In June 2012,the UN conference on sustainable development "Rio+20" summarized the work of the world community in this direction for last 20 years and outlined the tasks for the future.The UN website contains...In June 2012,the UN conference on sustainable development "Rio+20" summarized the work of the world community in this direction for last 20 years and outlined the tasks for the future.The UN website contains enough information to estimate the importance of global problems and "green economy," taking into account a very complicated state of the global and Russian markets which are balancing on the verge of crisis.Unfortunately,the website does not contain the materials of the 6th civilization forum "Long-Term Strategy for Sustainable Development on the Basis of Partnership of Civilizations:Concepts,Strategy,Programs and Projects" within the bounds of "Rio+20" which has considered the problems of a dialogue and partnership in conditions of extensive globalization.These problems are covered in the Partnership of Civilizations Journal which is issued in the Russian,English and Arabian languages,including its Internet version.The International Informatization Academy (it has the general consultative status at the Economic and Social Council of the UN) and its 20-year activity in the sphere of informatization of the world and Russia on the way to the partnership of civilizations,are presented there[1].展开更多
The rapidly evolving cybersecurity threat landscape exposes a critical flaw in traditional educational programs where static curricula cannot adapt swiftly to novel attack vectors.This creates a significant gap betwee...The rapidly evolving cybersecurity threat landscape exposes a critical flaw in traditional educational programs where static curricula cannot adapt swiftly to novel attack vectors.This creates a significant gap between theoretical knowledge and the practical defensive capabilities needed in the field.To address this,we propose TeachSecure-CTI,a novel framework for adaptive cybersecurity curriculumgeneration that integrates real-time Cyber Threat Intelligence(CTI)with AI-driven personalization.Our framework employs a layered architecture featuring a CTI ingestion and clusteringmodule,natural language processing for semantic concept extraction,and a reinforcement learning agent for adaptive content sequencing.Bydynamically aligning learningmaterialswithboththe evolving threat environment and individual learner profiles,TeachSecure-CTI ensures content remains current,relevant,and tailored.A 12-week study with 150 students across three institutions demonstrated that the framework improves learning gains by 34%,significantly exceeding the 12%–21%reported in recent literature.The system achieved 84.8%personalization accuracy,85.9%recognition accuracy for MITRE ATT&CK tactics,and a 31%faster competency development rate compared to static curricula.These findings have implications beyond academia,extending to workforce development,cyber range training,and certification programs.By bridging the gap between dynamic threats and static educational materials,TeachSecure-CTI offers an empirically validated,scalable solution for cultivating cybersecurity professionals capable of responding to modern threats.展开更多
Activation pruning reduces neural network complexity by eliminating low-importance neuron activations,yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally...Activation pruning reduces neural network complexity by eliminating low-importance neuron activations,yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally expensive and typically requires exhaustive search.We introduce a thermodynamics-inspired framework that treats activation distributions as energy-filtered physical systems and employs the free energy of activations as a principled evaluation metric.Phase-transition-like phenomena in the free-energy profile—such as extrema,inflection points,and curvature changes—yield reliable estimates of the critical pruning threshold,providing a theoretically grounded means of predicting sharp accuracy degradation.To further enhance efficiency,we propose a renormalized free energy technique that approximates full-evaluation free energy using only the activation distribution of the unpruned network.This eliminates repeated forward passes,dramatically reducing computational overhead and achieving speedups of up to 550×for MLPs.Extensive experiments across diverse vision architectures(MLP,CNN,ResNet,MobileNet,Vision Transformer)and text models(LSTM,BERT,ELECTRA,T5,GPT-2)on multiple datasets validate the generality,robustness,and computational efficiency of our approach.Overall,this work establishes a theoretically grounded and practically effective framework for activation pruning,bridging the gap between analytical understanding and efficient deployment of sparse neural networks.展开更多
This work addresses optimality aspects related to composite laminates having layers with different orientations.RegressionNeuralNetworks can model the mechanical behavior of these laminates,specifically the stressstra...This work addresses optimality aspects related to composite laminates having layers with different orientations.RegressionNeuralNetworks can model the mechanical behavior of these laminates,specifically the stressstrain relationship.If this model has strong generalization ability,it can be coupled with a metaheuristic algorithm–the PSO algorithm used in this article–to address an optimization problem(OP)related to the orientations of composite laminates.To solve OPs,this paper proposes an optimization framework(OFW)that connects the two components,the optimal solution search mechanism and the RNN model.The OFW has two modules:the search mechanism(Adaptive Hybrid Topology PSO)and the Prediction and Computation Module(PCM).The PCM undertakes all the activities concerning the OP at hand:the stress-strain model,constraints checking,and computation of the objective function.Two case studies about the layers’orientations of laminated specimens are conducted to validate the proposed framework.The specimens belong to“Off-axis oriented specimens”and are subjects of two OPs.The algorithms for AHTPSO and for the two PCMs(one for each problem)are proposed and implemented by MATLAB scripts and functions.Simulations are carried out for different initial conditions.The solutions demonstrated that the OFW is effective and has a highly acceptable computational complexity.The limitation of using the OFWis the generalization ability of the RNN model or any other regression models.To harness the RNN model efficiently,it must have a very good generalization power.If this condition ismet,the OFWcan be integrated into any design process to make optimal choices of the layers’orientations.展开更多
