Graph neural networks(GNN)have shown strong performance in node classification tasks,yet most existing models rely on uniform or shared weight aggregation,lacking flexibility in modeling the varying strength of relati...Graph neural networks(GNN)have shown strong performance in node classification tasks,yet most existing models rely on uniform or shared weight aggregation,lacking flexibility in modeling the varying strength of relationships among nodes.This paper proposes a novel graph coupling convolutional model that introduces an adaptive weighting mechanism to assign distinct importance to neighboring nodes based on their similarity to the central node.Unlike traditional methods,the proposed coupling strategy enhances the interpretability of node interactions while maintaining competitive classification performance.The model operates in the spatial domain,utilizing adjacency list structures for efficient convolution and addressing the limitations of weight sharing through a coupling-based similarity computation.Extensive experiments are conducted on five graph-structured datasets,including Cora,Citeseer,PubMed,Reddit,and BlogCatalog,as well as a custom topology dataset constructed from the Open University Learning Analytics Dataset(OULAD)educational platform.Results demonstrate that the proposed model achieves good classification accuracy,while significantly reducing training time through direct second-order neighbor fusion and data preprocessing.Moreover,analysis of neighborhood order reveals that considering third-order neighbors offers limited accuracy gains but introduces considerable computational overhead,confirming the efficiency of first-and second-order convolution in practical applications.Overall,the proposed graph coupling model offers a lightweight,interpretable,and effective framework for multi-label node classification in complex networks.展开更多
Aiming at the problem of multi-label classification, a multi-label classification algorithm based on label-specific features is proposed in this paper. In this algorithm, we compute feature density on the positive and...Aiming at the problem of multi-label classification, a multi-label classification algorithm based on label-specific features is proposed in this paper. In this algorithm, we compute feature density on the positive and negative instances set of each class firstly and then select mk features of high density from the positive and negative instances set of each class, respectively; the intersec- tion is taken as the label-specific features of the corresponding class. Finally, multi-label data are classified on the basis of la- bel-specific features. The algorithm can show the label-specific features of each class. Experiments show that our proposed method, the MLSF algorithm, performs significantly better than the other state-of-the-art multi-label learning approaches.展开更多
The Internet of Medical Things(IoMT)will come to be of great importance in the mediation of medical disputes,as it is emerging as the core of intelligent medical treatment.First,IoMT can track the entire medical treat...The Internet of Medical Things(IoMT)will come to be of great importance in the mediation of medical disputes,as it is emerging as the core of intelligent medical treatment.First,IoMT can track the entire medical treatment process in order to provide detailed trace data in medical dispute resolution.Second,IoMT can infiltrate the ongoing treatment and provide timely intelligent decision support to medical staff.This information includes recommendation of similar historical cases,guidance for medical treatment,alerting of hired dispute profiteers etc.The multi-label classification of medical dispute documents(MDDs)plays an important role as a front-end process for intelligent decision support,especially in the recommendation of similar historical cases.However,MDDs usually appear as long texts containing a large amount of redundant information,and there is a serious distribution imbalance in the dataset,which directly leads to weaker classification performance.Accordingly,in this paper,a multi-label classification method based on key sentence extraction is proposed for MDDs.The method is divided into two parts.First,the attention-based hierarchical bi-directional long short-term memory(BiLSTM)model is used to extract key sentences from documents;second,random comprehensive sampling Bagging(RCS-Bagging),which is an ensemble multi-label classification model,is employed to classify MDDs based on key sentence sets.The use of this approach greatly improves the classification performance.Experiments show that the performance of the two models proposed in this paper is remarkably better than that of the baseline methods.展开更多
Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on t...Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on the benchmark datasets have been proposed for multi-label classification task in the literature.Furthermore,several open-source tools implementing these approaches have also been developed.However,the characteristics of real-world multi-label patent and publication datasets are not completely in line with those of benchmark ones.Therefore,the main purpose of this paper is to evaluate comprehensively seven multi-label classification methods on real-world datasets.Research limitations:Three real-world datasets differ in the following aspects:statement,data quality,and purposes.Additionally,open-source tools designed for multi-label classification also have intrinsic differences in their approaches for data processing and feature selection,which in turn impacts the performance of a multi-label classification approach.In the near future,we will enhance experimental precision and reinforce the validity of conclusions by employing more rigorous control over variables through introducing expanded parameter settings.Practical implications:The observed Macro F1 and Micro F1 scores on real-world datasets typically fall short of those achieved on benchmark datasets,underscoring the complexity of real-world multi-label classification tasks.Approaches leveraging deep learning techniques offer promising solutions by accommodating the hierarchical relationships and interdependencies among labels.With ongoing enhancements in deep learning algorithms and large-scale models,it is expected that the efficacy of multi-label classification tasks will be significantly improved,reaching a level of practical utility in the foreseeable future.Originality/value:(1)Seven multi-label classification methods are comprehensively compared on three real-world datasets.(2)The TextCNN and TextRCNN models perform better on small-scale datasets with more complex hierarchical structure of labels and more balanced document-label distribution.(3)The MLkNN method works better on the larger-scale dataset with more unbalanced document-label distribution.展开更多
Phononic crystals,as artificial composite materials,have sparked significant interest due to their novel characteristics that emerge upon the introduction of nonlinearity.Among these properties,second-harmonic feature...Phononic crystals,as artificial composite materials,have sparked significant interest due to their novel characteristics that emerge upon the introduction of nonlinearity.Among these properties,second-harmonic features exhibit potential applications in acoustic frequency conversion,non-reciprocal wave propagation,and non-destructive testing.Precisely manipulating the harmonic band structure presents a major challenge in the design of nonlinear phononic crystals.Traditional design approaches based on parameter adjustments to meet specific application requirements are inefficient and often yield suboptimal performance.Therefore,this paper develops a design methodology using Softmax logistic regression and multi-label classification learning to inversely design the material distribution of nonlinear phononic crystals by exploiting information from harmonic transmission spectra.The results demonstrate that the neural network-based inverse design method can effectively tailor nonlinear phononic crystals with desired functionalities.This work establishes a mapping relationship between the band structure and the material distribution within phononic crystals,providing valuable insights into the inverse design of metamaterials.展开更多
