bit encoding was newly introduced to clearly quantify the correlation and complementarity between four-channel loose particle signals, thus four values were obtained. On this basis, fused feature vectors were constructed, the encoding localization data set was created. Test results on one type of spaceborne electronic equipment show that, compared with the localization data sets created by existing methods, multiple classifiers trained on the encoding localization data set achieves the highest classification accuracy. The feasibility and superiority of the proposed method are fully demonstrated. The in-depth validation results represented by random forest show that, it achieves the most significant and stable classification effect on the encoding localization data set, with the least time loss. The stability and efficiency of the proposed method are fully demonstrated. Currently, the proposed method achieves the highest classification accuracy in loose particle localization research, at 94.03%, an increase of 9.95% compared to previous.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127204"},"PeriodicalIF":7.5,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601666","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}
Pub Date : 2025-03-10DOI: 10.1016/j.eswa.2025.126904
Dequan Jin , Ruoge Li , Nan Xiang , Di Zhao , Xuanlu Xiang , Shihui Ying
Small-sample image classification is a hot topic in computer vision. Despite the progress made by some deep neural networks in solving the small-sample learning problem, there remain challenges in learning efficiently and robustly. These challenges can affect the overall performance and effectiveness of the model. To address these issues, we propose a hierarchical cognitive neural model (HCNM) based on the simulation of visual cognition to construct the sparse structure of the neural model from the perspective of semi-supervised learning. We use a deep learning network for feature extraction and two coupled dynamic neural field equations to simulate the encoding and classification functions in visual image recognition and classification. The model simulates macroscopic neural activation in object recognition and identifies representative point neurons (RPNs) by evaluating the magnitude of lateral interactions within the V4 neural field on an adaptive cognitive scale. Our approach provides an efficient small-sample image classification algorithm that does not require complex parameter tuning and maintains biological plausibility and interpretability. Experimental results using four real-world image datasets demonstrate the superior performance of our model and method for small-sample image classification compared to other state-of-the-art research methods.
{"title":"HCNM: Hierarchical cognitive neural model for small-sample image classification","authors":"Dequan Jin , Ruoge Li , Nan Xiang , Di Zhao , Xuanlu Xiang , Shihui Ying","doi":"10.1016/j.eswa.2025.126904","DOIUrl":"10.1016/j.eswa.2025.126904","url":null,"abstract":"<div><div>Small-sample image classification is a hot topic in computer vision. Despite the progress made by some deep neural networks in solving the small-sample learning problem, there remain challenges in learning efficiently and robustly. These challenges can affect the overall performance and effectiveness of the model. To address these issues, we propose a hierarchical cognitive neural model (HCNM) based on the simulation of visual cognition to construct the sparse structure of the neural model from the perspective of semi-supervised learning. We use a deep learning network for feature extraction and two coupled dynamic neural field equations to simulate the encoding and classification functions in visual image recognition and classification. The model simulates macroscopic neural activation in object recognition and identifies representative point neurons (RPNs) by evaluating the magnitude of lateral interactions within the V4 neural field on an adaptive cognitive scale. Our approach provides an efficient small-sample image classification algorithm that does not require complex parameter tuning and maintains biological plausibility and interpretability. Experimental results using four real-world image datasets demonstrate the superior performance of our model and method for small-sample image classification compared to other state-of-the-art research methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 126904"},"PeriodicalIF":7.5,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601744","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}
Pub Date : 2025-03-10DOI: 10.1016/j.eswa.2025.127108
Kai Zhou , Guanglu Sun , Jun Wang , Linsen Yu , Tianlin Li
The rapid advancement of Deepfake technology has rendered the generation of forged faces highly realistic, while simultaneously introducing significant societal security concerns. The accurate detection of forged facial images has thus emerged as an urgent issue and a formidable challenge. In this paper, we approach face forgery detection as a fine-grained classification problem due to the subtle differences between real and fake faces. We propose a detection framework termed the Mid-High Frequency Based Fine-Grained Network (MH-FFNet), which enhances the detection of forged faces by leveraging mid- and high-frequency information to capture fine-grained forgery cues. To better extract and utilize these cues, we devise two fine-grained feature enhancement modules: the Patch-based Fine-Grained Enhancement Module (P-FGEM) and the Feature-based Fine-Grained Enhancement Module (F-FGEM). The P-FGEM module focuses on extracting mid- and high-frequency information from shallow feature blocks, enhancing forgery representations in shallow features. This design effectively mitigates the loss of mid- and high-frequency cues as the network deepens, thereby improving the algorithm’s sensitivity to forgery cues. In contrast, the F-FGEM module captures mid- and high-frequency information from mid-level global features, further enriching forgery representations in these features and significantly enhancing their discriminative power. Experimental results indicate that our proposed method achieves an AUC of 99.44% on the FF++ (C23) dataset and 83.44% on the Celeb-DF (V2) dataset, demonstrating the algorithm’s superior detection capability and generalization performance. Additionally, we conduct experiments to comprehensively illustrate the robustness of the algorithm against common image post-processing attacks.
{"title":"MH-FFNet: Leveraging mid-high frequency information for robust fine-grained face forgery detection","authors":"Kai Zhou , Guanglu Sun , Jun Wang , Linsen Yu , Tianlin Li","doi":"10.1016/j.eswa.2025.127108","DOIUrl":"10.1016/j.eswa.2025.127108","url":null,"abstract":"<div><div>The rapid advancement of Deepfake technology has rendered the generation of forged faces highly realistic, while simultaneously introducing significant societal security concerns. The accurate detection of forged facial images has thus emerged as an urgent issue and a formidable challenge. In this paper, we approach face forgery detection as a fine-grained classification problem due to the subtle differences between real and fake faces. We propose a detection framework termed the Mid-High Frequency Based Fine-Grained Network (MH-FFNet), which enhances the detection of forged faces by leveraging mid- and high-frequency information to capture fine-grained forgery cues. To better extract and utilize these cues, we devise two fine-grained feature enhancement modules: the Patch-based Fine-Grained Enhancement Module (P-FGEM) and the Feature-based Fine-Grained Enhancement Module (F-FGEM). The P-FGEM module focuses on extracting mid- and high-frequency information from shallow feature blocks, enhancing forgery representations in shallow features. This design effectively mitigates the loss of mid- and high-frequency cues as the network deepens, thereby improving the algorithm’s sensitivity to forgery cues. In contrast, the F-FGEM module captures mid- and high-frequency information from mid-level global features, further enriching forgery representations in these features and significantly enhancing their discriminative power. Experimental results indicate that our proposed method achieves an AUC of 99.44% on the FF++ (C23) dataset and 83.44% on the Celeb-DF (V2) dataset, demonstrating the algorithm’s superior detection capability and generalization performance. Additionally, we conduct experiments to comprehensively illustrate the robustness of the algorithm against common image post-processing attacks.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"276 ","pages":"Article 127108"},"PeriodicalIF":7.5,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143611237","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}