End-to-end Temporal Action Detection(TAD)has achieved remarkable progress in recent years,driven by innovations in model architectures and the emergence of Video Foundation Models(VFMs).However,existing TAD methods th...End-to-end Temporal Action Detection(TAD)has achieved remarkable progress in recent years,driven by innovations in model architectures and the emergence of Video Foundation Models(VFMs).However,existing TAD methods that perform full fine-tuning of pretrained video models often incur substantial computational costs,which become particularly pronounced when processing long video sequences.Moreover,the need for precise temporal boundary annotations makes data labeling extremely expensive.In low-resource settings where annotated samples are scarce,direct fine-tuning tends to cause overfitting.To address these challenges,we introduce Dynamic LowRank Adapter(DyLoRA),a lightweight fine-tuning framework tailored specifically for the TAD task.Built upon the Low-Rank Adaptation(LoRA)architecture,DyLoRA adapts only the key layers of the pretrained model via low-rank decomposition,reducing the number of trainable parameters to less than 5%of full fine-tuning methods.This significantly lowers memory consumption and mitigates overfitting in low-resource settings.Notably,DyLoRA enhances the temporal modeling capability of pretrained models by optimizing temporal dimension weights,thereby alleviating the representation misalignment of temporal features.Experimental results demonstrate that DyLoRA-TAD achieves impressive performance,with 73.9%mAP on THUMOS14,39.52%on ActivityNet-1.3,and 28.2%on Charades,substantially surpassing the best traditional feature-based methods.展开更多
Customer churn is the rate at which customers discontinue doing business with a company over a given time period.It is an essential measure for businesses to monitor high churn rates,as they often indicate underlying ...Customer churn is the rate at which customers discontinue doing business with a company over a given time period.It is an essential measure for businesses to monitor high churn rates,as they often indicate underlying issues with services,products,or customer experience,resulting in considerable income loss.Prediction of customer churn is a crucial task aimed at retaining customers and maintaining revenue growth.Traditional machine learning(ML)models often struggle to capture complex temporal dependencies in client behavior data.To address this,an optimized deep learning(DL)approach using a Regularized Bidirectional Long Short-Term Memory(RBiLSTM)model is proposed to mitigate overfitting and improve generalization error.The model integrates dropout,L2-regularization,and early stopping to enhance predictive accuracy while preventing over-reliance on specific patterns.Moreover,this study investigates the effect of optimization techniques on boosting the training efficiency of the developed model.Experimental results on a recent public customer churn dataset demonstrate that the trained model outperforms the traditional ML models and some other DL models,such as Long Short-Term Memory(LSTM)and Deep Neural Network(DNN),in churn prediction performance and stability.The proposed approach achieves 96.1%accuracy,compared with LSTM and DNN,which attain 94.5%and 94.1%accuracy,respectively.These results confirm that the proposed approach can be used as a valuable tool for businesses to identify at-risk consumers proactively and implement targeted retention strategies.展开更多
Alzheimer’s disease(AD)is the most common form of dementia.In addition to the lack of effective treatments,there are limitations in diagnostic capabilities.The complexity of AD itself,together with a variety of other...Alzheimer’s disease(AD)is the most common form of dementia.In addition to the lack of effective treatments,there are limitations in diagnostic capabilities.The complexity of AD itself,together with a variety of other diseases often observed in a patient’s history in addition to their AD diagnosis,make deciphering the molecular mechanisms that underlie AD,even more important.Large datasets of single-cell RNA sequencing,single-nucleus RNA-sequencing(snRNA-seq),and spatial transcriptomics(ST)have become essential in guiding and supporting new investigations into the cellular and regional susceptibility of AD.However,with unique technology,software,and larger databases emerging;a lack of integration of these data can contribute to ineffective use of valuable knowledge.Importantly,there was no specialized database that concentrates on ST in AD that offers comprehensive differential analyses under various conditions,such as sex-specific,region-specific,and comparisons between AD and control groups until the new Single-cell and Spatial RNA-seq databasE for Alzheimer’s Disease(ssREAD)database(Wang et al.,2024)was introduced to meet the scientific community’s growing demand for comprehensive,integrated,and accessible data analysis.展开更多
Objective expertise evaluation of individuals,as a prerequisite stage for team formation,has been a long-term desideratum in large software development companies.With the rapid advancements in machine learning methods...Objective expertise evaluation of individuals,as a prerequisite stage for team formation,has been a long-term desideratum in large software development companies.With the rapid advancements in machine learning methods,based on reliable existing data stored in project management tools’datasets,automating this evaluation process becomes a natural step forward.In this context,our approach focuses on quantifying software developer expertise by using metadata from the task-tracking systems.For this,we mathematically formalize two categories of expertise:technology-specific expertise,which denotes the skills required for a particular technology,and general expertise,which encapsulates overall knowledge in the software industry.Afterward,we automatically classify the zones of expertise associated with each task a developer has worked on using Bidirectional Encoder Representations from Transformers(BERT)-like transformers to handle the unique characteristics of project tool datasets effectively.Finally,our method evaluates the proficiency of each software specialist across already completed projects from both technology-specific and general perspectives.The method was experimentally validated,yielding promising results.展开更多
The effect of plasma and charged particle interaction with spacecraft in a low Earth orbit(LEO)environment leads to ion focusing and the formation of an ion void in the downstream region as a result of charging.Simula...The effect of plasma and charged particle interaction with spacecraft in a low Earth orbit(LEO)environment leads to ion focusing and the formation of an ion void in the downstream region as a result of charging.Simulations and investigations using a fixed potential imposed on the spacecraft showed the nonsignificance of geophysical parameter changes to ion focusing.Variation of the temperature ratio(T_(r))contributed only to local ion focusing and manifested as two-ion streamers dispersed at the upper and lower edges of the spacecraft-the outermost layers of the satellite structure at the top and bottom,respectively.A simulation involving changing the ambient plasma density(N_(p))also showed the formation of local ion focusing,in which ions were more concentrated as the density increased.Furthermore,auroral electron density(N_(ae))variation had no clear impact on ion focusing,as indicated by static two-ion structures in the wake field.However,variation of the object potential(ϕ)strongly affected ion focusing formation,leading to distortion of the initial ion void region behind the spacecraft.The formation of ion focusing in this study was subject to the electric field produced by the object potential and the ambipolar electric field resulting from plasma expansion in the downstream region.展开更多