Sum-product networks(SPNs)are an expressive deep probabilistic architecture with solid theoretical foundations,which allows tractable and exact inference.SPNs always act as black-box inference machine in many artifici...Sum-product networks(SPNs)are an expressive deep probabilistic architecture with solid theoretical foundations,which allows tractable and exact inference.SPNs always act as black-box inference machine in many artificial intelligence tasks.Due to their recursive definition,SPNs can also be naturally employed as hierarchical feature extractors.Recently,SPNs have been successfully employed as autoencoder framework in representation learning.However,SPNs autoencoder ignores the model structural duality and trains the models separately and independently.In this work,we propose a Dual-SPNs autoencoder which designs two SPNs autoencoders to compose as a dual form.This approach trains the models simultaneously,and explicitly exploits the structural duality between them to enhance the training process.Experimental results on several multilabel classification problems demonstrate that Dual-SPNs autoencoder is very competitive against with state-of-the-art autoencoder architectures.展开更多
In the last few years,smartphone usage and driver sleepiness have been unanimously considered to lead to numerous road accidents,which causes many scholars to pay attention to autonomous driving.For this complexity sc...In the last few years,smartphone usage and driver sleepiness have been unanimously considered to lead to numerous road accidents,which causes many scholars to pay attention to autonomous driving.For this complexity scene,one of the major challenges is mining information comprehensively from massive features in vehicle video.This paper proposes a multi-label classification method MCM-VV(Multi-label Classification Method for Vehicle Video)for vehicle video to judge the label of road condition for unmanned system.Method MCM-VV includes a process of feature extraction and a process of multi-label classification.During feature extraction,grayscale,lane line and the edge of main object are extracted after video preprocessing.During the multi-label classification,the algorithm DR-ML-KNN(Multi-label K-nearest Neighbor Classification Algorithm based on Dimensionality Reduction)learns the training set to obtain multi-label classifier,then predicts the label of road condition according to maximum a posteriori principle,finally outputs labels and adds the new instance to training set for the optimization of classifier.Experimental results on five vehicle video datasets show that the method MCM-VV is effective and efficient.The DR-ML-KNN algorithm reduces the runtime by 50%.It also reduces the time complexity and improves the accuracy.展开更多
Multi-label learning deals with data associated with a set of labels simultaneously. Dimensionality reduction is an important but challenging task in multi-label learning. Feature selection is an efficient technique f...Multi-label learning deals with data associated with a set of labels simultaneously. Dimensionality reduction is an important but challenging task in multi-label learning. Feature selection is an efficient technique for dimensionality reduction to search an optimal feature subset preserving the most relevant information. In this paper, we propose an effective feature evaluation criterion for multi-label feature selection, called neighborhood relationship preserving score. This criterion is inspired by similarity preservation, which is widely used in single-label feature selection. It evaluates each feature subset by measuring its capability in preserving neighborhood relationship among samples. Unlike similarity preservation, we address the order of sample similarities which can well express the neighborhood relationship among samples, not just the pairwise sample similarity. With this criterion, we also design one ranking algorithm and one greedy algorithm for feature selection problem. The proposed algorithms are validated in six publicly available data sets from machine learning repository. Experimental results demonstrate their superiorities over the compared state-of-the-art methods.展开更多
In this paper a novel coupled attribute similarity learning method is proposed with the basis on the multi-label categorical data（CASonMLCD）.The CASonMLCD method not only computes the correlations between different ...In this paper a novel coupled attribute similarity learning method is proposed with the basis on the multi-label categorical data（CASonMLCD）.The CASonMLCD method not only computes the correlations between different attributes and multi-label sets using information gain,which can be regarded as the important degree of each attribute in the attribute learning method,but also further analyzes the intra-coupled and inter-coupled interactions between an attribute value pair for different attributes and multiple labels.The paper compared the CASonMLCD method with the OF distance and Jaccard similarity,which is based on the MLKNN algorithm according to 5common evaluation criteria.The experiment results demonstrated that the CASonMLCD method can mine the similarity relationship more accurately and comprehensively,it can obtain better performance than compared methods.展开更多
Multitasking and multioutput neural networks models jointly learn related classification tasks from a shared structure.Hard parameters sharing is a multitasking approach that shares hidden layers between multiple task...Multitasking and multioutput neural networks models jointly learn related classification tasks from a shared structure.Hard parameters sharing is a multitasking approach that shares hidden layers between multiple taskspecific outputs.The output layers’weights are essential in transforming aggregated neurons outputs into tasks labels.This paper redirects the multioutput network research to prove that the ensemble of output layers prediction can improve network performance in classifying multi-label classification tasks.The network’s output layers initialized with different weights simulate multiple semi-independent classifiers that can make non-identical label sets predictions for the same instance.The ensemble of a multi-output neural network that learns to classify the same multi-label classification task per output layer can outperform an individual output layer neural network.We propose an ensemble strategy of output layers components in the multi-output neural network for multi-label classification(ENSOCOM).The baseline and proposed models are selected based on the size of the hidden layer and the number of output layers to evaluate the proposed method comprehensively.The ENSOCOM method improved the performance of the neural networks on five different multi-label datasets based on several evaluation metrics.The methods presented in this work can substitute the standard labels representation and predictions generation of any neural network.展开更多
Automated and accurate movie genre classification is crucial for content organization,recommendation systems,and audience targeting in the film industry.Although most existing approaches focus on audiovisual features ...Automated and accurate movie genre classification is crucial for content organization,recommendation systems,and audience targeting in the film industry.Although most existing approaches focus on audiovisual features such as trailers and posters,the text-based classification remains underexplored despite its accessibility and semantic richness.This paper introduces the Genre Attention Model(GAM),a deep learning architecture that integrates transformer models with a hierarchical attention mechanism to extract and leverage contextual information from movie plots formulti-label genre classification.In order to assess its effectiveness,we assessmultiple transformer-based models,including Bidirectional Encoder Representations fromTransformers(BERT),ALite BERT(ALBERT),Distilled BERT(DistilBERT),Robustly Optimized BERT Pretraining Approach(RoBERTa),Efficiently Learning an Encoder that Classifies Token Replacements Accurately(ELECTRA),eXtreme Learning Network(XLNet)and Decodingenhanced BERT with Disentangled Attention(DeBERTa).Experimental results demonstrate the superior performance of DeBERTa-based GAM,which employs a two-tier hierarchical attention mechanism:word-level attention highlights key terms,while sentence-level attention captures critical narrative segments,ensuring a refined and interpretable representation of movie plots.Evaluated on three benchmark datasets Trailers12K,Large Movie Trailer Dataset-9(LMTD-9),and MovieLens37K.GAM achieves micro-average precision scores of 83.63%,83.32%,and 83.34%,respectively,surpassing state-of-the-artmodels.Additionally,GAMis computationally efficient,requiring just 6.10Giga Floating Point Operations Per Second(GFLOPS),making it a scalable and cost-effective solution.These results highlight the growing potential of text-based deep learning models in genre classification and GAM’s effectiveness in improving predictive accuracy while maintaining computational efficiency.With its robust performance,GAM offers a versatile and scalable framework for content recommendation,film indexing,and media analytics,providing an interpretable alternative to traditional audiovisual-based classification techniques.展开更多
In its 2023 global health statistics,the World Health Organization noted that noncommunicable diseases(NCDs)remain the leading cause of disease burden worldwide,with cardiovascular diseases(CVDs)resulting in more deat...In its 2023 global health statistics,the World Health Organization noted that noncommunicable diseases(NCDs)remain the leading cause of disease burden worldwide,with cardiovascular diseases(CVDs)resulting in more deaths than the three other major NCDs combined.In this study,we developed a method that can comprehensively detect which CVDs are present in a patient.Specifically,we propose a multi-label classification method that utilizes photoplethysmography(PPG)signals and physiological characteristics from public datasets to classify four types of CVDs and related conditions:hypertension,diabetes,cerebral infarction,and cerebrovascular disease.Our approach to multi-disease classification of cardiovascular diseases(CVDs)using PPG signals achieves the highest classification performance when encompassing the broadest range of disease categories,thereby offering a more comprehensive assessment of human health.We employ a multi-label classification strategy to simultaneously predict the presence or absence of multiple diseases.Specifically,we first apply the Savitzky-Golay(S-G)filter to the PPG signals to reduce noise and then transform into statistical features.We integrate processed PPG signals with individual physiological features as a multimodal input,thereby expanding the learned feature space.Notably,even with a simple machine learning method,this approach can achieve relatively high accuracy.The proposed method achieved a maximum F1-score of 0.91,minimum Hamming loss of 0.04,and an accuracy of 0.95.Thus,our method represents an effective and rapid solution for detecting multiple diseases simultaneously,which is beneficial for comprehensively managing CVDs.展开更多
Automated cartoon character recognition is crucial for applications in content indexing,filtering,and copyright protection,yet it faces a significant challenge in animated media due to high intra-class visual variabil...Automated cartoon character recognition is crucial for applications in content indexing,filtering,and copyright protection,yet it faces a significant challenge in animated media due to high intra-class visual variability,where characters frequently alter their appearance.To address this problem,we introduce the novel Kral Sakir dataset,a public benchmark of 16,725 images specifically curated for the task of multi-label cartoon character classification under these varied conditions.This paper conducts a comprehensive benchmark study,evaluating the performance of state-of-the-art pretrained Convolutional Neural Networks(CNNs),including DenseNet,ResNet,and VGG,against a custom baseline model trained from scratch.Our experiments,evaluated using metrics of F1-Score,accuracy,and Area Under the ROC Curve(AUC),demonstrate that fine-tuning pretrained models is a highly effective strategy.The best-performing model,DenseNet121,achieved an F1-Score of 0.9890 and an accuracy of 0.9898,significantly outperforming our baseline CNN(F1-Score of 0.9545).The findings validate the power of transfer learning for this domain and establish a strong performance benchmark.The introduced dataset provides a valuable resource for future research into developing robust and accurate character recognition systems.展开更多
In recent decades,the proliferation of email communication has markedly escalated,resulting in a concomitant surge in spam emails that congest networks and presenting security risks.This study introduces an innovative...In recent decades,the proliferation of email communication has markedly escalated,resulting in a concomitant surge in spam emails that congest networks and presenting security risks.This study introduces an innovative spam detection method utilizing the Horse Herd Optimization Algorithm(HHOA),designed for binary classification within multi⁃objective framework.The method proficiently identifies essential features,minimizing redundancy and improving classification precision.The suggested HHOA attained an impressive accuracy of 97.21%on the Kaggle email dataset,with precision of 94.30%,recall of 90.50%,and F1⁃score of 92.80%.Compared to conventional techniques,such as Support Vector Machine(93.89%accuracy),Random Forest(96.14%accuracy),and K⁃Nearest Neighbours(92.08%accuracy),HHOA exhibited enhanced performance with reduced computing complexity.The suggested method demonstrated enhanced feature selection efficiency,decreasing the number of selected features while maintaining high classification accuracy.The results underscore the efficacy of HHOA in spam identification and indicate its potential for further applications in practical email filtering systems.展开更多
Visual diagnosis of skin cancer is challenging due to subtle inter-class similarities,variations in skin texture,the presence of hair,and inconsistent illumination.Deep learning models have shown promise in assisting ...Visual diagnosis of skin cancer is challenging due to subtle inter-class similarities,variations in skin texture,the presence of hair,and inconsistent illumination.Deep learning models have shown promise in assisting early detection,yet their performance is often limited by the severe class imbalance present in dermoscopic datasets.This paper proposes CANNSkin,a skin cancer classification framework that integrates a convolutional autoencoder with latent-space oversampling to address this imbalance.The autoencoder is trained to reconstruct lesion images,and its latent embeddings are used as features for classification.To enhance minority-class representation,the Synthetic Minority Oversampling Technique(SMOTE)is applied directly to the latent vectors before classifier training.The encoder and classifier are first trained independently and later fine-tuned end-to-end.On the HAM10000 dataset,CANNSkin achieves an accuracy of 93.01%,a macro-F1 of 88.54%,and an ROC–AUC of 98.44%,demonstrating strong robustness across ten test subsets.Evaluation on the more complex ISIC 2019 dataset further confirms the model’s effectiveness,where CANNSkin achieves 94.27%accuracy,93.95%precision,94.09%recall,and 99.02%F1-score,supported by high reconstruction fidelity(PSNR 35.03 dB,SSIM 0.86).These results demonstrate the effectiveness of our proposed latent-space balancing and fine-tuned representation learning as a new benchmark method for robust and accurate skin cancer classification across heterogeneous datasets.展开更多
Arrhythmias are a frequently occurring phenomenon in clinical practice,but how to accurately dis-tinguish subtle rhythm abnormalities remains an ongoing difficulty faced by the entire research community when conductin...Arrhythmias are a frequently occurring phenomenon in clinical practice,but how to accurately dis-tinguish subtle rhythm abnormalities remains an ongoing difficulty faced by the entire research community when conducting ECG-based studies.From a review of existing studies,two main factors appear to contribute to this problem:the uneven distribution of arrhythmia classes and the limited expressiveness of features learned by current models.To overcome these limitations,this study proposes a dual-path multimodal framework,termed DM-EHC(Dual-Path Multimodal ECG Heartbeat Classifier),for ECG-based heartbeat classification.The proposed framework links 1D ECG temporal features with 2D time–frequency features.By setting up the dual paths described above,the model can process more dimensions of feature information.The MIT-BIH arrhythmia database was selected as the baseline dataset for the experiments.Experimental results show that the proposed method outperforms single modalities and performs better for certain specific types of arrhythmias.The model achieved mean precision,recall,and F1 score of 95.14%,92.26%,and 93.65%,respectively.These results indicate that the framework is robust and has potential value in automated arrhythmia classification.展开更多