Weakly Supervised Semantic Segmentation(WSSS),which relies only on image-level labels,has attracted significant attention for its cost-effectiveness and scalability.Existing methods mainly enhance inter-class distinct...Weakly Supervised Semantic Segmentation(WSSS),which relies only on image-level labels,has attracted significant attention for its cost-effectiveness and scalability.Existing methods mainly enhance inter-class distinctions and employ data augmentation to mitigate semantic ambiguity and reduce spurious activations.However,they often neglect the complex contextual dependencies among image patches,resulting in incomplete local representations and limited segmentation accuracy.To address these issues,we propose the Context Patch Fusion with Class Token Enhancement(CPF-CTE)framework,which exploits contextual relations among patches to enrich feature repre-sentations and improve segmentation.At its core,the Contextual-Fusion Bidirectional Long Short-Term Memory(CF-BiLSTM)module captures spatial dependencies between patches and enables bidirectional information flow,yield-ing a more comprehensive understanding of spatial correlations.This strengthens feature learning and segmentation robustness.Moreover,we introduce learnable class tokens that dynamically encode and refine class-specific semantics,enhancing discriminative capability.By effectively integrating spatial and semantic cues,CPF-CTE produces richer and more accurate representations of image content.Extensive experiments on PASCAL VOC 2012 and MS COCO 2014 validate that CPF-CTE consistently surpasses prior WSSS methods.展开更多
Alzheimer’s disease(AD)is the most common form of dementia,affecting over 50 million people worldwide.This figure is projected to nearly double every 20 years,reaching 82 million by 2030 and 152 million by 2050(Alzhe...Alzheimer’s disease(AD)is the most common form of dementia,affecting over 50 million people worldwide.This figure is projected to nearly double every 20 years,reaching 82 million by 2030 and 152 million by 2050(Alzheimer’s Disease International).The apolipoproteinε4(APOE4)allele is the strongest genetic risk factor for late-onset AD(after age 65 years).Apolipoprotein E,a lipid transporter,exists in three variants:ε2,ε3,andε4.APOEε2(APOE2)is protective against AD,APOEε3(APOE3)is neutral,while APOE4 significantly increases the risk.Individuals with one copy of APOE4 have a 4-fold greater risk of developing AD,and those with two copies face an 8-fold risk compared to non-carriers.Even in cognitively normal individuals,APOE4 carriers exhibit brain metabolic and vascular deficits decades before amyloid-beta(Aβ)plaques and neurofibrillary tau tangles emerge-the hallmark pathologies of AD(Reiman et al.,2001,2005;Thambisetty et al.,2010).Notably,studies have demonstrated reduced glucose uptake,or hypometabolism,in brain regions vulnerable to AD in asymptomatic middle-aged APOE4 carriers,long before clinical symptoms arise(Reiman et al.,2001,2005).展开更多
The newly formulated non-Newtonian rivulet flows streaming down an inclined planar surface,with additional periodic perturbations arising from the application of the 2nd Stokes problem to the investigation of rivulet ...The newly formulated non-Newtonian rivulet flows streaming down an inclined planar surface,with additional periodic perturbations arising from the application of the 2nd Stokes problem to the investigation of rivulet dynamics,are demonstrated in the current research.Hereby,the 2nd Stokes problem assumes that the surface,with a thin shared layer of the fluid on it,oscillates in a harmonic manner along the x-axis of the rivulet flow,which coincides with the main flow direction streaming down the underlying surface.We obtain the exact extension of the rivulet flow family,clarifying the structure of the pressure field,which fully absorbs the arising perturbation.The profile of the velocity field is assumed to be Gaussian-type with a non-zero level of plasticity.Hence,the absolutely non-Newtonian case of the viscoplastic flow solution,which satisfies the motion and continuity equations,is considered(with particular cases of exact solutions for pressure).The perturbed governing equations of motion for rivulet flows then result in the Riccati-type ordinary differential equation(ODE),describing the dynamics of the coordinate x(t).The approximated schematic dynamics are presented in graphical plots.展开更多
[Significance]In alignment with the national germplasm security strategy,current research efforts are accelerating the adoption of precision breeding in sheep.Within the whole-genome selection,accurate phenotyping of ...[Significance]In alignment with the national germplasm security strategy,current research efforts are accelerating the adoption of precision breeding in sheep.Within the whole-genome selection,accurate phenotyping of body morphometrics is critical for assessing growth performance and breeding value.Traditional manual measurements are inefficient,prone to human error,and may cause stress to sheep,limiting their suitability for precision sheep management.By summarizing the applications of sheep body size measurement technologies and analyzing their development directions,this paper provides theoretical references and practical guidance for the research and application of non contact sheep body size measurement.[Progress]This review synthesizes progress across three principal methodological paradigms:two-dimensional(2D)image-based techniques,three-dimensional(3D)point cloud-based approaches,and integrated 2D-3D fusion systems.2D methods,employing either handcrafted geometric features or deep learning-based keypoint detector algorithms,are cost-effective and operationally simple but sensitive to variation in imaging conditions and unable to capture critical circumference metrics.3D point-cloud approaches enable precise reconstruction of full animal morphology,supporting comprehensive body-size acquisition with higher accuracy,yet face challenges including high hardware costs,complex data workflows,and sensitivity to posture variability.Hybrid 2D-3D fusion systems combine semantic richness from RGB imagery with geometric completeness from point clouds.Having been effectively validated in other livestock specise,e.g.,cattle and pigs,these fusion systems have demonstrated excellent performance,providing important technical references and practical insights for sheep body size measurement.