Near-Earth objects are important not only in studying the early formation of the Solar System,but also because they pose a serious hazard to humanity when they make close approaches to the Earth.Study of their physica...Near-Earth objects are important not only in studying the early formation of the Solar System,but also because they pose a serious hazard to humanity when they make close approaches to the Earth.Study of their physical properties can provide useful information on their origin,evolution,and hazard to human beings.However,it remains challenging to investigate small,newly discovered,near-Earth objects because of our limited observational window.This investigation seeks to determine the visible colors of near-Earth asteroids(NEAs),perform an initial taxonomic classification based on visible colors and analyze possible correlations between the distribution of taxonomic classification and asteroid size or orbital parameters.Observations were performed in the broadband BVRI Johnson−Cousins photometric system,applied to images from the Yaoan High Precision Telescope and the 1.88 m telescope at the Kottamia Astronomical Observatory.We present new photometric observations of 84 near-Earth asteroids,and classify 80 of them taxonomically,based on their photometric colors.We find that nearly half(46.3%)of the objects in our sample can be classified as S-complex,26.3%as C-complex,6%as D-complex,and 15.0%as X-complex;the remaining belong to the A-or V-types.Additionally,we identify three P-type NEAs in our sample,according to the Tholen scheme.The fractional abundances of the C/X-complex members with absolute magnitude H≥17.0 were more than twice as large as those with H<17.0.However,the fractions of C-and S-complex members with diameters≤1 km and>1 km are nearly equal,while X-complex members tend to have sub-kilometer diameters.In our sample,the C/D-complex objects are predominant among those with a Jovian Tisserand parameter of T_(J)<3.1.These bodies could have a cometary origin.C-and S-complex members account for a considerable proportion of the asteroids that are potentially hazardous.展开更多
Graph Neural Networks(GNNs)have proven highly effective for graph classification across diverse fields such as social networks,bioinformatics,and finance,due to their capability to learn complex graph structures.Howev...Graph Neural Networks(GNNs)have proven highly effective for graph classification across diverse fields such as social networks,bioinformatics,and finance,due to their capability to learn complex graph structures.However,despite their success,GNNs remain vulnerable to adversarial attacks that can significantly degrade their classification accuracy.Existing adversarial attack strategies primarily rely on label information to guide the attacks,which limits their applicability in scenarios where such information is scarce or unavailable.This paper introduces an innovative unsupervised attack method for graph classification,which operates without relying on label information,thereby enhancing its applicability in a broad range of scenarios.Specifically,our method first leverages a graph contrastive learning loss to learn high-quality graph embeddings by comparing different stochastic augmented views of the graphs.To effectively perturb the graphs,we then introduce an implicit estimator that measures the impact of various modifications on graph structures.The proposed strategy identifies and flips edges with the top-K highest scores,determined by the estimator,to maximize the degradation of the model’s performance.In addition,to defend against such attack,we propose a lightweight regularization-based defense mechanism that is specifically tailored to mitigate the structural perturbations introduced by our attack strategy.It enhances model robustness by enforcing embedding consistency and edge-level smoothness during training.We conduct experiments on six public TU graph classification datasets:NCI1,NCI109,Mutagenicity,ENZYMES,COLLAB,and DBLP_v1,to evaluate the effectiveness of our attack and defense strategies.Under an attack budget of 3,the maximum reduction in model accuracy reaches 6.67%on the Graph Convolutional Network(GCN)and 11.67%on the Graph Attention Network(GAT)across different datasets,indicating that our unsupervised method induces degradation comparable to state-of-the-art supervised attacks.Meanwhile,our defense achieves the highest accuracy recovery of 3.89%(GCN)and 5.00%(GAT),demonstrating improved robustness against structural perturbations.展开更多
Skin diseases affect millions worldwide.Early detection is key to preventing disfigurement,lifelong disability,or death.Dermoscopic images acquired in primary-care settings show high intra-class visual similarity and ...Skin diseases affect millions worldwide.Early detection is key to preventing disfigurement,lifelong disability,or death.Dermoscopic images acquired in primary-care settings show high intra-class visual similarity and severe class imbalance,and occasional imaging artifacts can create ambiguity for state-of-the-art convolutional neural networks(CNNs).We frame skin lesion recognition as graph-based reasoning and,to ensure fair evaluation and avoid data leakage,adopt a strict lesion-level partitioning strategy.Each image is first over-segmented using SLIC(Simple Linear Iterative Clustering)to produce perceptually homogeneous superpixels.These superpixels form the nodes of a region-adjacency graph whose edges encode spatial continuity.Node attributes are 1280-dimensional embeddings extracted with a lightweight yet expressive EfficientNet-B0 backbone,providing strong representational power at modest computational cost.The resulting graphs are processed by a five-layer Graph Attention Network(GAT)that learns to weight inter-node relationships dynamically and aggregates multi-hop context before classifying lesions into seven classes with a log-softmax output.Extensive experiments on the DermaMNIST benchmark show the proposed pipeline achieves 88.35%accuracy and 98.04%AUC,outperforming contemporary CNNs,AutoML approaches,and alternative graph neural networks.An ablation study indicates EfficientNet-B0 produces superior node descriptors compared with ResNet-18 and DenseNet,and that roughly five GAT layers strike a good balance between being too shallow and over-deep while avoiding oversmoothing.The method requires no data augmentation or external metadata,making it a drop-in upgrade for clinical computer-aided diagnosis systems.展开更多
With the recent increase in data volume and diversity,traditional text representation techniques are struggling to capture context,particularly in environments with sparse data.To address these challenges,this study p...With the recent increase in data volume and diversity,traditional text representation techniques are struggling to capture context,particularly in environments with sparse data.To address these challenges,this study proposes a new model,the Masked Joint Representation Model(MJRM).MJRM approximates the original hypothesis by leveraging multiple elements in a limited context.It dynamically adapts to changes in characteristics based on data distribution through three main components.First,masking-based representation learning,termed selective dynamic masking,integrates topic modeling and sentiment clustering to generate and train multiple instances across different data subsets,whose predictions are then aggregated with optimized weights.This design alleviates sparsity,suppresses noise,and preserves contextual structures.Second,regularization-based improvements are applied.Third,techniques for addressing sparse data are used to perform final inference.As a result,MJRM improves performance by up to 4%compared to existing AI techniques.In our experiments,we analyzed the contribution of each factor,demonstrating that masking,dynamic learning,and aggregating multiple instances complement each other to improve performance.This demonstrates that a masking-based multi-learning strategy is effective for context-aware sparse text classification,and can be useful even in challenging situations such as data shortage or data distribution variations.We expect that the approach can be extended to diverse fields such as sentiment analysis,spam filtering,and domain-specific document classification.展开更多
基金Support by Sichuan Science and Technology Program[2023YFSY0026,2023YFH0004]Guangzhou Huashang University[2024HSZD01,HS2023JYSZH01].
文摘Graph neural networks(GNN)have shown strong performance in node classification tasks,yet most existing models rely on uniform or shared weight aggregation,lacking flexibility in modeling the varying strength of relationships among nodes.This paper proposes a novel graph coupling convolutional model that introduces an adaptive weighting mechanism to assign distinct importance to neighboring nodes based on their similarity to the central node.Unlike traditional methods,the proposed coupling strategy enhances the interpretability of node interactions while maintaining competitive classification performance.The model operates in the spatial domain,utilizing adjacency list structures for efficient convolution and addressing the limitations of weight sharing through a coupling-based similarity computation.Extensive experiments are conducted on five graph-structured datasets,including Cora,Citeseer,PubMed,Reddit,and BlogCatalog,as well as a custom topology dataset constructed from the Open University Learning Analytics Dataset(OULAD)educational platform.Results demonstrate that the proposed model achieves good classification accuracy,while significantly reducing training time through direct second-order neighbor fusion and data preprocessing.Moreover,analysis of neighborhood order reveals that considering third-order neighbors offers limited accuracy gains but introduces considerable computational overhead,confirming the efficiency of first-and second-order convolution in practical applications.Overall,the proposed graph coupling model offers a lightweight,interpretable,and effective framework for multi-label node classification in complex networks.
基金Supported by the Opening Fund of Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education (93K-17-2010-K02)the Opening Fund of Key Discipline of Computer Soft-Ware and Theory of Zhejiang Province at Zhejiang Normal University (ZSDZZZZXK05)
文摘Aiming at the problem of multi-label classification, a multi-label classification algorithm based on label-specific features is proposed in this paper. In this algorithm, we compute feature density on the positive and negative instances set of each class firstly and then select mk features of high density from the positive and negative instances set of each class, respectively; the intersec- tion is taken as the label-specific features of the corresponding class. Finally, multi-label data are classified on the basis of la- bel-specific features. The algorithm can show the label-specific features of each class. Experiments show that our proposed method, the MLSF algorithm, performs significantly better than the other state-of-the-art multi-label learning approaches.
基金supported by the National Key R&D Program of China(2018YFC0830200,Zhang,B,www.most.gov.cn)the Fundamental Research Funds for the Central Universities(2242018S30021 and 2242017S30023,Zhou S,www.seu.edu.cn)the Open Research Fund from Key Laboratory of Computer Network and Information Integration In Southeast University,Ministry of Education,China(3209012001C3,Zhang B,www.seu.edu.cn).
文摘The Internet of Medical Things(IoMT)will come to be of great importance in the mediation of medical disputes,as it is emerging as the core of intelligent medical treatment.First,IoMT can track the entire medical treatment process in order to provide detailed trace data in medical dispute resolution.Second,IoMT can infiltrate the ongoing treatment and provide timely intelligent decision support to medical staff.This information includes recommendation of similar historical cases,guidance for medical treatment,alerting of hired dispute profiteers etc.The multi-label classification of medical dispute documents(MDDs)plays an important role as a front-end process for intelligent decision support,especially in the recommendation of similar historical cases.However,MDDs usually appear as long texts containing a large amount of redundant information,and there is a serious distribution imbalance in the dataset,which directly leads to weaker classification performance.Accordingly,in this paper,a multi-label classification method based on key sentence extraction is proposed for MDDs.The method is divided into two parts.First,the attention-based hierarchical bi-directional long short-term memory(BiLSTM)model is used to extract key sentences from documents;second,random comprehensive sampling Bagging(RCS-Bagging),which is an ensemble multi-label classification model,is employed to classify MDDs based on key sentence sets.The use of this approach greatly improves the classification performance.Experiments show that the performance of the two models proposed in this paper is remarkably better than that of the baseline methods.
基金the Natural Science Foundation of China(Grant Numbers 72074014 and 72004012).
文摘Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on the benchmark datasets have been proposed for multi-label classification task in the literature.Furthermore,several open-source tools implementing these approaches have also been developed.However,the characteristics of real-world multi-label patent and publication datasets are not completely in line with those of benchmark ones.Therefore,the main purpose of this paper is to evaluate comprehensively seven multi-label classification methods on real-world datasets.Research limitations:Three real-world datasets differ in the following aspects:statement,data quality,and purposes.Additionally,open-source tools designed for multi-label classification also have intrinsic differences in their approaches for data processing and feature selection,which in turn impacts the performance of a multi-label classification approach.In the near future,we will enhance experimental precision and reinforce the validity of conclusions by employing more rigorous control over variables through introducing expanded parameter settings.Practical implications:The observed Macro F1 and Micro F1 scores on real-world datasets typically fall short of those achieved on benchmark datasets,underscoring the complexity of real-world multi-label classification tasks.Approaches leveraging deep learning techniques offer promising solutions by accommodating the hierarchical relationships and interdependencies among labels.With ongoing enhancements in deep learning algorithms and large-scale models,it is expected that the efficacy of multi-label classification tasks will be significantly improved,reaching a level of practical utility in the foreseeable future.Originality/value:(1)Seven multi-label classification methods are comprehensively compared on three real-world datasets.(2)The TextCNN and TextRCNN models perform better on small-scale datasets with more complex hierarchical structure of labels and more balanced document-label distribution.(3)The MLkNN method works better on the larger-scale dataset with more unbalanced document-label distribution.
基金supported by the National Key Research and Development Program of China(Grant No.2020YFA0211400)the State Key Program of the National Natural Science of China(Grant No.11834008)+2 种基金the National Natural Science Foundation of China(Grant Nos.12174192,12174188,and 11974176)the State Key Laboratory of Acoustics,Chinese Academy of Sciences(Grant No.SKLA202410)the Fund from the Key Laboratory of Underwater Acoustic Environment,Chinese Academy of Sciences(Grant No.SSHJ-KFKT-1701).
文摘Phononic crystals,as artificial composite materials,have sparked significant interest due to their novel characteristics that emerge upon the introduction of nonlinearity.Among these properties,second-harmonic features exhibit potential applications in acoustic frequency conversion,non-reciprocal wave propagation,and non-destructive testing.Precisely manipulating the harmonic band structure presents a major challenge in the design of nonlinear phononic crystals.Traditional design approaches based on parameter adjustments to meet specific application requirements are inefficient and often yield suboptimal performance.Therefore,this paper develops a design methodology using Softmax logistic regression and multi-label classification learning to inversely design the material distribution of nonlinear phononic crystals by exploiting information from harmonic transmission spectra.The results demonstrate that the neural network-based inverse design method can effectively tailor nonlinear phononic crystals with desired functionalities.This work establishes a mapping relationship between the band structure and the material distribution within phononic crystals,providing valuable insights into the inverse design of metamaterials.