[Conclusions and Prospects]Firstly,future research should focus on constructing large-scale,high-quality datasets for sheep body size measurement that encompass diverse breeds,growth stages,and environmental conditions,thereby enhancing model robustness and generalization.Secondly,the development of lightweight artificial intelligence models is essential.Techniques such as model compression,quantization,and algorithmic optimization can substantially reduce computational complexity and storage requirements,facilitating deployment in resource-constrained environments.Thirdly,the 3D point cloud processing pipeline should be streamlined to improve the efficiency of data acquisition,filtering,registration,and segmentation,while promoting the integration of low-cost,high-resilience vision systems into practical farming scenarios.Fourthly,specific emphasis should be placed on improving the accuracy of curved-dimensional measurements,such as chest circumference,abdominal circumference,and shank circumference,through advances in pose standardization,refined 3D segmentation strategies,and multimodal data fusion.Finally,the cross-fertilization of sheep body size measurement technologies with analogous methods for other livestock species offers a promising pathway for mutual learning and collaborative innovation,accelerating the industrialization of automated sheep morphometric systems and supporting the development of intelligent,data-driven pasture management practices.展开更多
Accurate parameter extraction of photovoltaic(PV)models plays a critical role in enabling precise performance prediction,optimal system sizing,and effective operational control under diverse environmental conditions.W...Accurate parameter extraction of photovoltaic(PV)models plays a critical role in enabling precise performance prediction,optimal system sizing,and effective operational control under diverse environmental conditions.While a wide range of metaheuristic optimisation techniques have been applied to this problem,many existing methods are hindered by slow convergence rates,susceptibility to premature stagnation,and reduced accuracy when applied to complex multi-diode PV configurations.These limitations can lead to suboptimal modelling,reducing the efficiency of PV system design and operation.In this work,we propose an enhanced hybrid optimisation approach,the modified Spider Wasp Optimization(mSWO)with Opposition-Based Learning algorithm,which integrates the exploration and exploitation capabilities of the Spider Wasp Optimization(SWO)metaheuristic with the diversityenhancing mechanism of Opposition-Based Learning(OBL).The hybridisation is designed to dynamically expand the search space coverage,avoid premature convergence,and improve both convergence speed and precision in highdimensional optimisation tasks.The mSWO algorithm is applied to three well-established PV configurations:the single diode model(SDM),the double diode model(DDM),and the triple diode model(TDM).Real experimental current-voltage(I-V)datasets from a commercial PV module under standard test conditions(STC)are used for evaluation.Comparative analysis is conducted against eighteen advanced metaheuristic algorithms,including BSDE,RLGBO,GWOCS,MFO,EO,TSA,and SCA.Performance metrics include minimum,mean,and maximum root mean square error(RMSE),standard deviation(SD),and convergence behaviour over 30 independent runs.The results reveal that mSWO consistently delivers superior accuracy and robustness across all PV models,achieving the lowest RMSE values of 0.000986022(SDM),0.000982884(DDM),and 0.000982529(TDM),with minimal SD values,indicating remarkable repeatability.Convergence analyses further show that mSWO reaches optimal solutions more rapidly and with fewer oscillations than all competing methods,with the performance gap widening as model complexity increases.These findings demonstrate that mSWO provides a scalable,computationally efficient,and highly reliable framework for PV parameter extraction.Its adaptability to models of growing complexity suggests strong potential for broader applications in renewable energy systems,including performance monitoring,fault detection,and intelligent control,thereby contributing to the optimisation of next-generation solar energy solutions.展开更多
Mine surveying is an indispensable and crucial basic technical work in the process of mineral resource development.It plays an important role throughout the entire life cycle of a mine,from exploration,design,construc...Mine surveying is an indispensable and crucial basic technical work in the process of mineral resource development.It plays an important role throughout the entire life cycle of a mine,from exploration,design,construction,and production to closure,and is known as the“eyes of the mine”.With the rapid development of satellite technology,computer science,artificial intelligence,robotics,and spatiotemporal big data,mine surveying science and technology supported by spatial information technology is increasingly playing the role of the“brain of the mine”.This paper systematically summarizes the characteristics of mining surveying science and technology in contemporary and future mining development.First,based on the requirements of safe,efficient,and green development in modern mining,an analysis is conducted on the innovative practices of intelligent mining methods;secondly,it explains the transformation of regional economic and mining economic integration towards lengthening the industrial chain and scientific and technological innovation.Regarding intelligent mining,this paper discusses three technical dimensions:(1)By establishing a spatiotemporal data model of the mine,real-time perception and remote intelligent control of the production system are realized;(2)Based on the transparent mine three-dimensional geological modelling technology,the accuracy of geological condition prediction and the scientific nature of mining decisions are significantly improved;(3)By integrating multi-source remote sensing data and deep learning algorithms,a high-precision coal and rock identification system is constructed.The study further revealed the innovative application value of mine surveying in the post-mining era,including:diversified utilization of underground space in mining areas(tourism development,geothermal energy storage,pumped storage,etc.),multi-platform remote sensing coordinated ecological restoration monitoring,and optimized land space planning in mining areas.Practice has proved that mine surveying technology is an important technical engine for promoting green transformation and high-quality development in resource-based regions,and has irreplaceable strategic significance for achieving coordinated development of energy,economy,and environment.展开更多