基金the National Natural Science Foundation of China(No.61472161)the Science&Technology Development Project of Jilin Province(Nos.20180101334JC and 20160520099JH)。
文摘Sum-product networks(SPNs)are an expressive deep probabilistic architecture with solid theoretical foundations,which allows tractable and exact inference.SPNs always act as black-box inference machine in many artificial intelligence tasks.Due to their recursive definition,SPNs can also be naturally employed as hierarchical feature extractors.Recently,SPNs have been successfully employed as autoencoder framework in representation learning.However,SPNs autoencoder ignores the model structural duality and trains the models separately and independently.In this work,we propose a Dual-SPNs autoencoder which designs two SPNs autoencoders to compose as a dual form.This approach trains the models simultaneously,and explicitly exploits the structural duality between them to enhance the training process.Experimental results on several multilabel classification problems demonstrate that Dual-SPNs autoencoder is very competitive against with state-of-the-art autoencoder architectures.
基金This article was funded by the National Natural Science Foundation of China(Nos.61702270,41971343)the China Postdoctoral Science Foundation(No.2017M621592).
文摘In the last few years,smartphone usage and driver sleepiness have been unanimously considered to lead to numerous road accidents,which causes many scholars to pay attention to autonomous driving.For this complexity scene,one of the major challenges is mining information comprehensively from massive features in vehicle video.This paper proposes a multi-label classification method MCM-VV(Multi-label Classification Method for Vehicle Video)for vehicle video to judge the label of road condition for unmanned system.Method MCM-VV includes a process of feature extraction and a process of multi-label classification.During feature extraction,grayscale,lane line and the edge of main object are extracted after video preprocessing.During the multi-label classification,the algorithm DR-ML-KNN(Multi-label K-nearest Neighbor Classification Algorithm based on Dimensionality Reduction)learns the training set to obtain multi-label classifier,then predicts the label of road condition according to maximum a posteriori principle,finally outputs labels and adds the new instance to training set for the optimization of classifier.Experimental results on five vehicle video datasets show that the method MCM-VV is effective and efficient.The DR-ML-KNN algorithm reduces the runtime by 50%.It also reduces the time complexity and improves the accuracy.
基金supported in part by the National Natural Science Foundation of China(61379049,61772120)
文摘Multi-label learning deals with data associated with a set of labels simultaneously. Dimensionality reduction is an important but challenging task in multi-label learning. Feature selection is an efficient technique for dimensionality reduction to search an optimal feature subset preserving the most relevant information. In this paper, we propose an effective feature evaluation criterion for multi-label feature selection, called neighborhood relationship preserving score. This criterion is inspired by similarity preservation, which is widely used in single-label feature selection. It evaluates each feature subset by measuring its capability in preserving neighborhood relationship among samples. Unlike similarity preservation, we address the order of sample similarities which can well express the neighborhood relationship among samples, not just the pairwise sample similarity. With this criterion, we also design one ranking algorithm and one greedy algorithm for feature selection problem. The proposed algorithms are validated in six publicly available data sets from machine learning repository. Experimental results demonstrate their superiorities over the compared state-of-the-art methods.
基金Supported by Australian Research Council Discovery(DP130102691)the National Science Foundation of China(61302157)+1 种基金China National 863 Project(2012AA12A308)China Pre-research Project of Nuclear Industry(FZ1402-08)
文摘In this paper a novel coupled attribute similarity learning method is proposed with the basis on the multi-label categorical data（CASonMLCD）.The CASonMLCD method not only computes the correlations between different attributes and multi-label sets using information gain,which can be regarded as the important degree of each attribute in the attribute learning method,but also further analyzes the intra-coupled and inter-coupled interactions between an attribute value pair for different attributes and multiple labels.The paper compared the CASonMLCD method with the OF distance and Jaccard similarity,which is based on the MLKNN algorithm according to 5common evaluation criteria.The experiment results demonstrated that the CASonMLCD method can mine the similarity relationship more accurately and comprehensively,it can obtain better performance than compared methods.
基金The authors would like to thank the Deanship of Scientific Research at UmmAl-Qura University for supporting this work by Grant Code: (22UQU4340018DSR02).
文摘Multitasking and multioutput neural networks models jointly learn related classification tasks from a shared structure.Hard parameters sharing is a multitasking approach that shares hidden layers between multiple taskspecific outputs.The output layers’weights are essential in transforming aggregated neurons outputs into tasks labels.This paper redirects the multioutput network research to prove that the ensemble of output layers prediction can improve network performance in classifying multi-label classification tasks.The network’s output layers initialized with different weights simulate multiple semi-independent classifiers that can make non-identical label sets predictions for the same instance.The ensemble of a multi-output neural network that learns to classify the same multi-label classification task per output layer can outperform an individual output layer neural network.We propose an ensemble strategy of output layers components in the multi-output neural network for multi-label classification(ENSOCOM).The baseline and proposed models are selected based on the size of the hidden layer and the number of output layers to evaluate the proposed method comprehensively.The ENSOCOM method improved the performance of the neural networks on five different multi-label datasets based on several evaluation metrics.The methods presented in this work can substitute the standard labels representation and predictions generation of any neural network.
基金would like to thank the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support(QU-APC-2025).
文摘Automated and accurate movie genre classification is crucial for content organization,recommendation systems,and audience targeting in the film industry.Although most existing approaches focus on audiovisual features such as trailers and posters,the text-based classification remains underexplored despite its accessibility and semantic richness.This paper introduces the Genre Attention Model(GAM),a deep learning architecture that integrates transformer models with a hierarchical attention mechanism to extract and leverage contextual information from movie plots formulti-label genre classification.In order to assess its effectiveness,we assessmultiple transformer-based models,including Bidirectional Encoder Representations fromTransformers(BERT),ALite BERT(ALBERT),Distilled BERT(DistilBERT),Robustly Optimized BERT Pretraining Approach(RoBERTa),Efficiently Learning an Encoder that Classifies Token Replacements Accurately(ELECTRA),eXtreme Learning Network(XLNet)and Decodingenhanced BERT with Disentangled Attention(DeBERTa).Experimental results demonstrate the superior performance of DeBERTa-based GAM,which employs a two-tier hierarchical attention mechanism:word-level attention highlights key terms,while sentence-level attention captures critical narrative segments,ensuring a refined and interpretable representation of movie plots.Evaluated on three benchmark datasets Trailers12K,Large Movie Trailer Dataset-9(LMTD-9),and MovieLens37K.GAM achieves micro-average precision scores of 83.63%,83.32%,and 83.34%,respectively,surpassing state-of-the-artmodels.Additionally,GAMis computationally efficient,requiring just 6.10Giga Floating Point Operations Per Second(GFLOPS),making it a scalable and cost-effective solution.These results highlight the growing potential of text-based deep learning models in genre classification and GAM’s effectiveness in improving predictive accuracy while maintaining computational efficiency.With its robust performance,GAM offers a versatile and scalable framework for content recommendation,film indexing,and media analytics,providing an interpretable alternative to traditional audiovisual-based classification techniques.