Automated grading of dandruff severity is a clinically significant but challenging task due to the inherent ordinal nature of severity levels and the high prevalence of label noise from subjective expert annotations.S...Automated grading of dandruff severity is a clinically significant but challenging task due to the inherent ordinal nature of severity levels and the high prevalence of label noise from subjective expert annotations.Standard classification methods fail to address these dual challenges,limiting their real-world performance.In this paper,a novel,three-phase training framework is proposed that learns a robust ordinal classifier directly from noisy labels.The approach synergistically combines a rank-based ordinal regression backbone with a cooperative,semi-supervised learning strategy to dynamically partition the data into clean and noisy subsets.A hybrid training objective is then employed,applying a supervised ordinal loss to the clean set.The noisy set is simultaneously trained using a dualobjective that combines a semi-supervised ordinal loss with a parallel,label-agnostic contrastive loss.This design allows themodel to learn fromthe entire noisy subset while using contrastive learning to mitigate the risk of error propagation frompotentially corrupt supervision.Extensive experiments on a new,large-scale,multi-site clinical dataset validate our approach.Themethod achieves state-of-the-art performance with 80.71%accuracy and a 76.86%F1-score,significantly outperforming existing approaches,including a 2.26%improvement over the strongest baseline method.This work provides not only a robust solution for a practical medical imaging problem but also a generalizable framework for other tasks plagued by noisy ordinal labels.展开更多
The rapid expansion of the Internet of Things(IoT)and Edge Artificial Intelligence(AI)has redefined automation and connectivity acrossmodern networks.However,the heterogeneity and limited resources of IoT devices expo...The rapid expansion of the Internet of Things(IoT)and Edge Artificial Intelligence(AI)has redefined automation and connectivity acrossmodern networks.However,the heterogeneity and limited resources of IoT devices expose them to increasingly sophisticated and persistentmalware attacks.These adaptive and stealthy threats can evade conventional detection,establish remote control,propagate across devices,exfiltrate sensitive data,and compromise network integrity.This study presents a Software-Defined Internet of Things(SD-IoT)control-plane-based,AI-driven framework that integrates Gated Recurrent Units(GRU)and Long Short-TermMemory(LSTM)networks for efficient detection of evolving multi-vector,malware-driven botnet attacks.The proposed CUDA-enabled hybrid deep learning(DL)framework performs centralized real-time detection without adding computational overhead to IoT nodes.A feature selection strategy combining variable clustering,attribute evaluation,one-R attribute evaluation,correlation analysis,and principal component analysis(PCA)enhances detection accuracy and reduces complexity.The framework is rigorously evaluated using the N_BaIoT dataset under k-fold cross-validation.Experimental results achieve 99.96%detection accuracy,a false positive rate(FPR)of 0.0035%,and a detection latency of 0.18 ms,confirming its high efficiency and scalability.The findings demonstrate the framework’s potential as a robust and intelligent security solution for next-generation IoT ecosystems.展开更多
Spam emails remain one of the most persistent threats to digital communication,necessitating effective detection solutions that safeguard both individuals and organisations.We propose a spam email classification frame...Spam emails remain one of the most persistent threats to digital communication,necessitating effective detection solutions that safeguard both individuals and organisations.We propose a spam email classification frame-work that uses Bidirectional Encoder Representations from Transformers(BERT)for contextual feature extraction and a multiple-window Convolutional Neural Network(CNN)for classification.To identify semantic nuances in email content,BERT embeddings are used,and CNN filters extract discriminative n-gram patterns at various levels of detail,enabling accurate spam identification.The proposed model outperformed Word2Vec-based baselines on a sample of 5728 labelled emails,achieving an accuracy of 98.69%,AUC of 0.9981,F1 Score of 0.9724,and MCC of 0.9639.With a medium kernel size of(6,9)and compact multi-window CNN architectures,it improves performance.Cross-validation illustrates stability and generalization across folds.By balancing high recall with minimal false positives,our method provides a reliable and scalable solution for current spam detection in advanced deep learning.By combining contextual embedding and a neural architecture,this study develops a security analysis method.展开更多
The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)an...The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)and Deep Learning(DL)techniques have demonstrated promising early detection capabilities.However,their effectiveness is limited when handling the vast volumes of IoT-generated data due to scalability constraints,high computational costs,and the costly time-intensive process of data labeling.To address these challenges,this study proposes a Federated Learning(FL)framework that leverages collaborative and hybrid supervised learning to enhance cyber threat detection in IoT networks.By employing Deep Neural Networks(DNNs)and decentralized model training,the approach reduces computational complexity while improving detection accuracy.The proposed model demonstrates robust performance,achieving accuracies of 94.34%,99.95%,and 87.94%on the publicly available kitsune,Bot-IoT,and UNSW-NB15 datasets,respectively.Furthermore,its ability to detect zero-day attacks is validated through evaluations on two additional benchmark datasets,TON-IoT and IoT-23,using a Deep Federated Learning(DFL)framework,underscoring the generalization and effectiveness of the model in heterogeneous and decentralized IoT environments.Experimental results demonstrate superior performance over existing methods,establishing the proposed framework as an efficient and scalable solution for IoT security.展开更多
文摘In June 2012,the UN conference on sustainable development "Rio+20" summarized the work of the world community in this direction for last 20 years and outlined the tasks for the future.The UN website contains enough information to estimate the importance of global problems and "green economy," taking into account a very complicated state of the global and Russian markets which are balancing on the verge of crisis.Unfortunately,the website does not contain the materials of the 6th civilization forum "Long-Term Strategy for Sustainable Development on the Basis of Partnership of Civilizations:Concepts,Strategy,Programs and Projects" within the bounds of "Rio+20" which has considered the problems of a dialogue and partnership in conditions of extensive globalization.These problems are covered in the Partnership of Civilizations Journal which is issued in the Russian,English and Arabian languages,including its Internet version.The International Informatization Academy (it has the general consultative status at the Economic and Social Council of the UN) and its 20-year activity in the sphere of informatization of the world and Russia on the way to the partnership of civilizations,are presented there[1].