基金supporting of the National Science and Technology Council NSTC(grant nos.NSTC 112-2221-E-019-023,NSTC 113-2221-E-019-039)Taiwan University of Science and Technology.
文摘In its 2023 global health statistics,the World Health Organization noted that noncommunicable diseases(NCDs)remain the leading cause of disease burden worldwide,with cardiovascular diseases(CVDs)resulting in more deaths than the three other major NCDs combined.In this study,we developed a method that can comprehensively detect which CVDs are present in a patient.Specifically,we propose a multi-label classification method that utilizes photoplethysmography(PPG)signals and physiological characteristics from public datasets to classify four types of CVDs and related conditions:hypertension,diabetes,cerebral infarction,and cerebrovascular disease.Our approach to multi-disease classification of cardiovascular diseases(CVDs)using PPG signals achieves the highest classification performance when encompassing the broadest range of disease categories,thereby offering a more comprehensive assessment of human health.We employ a multi-label classification strategy to simultaneously predict the presence or absence of multiple diseases.Specifically,we first apply the Savitzky-Golay(S-G)filter to the PPG signals to reduce noise and then transform into statistical features.We integrate processed PPG signals with individual physiological features as a multimodal input,thereby expanding the learned feature space.Notably,even with a simple machine learning method,this approach can achieve relatively high accuracy.The proposed method achieved a maximum F1-score of 0.91,minimum Hamming loss of 0.04,and an accuracy of 0.95.Thus,our method represents an effective and rapid solution for detecting multiple diseases simultaneously,which is beneficial for comprehensively managing CVDs.
文摘Automated cartoon character recognition is crucial for applications in content indexing,filtering,and copyright protection,yet it faces a significant challenge in animated media due to high intra-class visual variability,where characters frequently alter their appearance.To address this problem,we introduce the novel Kral Sakir dataset,a public benchmark of 16,725 images specifically curated for the task of multi-label cartoon character classification under these varied conditions.This paper conducts a comprehensive benchmark study,evaluating the performance of state-of-the-art pretrained Convolutional Neural Networks(CNNs),including DenseNet,ResNet,and VGG,against a custom baseline model trained from scratch.Our experiments,evaluated using metrics of F1-Score,accuracy,and Area Under the ROC Curve(AUC),demonstrate that fine-tuning pretrained models is a highly effective strategy.The best-performing model,DenseNet121,achieved an F1-Score of 0.9890 and an accuracy of 0.9898,significantly outperforming our baseline CNN(F1-Score of 0.9545).The findings validate the power of transfer learning for this domain and establish a strong performance benchmark.The introduced dataset provides a valuable resource for future research into developing robust and accurate character recognition systems.
文摘In recent decades,the proliferation of email communication has markedly escalated,resulting in a concomitant surge in spam emails that congest networks and presenting security risks.This study introduces an innovative spam detection method utilizing the Horse Herd Optimization Algorithm(HHOA),designed for binary classification within multi⁃objective framework.The method proficiently identifies essential features,minimizing redundancy and improving classification precision.The suggested HHOA attained an impressive accuracy of 97.21%on the Kaggle email dataset,with precision of 94.30%,recall of 90.50%,and F1⁃score of 92.80%.Compared to conventional techniques,such as Support Vector Machine(93.89%accuracy),Random Forest(96.14%accuracy),and K⁃Nearest Neighbours(92.08%accuracy),HHOA exhibited enhanced performance with reduced computing complexity.The suggested method demonstrated enhanced feature selection efficiency,decreasing the number of selected features while maintaining high classification accuracy.The results underscore the efficacy of HHOA in spam identification and indicate its potential for further applications in practical email filtering systems.
基金supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University(IMSIU)(grant number IMSIU-DDRSP2601).
文摘Visual diagnosis of skin cancer is challenging due to subtle inter-class similarities,variations in skin texture,the presence of hair,and inconsistent illumination.Deep learning models have shown promise in assisting early detection,yet their performance is often limited by the severe class imbalance present in dermoscopic datasets.This paper proposes CANNSkin,a skin cancer classification framework that integrates a convolutional autoencoder with latent-space oversampling to address this imbalance.The autoencoder is trained to reconstruct lesion images,and its latent embeddings are used as features for classification.To enhance minority-class representation,the Synthetic Minority Oversampling Technique(SMOTE)is applied directly to the latent vectors before classifier training.The encoder and classifier are first trained independently and later fine-tuned end-to-end.On the HAM10000 dataset,CANNSkin achieves an accuracy of 93.01%,a macro-F1 of 88.54%,and an ROC–AUC of 98.44%,demonstrating strong robustness across ten test subsets.Evaluation on the more complex ISIC 2019 dataset further confirms the model’s effectiveness,where CANNSkin achieves 94.27%accuracy,93.95%precision,94.09%recall,and 99.02%F1-score,supported by high reconstruction fidelity(PSNR 35.03 dB,SSIM 0.86).These results demonstrate the effectiveness of our proposed latent-space balancing and fine-tuned representation learning as a new benchmark method for robust and accurate skin cancer classification across heterogeneous datasets.
基金supported by the Innovative Human Resource Development for Local Intel-lectualization program through the Institute of Information&Communications Technology Planning&Evaluation(IITP)grant funded by the Korea government(MSIT)(No.IITP-2026-2020-0-01741)the research fund of Hanyang University(HY-2025-1110).
文摘Arrhythmias are a frequently occurring phenomenon in clinical practice,but how to accurately dis-tinguish subtle rhythm abnormalities remains an ongoing difficulty faced by the entire research community when conducting ECG-based studies.From a review of existing studies,two main factors appear to contribute to this problem:the uneven distribution of arrhythmia classes and the limited expressiveness of features learned by current models.To overcome these limitations,this study proposes a dual-path multimodal framework,termed DM-EHC(Dual-Path Multimodal ECG Heartbeat Classifier),for ECG-based heartbeat classification.The proposed framework links 1D ECG temporal features with 2D time–frequency features.By setting up the dual paths described above,the model can process more dimensions of feature information.The MIT-BIH arrhythmia database was selected as the baseline dataset for the experiments.Experimental results show that the proposed method outperforms single modalities and performs better for certain specific types of arrhythmias.The model achieved mean precision,recall,and F1 score of 95.14%,92.26%,and 93.65%,respectively.These results indicate that the framework is robust and has potential value in automated arrhythmia classification.