文摘The rapidly evolving cybersecurity threat landscape exposes a critical flaw in traditional educational programs where static curricula cannot adapt swiftly to novel attack vectors.This creates a significant gap between theoretical knowledge and the practical defensive capabilities needed in the field.To address this,we propose TeachSecure-CTI,a novel framework for adaptive cybersecurity curriculumgeneration that integrates real-time Cyber Threat Intelligence(CTI)with AI-driven personalization.Our framework employs a layered architecture featuring a CTI ingestion and clusteringmodule,natural language processing for semantic concept extraction,and a reinforcement learning agent for adaptive content sequencing.Bydynamically aligning learningmaterialswithboththe evolving threat environment and individual learner profiles,TeachSecure-CTI ensures content remains current,relevant,and tailored.A 12-week study with 150 students across three institutions demonstrated that the framework improves learning gains by 34%,significantly exceeding the 12%–21%reported in recent literature.The system achieved 84.8%personalization accuracy,85.9%recognition accuracy for MITRE ATT&CK tactics,and a 31%faster competency development rate compared to static curricula.These findings have implications beyond academia,extending to workforce development,cyber range training,and certification programs.By bridging the gap between dynamic threats and static educational materials,TeachSecure-CTI offers an empirically validated,scalable solution for cultivating cybersecurity professionals capable of responding to modern threats.
基金output of a research project implemented as part of the Basic Research Program at HSE University。
文摘Activation pruning reduces neural network complexity by eliminating low-importance neuron activations,yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally expensive and typically requires exhaustive search.We introduce a thermodynamics-inspired framework that treats activation distributions as energy-filtered physical systems and employs the free energy of activations as a principled evaluation metric.Phase-transition-like phenomena in the free-energy profile—such as extrema,inflection points,and curvature changes—yield reliable estimates of the critical pruning threshold,providing a theoretically grounded means of predicting sharp accuracy degradation.To further enhance efficiency,we propose a renormalized free energy technique that approximates full-evaluation free energy using only the activation distribution of the unpruned network.This eliminates repeated forward passes,dramatically reducing computational overhead and achieving speedups of up to 550×for MLPs.Extensive experiments across diverse vision architectures(MLP,CNN,ResNet,MobileNet,Vision Transformer)and text models(LSTM,BERT,ELECTRA,T5,GPT-2)on multiple datasets validate the generality,robustness,and computational efficiency of our approach.Overall,this work establishes a theoretically grounded and practically effective framework for activation pruning,bridging the gap between analytical understanding and efficient deployment of sparse neural networks.
基金supported by the Ministry of Research,Innovation and Digitization,CNCS/CCCDI–UEFISCDI(Romania),Nr.11/2024,within PNCDI IV.The APC received no external funding.
文摘This work addresses optimality aspects related to composite laminates having layers with different orientations.RegressionNeuralNetworks can model the mechanical behavior of these laminates,specifically the stressstrain relationship.If this model has strong generalization ability,it can be coupled with a metaheuristic algorithm–the PSO algorithm used in this article–to address an optimization problem(OP)related to the orientations of composite laminates.To solve OPs,this paper proposes an optimization framework(OFW)that connects the two components,the optimal solution search mechanism and the RNN model.The OFW has two modules:the search mechanism(Adaptive Hybrid Topology PSO)and the Prediction and Computation Module(PCM).The PCM undertakes all the activities concerning the OP at hand:the stress-strain model,constraints checking,and computation of the objective function.Two case studies about the layers’orientations of laminated specimens are conducted to validate the proposed framework.The specimens belong to“Off-axis oriented specimens”and are subjects of two OPs.The algorithms for AHTPSO and for the two PCMs(one for each problem)are proposed and implemented by MATLAB scripts and functions.Simulations are carried out for different initial conditions.The solutions demonstrated that the OFW is effective and has a highly acceptable computational complexity.The limitation of using the OFWis the generalization ability of the RNN model or any other regression models.To harness the RNN model efficiently,it must have a very good generalization power.If this condition ismet,the OFWcan be integrated into any design process to make optimal choices of the layers’orientations.
基金supported by the National Natural Science Foundation of China(Grant No.62266054)the Major Science and Technology Project of Yunnan Province(Grant No.202402AD080002)the Scientific Research Fund of the Yunnan Provincial Department of Education(Grant No.2025Y0302).
文摘End-to-end Temporal Action Detection(TAD)has achieved remarkable progress in recent years,driven by innovations in model architectures and the emergence of Video Foundation Models(VFMs).However,existing TAD methods that perform full fine-tuning of pretrained video models often incur substantial computational costs,which become particularly pronounced when processing long video sequences.Moreover,the need for precise temporal boundary annotations makes data labeling extremely expensive.In low-resource settings where annotated samples are scarce,direct fine-tuning tends to cause overfitting.To address these challenges,we introduce Dynamic LowRank Adapter(DyLoRA),a lightweight fine-tuning framework tailored specifically for the TAD task.Built upon the Low-Rank Adaptation(LoRA)architecture,DyLoRA adapts only the key layers of the pretrained model via low-rank decomposition,reducing the number of trainable parameters to less than 5%of full fine-tuning methods.This significantly lowers memory consumption and mitigates overfitting in low-resource settings.Notably,DyLoRA enhances the temporal modeling capability of pretrained models by optimizing temporal dimension weights,thereby alleviating the representation misalignment of temporal features.Experimental results demonstrate that DyLoRA-TAD achieves impressive performance,with 73.9%mAP on THUMOS14,39.52%on ActivityNet-1.3,and 28.2%on Charades,substantially surpassing the best traditional feature-based methods.
文摘Customer churn is the rate at which customers discontinue doing business with a company over a given time period.It is an essential measure for businesses to monitor high churn rates,as they often indicate underlying issues with services,products,or customer experience,resulting in considerable income loss.Prediction of customer churn is a crucial task aimed at retaining customers and maintaining revenue growth.Traditional machine learning(ML)models often struggle to capture complex temporal dependencies in client behavior data.To address this,an optimized deep learning(DL)approach using a Regularized Bidirectional Long Short-Term Memory(RBiLSTM)model is proposed to mitigate overfitting and improve generalization error.The model integrates dropout,L2-regularization,and early stopping to enhance predictive accuracy while preventing over-reliance on specific patterns.Moreover,this study investigates the effect of optimization techniques on boosting the training efficiency of the developed model.Experimental results on a recent public customer churn dataset demonstrate that the trained model outperforms the traditional ML models and some other DL models,such as Long Short-Term Memory(LSTM)and Deep Neural Network(DNN),in churn prediction performance and stability.The proposed approach achieves 96.1%accuracy,compared with LSTM and DNN,which attain 94.5%and 94.1%accuracy,respectively.These results confirm that the proposed approach can be used as a valuable tool for businesses to identify at-risk consumers proactively and implement targeted retention strategies.
文摘Alzheimer’s disease(AD)is the most common form of dementia.In addition to the lack of effective treatments,there are limitations in diagnostic capabilities.The complexity of AD itself,together with a variety of other diseases often observed in a patient’s history in addition to their AD diagnosis,make deciphering the molecular mechanisms that underlie AD,even more important.Large datasets of single-cell RNA sequencing,single-nucleus RNA-sequencing(snRNA-seq),and spatial transcriptomics(ST)have become essential in guiding and supporting new investigations into the cellular and regional susceptibility of AD.However,with unique technology,software,and larger databases emerging;a lack of integration of these data can contribute to ineffective use of valuable knowledge.Importantly,there was no specialized database that concentrates on ST in AD that offers comprehensive differential analyses under various conditions,such as sex-specific,region-specific,and comparisons between AD and control groups until the new Single-cell and Spatial RNA-seq databasE for Alzheimer’s Disease(ssREAD)database(Wang et al.,2024)was introduced to meet the scientific community’s growing demand for comprehensive,integrated,and accessible data analysis.
基金supported by the project“Romanian Hub for Artificial Intelligence-HRIA”,Smart Growth,Digitization and Financial Instruments Program,2021–2027,MySMIS No.334906.
文摘Objective expertise evaluation of individuals,as a prerequisite stage for team formation,has been a long-term desideratum in large software development companies.With the rapid advancements in machine learning methods,based on reliable existing data stored in project management tools’datasets,automating this evaluation process becomes a natural step forward.In this context,our approach focuses on quantifying software developer expertise by using metadata from the task-tracking systems.For this,we mathematically formalize two categories of expertise:technology-specific expertise,which denotes the skills required for a particular technology,and general expertise,which encapsulates overall knowledge in the software industry.Afterward,we automatically classify the zones of expertise associated with each task a developer has worked on using Bidirectional Encoder Representations from Transformers(BERT)-like transformers to handle the unique characteristics of project tool datasets effectively.Finally,our method evaluates the proficiency of each software specialist across already completed projects from both technology-specific and general perspectives.The method was experimentally validated,yielding promising results.
基金Kobe Universitythe National Research and Innovation Agency (BRIN)
文摘The effect of plasma and charged particle interaction with spacecraft in a low Earth orbit(LEO)environment leads to ion focusing and the formation of an ion void in the downstream region as a result of charging.Simulations and investigations using a fixed potential imposed on the spacecraft showed the nonsignificance of geophysical parameter changes to ion focusing.Variation of the temperature ratio(T_(r))contributed only to local ion focusing and manifested as two-ion streamers dispersed at the upper and lower edges of the spacecraft-the outermost layers of the satellite structure at the top and bottom,respectively.A simulation involving changing the ambient plasma density(N_(p))also showed the formation of local ion focusing,in which ions were more concentrated as the density increased.Furthermore,auroral electron density(N_(ae))variation had no clear impact on ion focusing,as indicated by static two-ion structures in the wake field.However,variation of the object potential(ϕ)strongly affected ion focusing formation,leading to distortion of the initial ion void region behind the spacecraft.The formation of ion focusing in this study was subject to the electric field produced by the object potential and the ambipolar electric field resulting from plasma expansion in the downstream region.