基金funded by the China National Space Administration(KJSP2023020105)supported by the National Key R&D Program of China(Grant No.2023YFA1608100)+2 种基金the NSFC(Grant No.62227901)the Minor Planet Foundationsupported by the Egyptian Science,Technology&Innovation Funding Authority(STDF)under Grant No.48102.
文摘Near-Earth objects are important not only in studying the early formation of the Solar System,but also because they pose a serious hazard to humanity when they make close approaches to the Earth.Study of their physical properties can provide useful information on their origin,evolution,and hazard to human beings.However,it remains challenging to investigate small,newly discovered,near-Earth objects because of our limited observational window.This investigation seeks to determine the visible colors of near-Earth asteroids(NEAs),perform an initial taxonomic classification based on visible colors and analyze possible correlations between the distribution of taxonomic classification and asteroid size or orbital parameters.Observations were performed in the broadband BVRI Johnson−Cousins photometric system,applied to images from the Yaoan High Precision Telescope and the 1.88 m telescope at the Kottamia Astronomical Observatory.We present new photometric observations of 84 near-Earth asteroids,and classify 80 of them taxonomically,based on their photometric colors.We find that nearly half(46.3%)of the objects in our sample can be classified as S-complex,26.3%as C-complex,6%as D-complex,and 15.0%as X-complex;the remaining belong to the A-or V-types.Additionally,we identify three P-type NEAs in our sample,according to the Tholen scheme.The fractional abundances of the C/X-complex members with absolute magnitude H≥17.0 were more than twice as large as those with H<17.0.However,the fractions of C-and S-complex members with diameters≤1 km and>1 km are nearly equal,while X-complex members tend to have sub-kilometer diameters.In our sample,the C/D-complex objects are predominant among those with a Jovian Tisserand parameter of T_(J)<3.1.These bodies could have a cometary origin.C-and S-complex members account for a considerable proportion of the asteroids that are potentially hazardous.
基金funded by the National Key Research and Development Program of China(Grant No.2024YFE0209000)the NSFC(Grant No.U23B2019).
文摘Graph Neural Networks(GNNs)have proven highly effective for graph classification across diverse fields such as social networks,bioinformatics,and finance,due to their capability to learn complex graph structures.However,despite their success,GNNs remain vulnerable to adversarial attacks that can significantly degrade their classification accuracy.Existing adversarial attack strategies primarily rely on label information to guide the attacks,which limits their applicability in scenarios where such information is scarce or unavailable.This paper introduces an innovative unsupervised attack method for graph classification,which operates without relying on label information,thereby enhancing its applicability in a broad range of scenarios.Specifically,our method first leverages a graph contrastive learning loss to learn high-quality graph embeddings by comparing different stochastic augmented views of the graphs.To effectively perturb the graphs,we then introduce an implicit estimator that measures the impact of various modifications on graph structures.The proposed strategy identifies and flips edges with the top-K highest scores,determined by the estimator,to maximize the degradation of the model’s performance.In addition,to defend against such attack,we propose a lightweight regularization-based defense mechanism that is specifically tailored to mitigate the structural perturbations introduced by our attack strategy.It enhances model robustness by enforcing embedding consistency and edge-level smoothness during training.We conduct experiments on six public TU graph classification datasets:NCI1,NCI109,Mutagenicity,ENZYMES,COLLAB,and DBLP_v1,to evaluate the effectiveness of our attack and defense strategies.Under an attack budget of 3,the maximum reduction in model accuracy reaches 6.67%on the Graph Convolutional Network(GCN)and 11.67%on the Graph Attention Network(GAT)across different datasets,indicating that our unsupervised method induces degradation comparable to state-of-the-art supervised attacks.Meanwhile,our defense achieves the highest accuracy recovery of 3.89%(GCN)and 5.00%(GAT),demonstrating improved robustness against structural perturbations.
基金funded by the Deanship of Graduate Studies and Scientific Research at Jouf University under grant No.(DGSSR-2025-02-01296).
文摘Skin diseases affect millions worldwide.Early detection is key to preventing disfigurement,lifelong disability,or death.Dermoscopic images acquired in primary-care settings show high intra-class visual similarity and severe class imbalance,and occasional imaging artifacts can create ambiguity for state-of-the-art convolutional neural networks(CNNs).We frame skin lesion recognition as graph-based reasoning and,to ensure fair evaluation and avoid data leakage,adopt a strict lesion-level partitioning strategy.Each image is first over-segmented using SLIC(Simple Linear Iterative Clustering)to produce perceptually homogeneous superpixels.These superpixels form the nodes of a region-adjacency graph whose edges encode spatial continuity.Node attributes are 1280-dimensional embeddings extracted with a lightweight yet expressive EfficientNet-B0 backbone,providing strong representational power at modest computational cost.The resulting graphs are processed by a five-layer Graph Attention Network(GAT)that learns to weight inter-node relationships dynamically and aggregates multi-hop context before classifying lesions into seven classes with a log-softmax output.Extensive experiments on the DermaMNIST benchmark show the proposed pipeline achieves 88.35%accuracy and 98.04%AUC,outperforming contemporary CNNs,AutoML approaches,and alternative graph neural networks.An ablation study indicates EfficientNet-B0 produces superior node descriptors compared with ResNet-18 and DenseNet,and that roughly five GAT layers strike a good balance between being too shallow and over-deep while avoiding oversmoothing.The method requires no data augmentation or external metadata,making it a drop-in upgrade for clinical computer-aided diagnosis systems.
基金supported by the SungKyunKwan University and the BK21 FOUR(Graduate School Innovation)funded by the Ministry of Education(MOE,Korea)and National Research Foundation of Korea(NRF).
文摘With the recent increase in data volume and diversity,traditional text representation techniques are struggling to capture context,particularly in environments with sparse data.To address these challenges,this study proposes a new model,the Masked Joint Representation Model(MJRM).MJRM approximates the original hypothesis by leveraging multiple elements in a limited context.It dynamically adapts to changes in characteristics based on data distribution through three main components.First,masking-based representation learning,termed selective dynamic masking,integrates topic modeling and sentiment clustering to generate and train multiple instances across different data subsets,whose predictions are then aggregated with optimized weights.This design alleviates sparsity,suppresses noise,and preserves contextual structures.Second,regularization-based improvements are applied.Third,techniques for addressing sparse data are used to perform final inference.As a result,MJRM improves performance by up to 4%compared to existing AI techniques.In our experiments,we analyzed the contribution of each factor,demonstrating that masking,dynamic learning,and aggregating multiple instances complement each other to improve performance.This demonstrates that a masking-based multi-learning strategy is effective for context-aware sparse text classification,and can be useful even in challenging situations such as data shortage or data distribution variations.We expect that the approach can be extended to diverse fields such as sentiment analysis,spam filtering,and domain-specific document classification.