文摘Weakly Supervised Semantic Segmentation(WSSS),which relies only on image-level labels,has attracted significant attention for its cost-effectiveness and scalability.Existing methods mainly enhance inter-class distinctions and employ data augmentation to mitigate semantic ambiguity and reduce spurious activations.However,they often neglect the complex contextual dependencies among image patches,resulting in incomplete local representations and limited segmentation accuracy.To address these issues,we propose the Context Patch Fusion with Class Token Enhancement(CPF-CTE)framework,which exploits contextual relations among patches to enrich feature repre-sentations and improve segmentation.At its core,the Contextual-Fusion Bidirectional Long Short-Term Memory(CF-BiLSTM)module captures spatial dependencies between patches and enables bidirectional information flow,yield-ing a more comprehensive understanding of spatial correlations.This strengthens feature learning and segmentation robustness.Moreover,we introduce learnable class tokens that dynamically encode and refine class-specific semantics,enhancing discriminative capability.By effectively integrating spatial and semantic cues,CPF-CTE produces richer and more accurate representations of image content.Extensive experiments on PASCAL VOC 2012 and MS COCO 2014 validate that CPF-CTE consistently surpasses prior WSSS methods.
基金supported by National Institute on Aging(NIH-NIA)R01AG054459(to ALL).
文摘Alzheimer’s disease(AD)is the most common form of dementia,affecting over 50 million people worldwide.This figure is projected to nearly double every 20 years,reaching 82 million by 2030 and 152 million by 2050(Alzheimer’s Disease International).The apolipoproteinε4(APOE4)allele is the strongest genetic risk factor for late-onset AD(after age 65 years).Apolipoprotein E,a lipid transporter,exists in three variants:ε2,ε3,andε4.APOEε2(APOE2)is protective against AD,APOEε3(APOE3)is neutral,while APOE4 significantly increases the risk.Individuals with one copy of APOE4 have a 4-fold greater risk of developing AD,and those with two copies face an 8-fold risk compared to non-carriers.Even in cognitively normal individuals,APOE4 carriers exhibit brain metabolic and vascular deficits decades before amyloid-beta(Aβ)plaques and neurofibrillary tau tangles emerge-the hallmark pathologies of AD(Reiman et al.,2001,2005;Thambisetty et al.,2010).Notably,studies have demonstrated reduced glucose uptake,or hypometabolism,in brain regions vulnerable to AD in asymptomatic middle-aged APOE4 carriers,long before clinical symptoms arise(Reiman et al.,2001,2005).
文摘The newly formulated non-Newtonian rivulet flows streaming down an inclined planar surface,with additional periodic perturbations arising from the application of the 2nd Stokes problem to the investigation of rivulet dynamics,are demonstrated in the current research.Hereby,the 2nd Stokes problem assumes that the surface,with a thin shared layer of the fluid on it,oscillates in a harmonic manner along the x-axis of the rivulet flow,which coincides with the main flow direction streaming down the underlying surface.We obtain the exact extension of the rivulet flow family,clarifying the structure of the pressure field,which fully absorbs the arising perturbation.The profile of the velocity field is assumed to be Gaussian-type with a non-zero level of plasticity.Hence,the absolutely non-Newtonian case of the viscoplastic flow solution,which satisfies the motion and continuity equations,is considered(with particular cases of exact solutions for pressure).The perturbed governing equations of motion for rivulet flows then result in the Riccati-type ordinary differential equation(ODE),describing the dynamics of the coordinate x(t).The approximated schematic dynamics are presented in graphical plots.
文摘[Significance]In alignment with the national germplasm security strategy,current research efforts are accelerating the adoption of precision breeding in sheep.Within the whole-genome selection,accurate phenotyping of body morphometrics is critical for assessing growth performance and breeding value.Traditional manual measurements are inefficient,prone to human error,and may cause stress to sheep,limiting their suitability for precision sheep management.By summarizing the applications of sheep body size measurement technologies and analyzing their development directions,this paper provides theoretical references and practical guidance for the research and application of non contact sheep body size measurement.[Progress]This review synthesizes progress across three principal methodological paradigms:two-dimensional(2D)image-based techniques,three-dimensional(3D)point cloud-based approaches,and integrated 2D-3D fusion systems.2D methods,employing either handcrafted geometric features or deep learning-based keypoint detector algorithms,are cost-effective and operationally simple but sensitive to variation in imaging conditions and unable to capture critical circumference metrics.3D point-cloud approaches enable precise reconstruction of full animal morphology,supporting comprehensive body-size acquisition with higher accuracy,yet face challenges including high hardware costs,complex data workflows,and sensitivity to posture variability.Hybrid 2D-3D fusion systems combine semantic richness from RGB imagery with geometric completeness from point clouds.Having been effectively validated in other livestock specise,e.g.,cattle and pigs,these fusion systems have demonstrated excellent performance,providing important technical references and practical insights for sheep body size measurement.[Conclusions and Prospects]Firstly,future research should focus on constructing large-scale,high-quality datasets for sheep body size measurement that encompass diverse breeds,growth stages,and environmental conditions,thereby enhancing model robustness and generalization.Secondly,the development of lightweight artificial intelligence models is essential.Techniques such as model compression,quantization,and algorithmic optimization can substantially reduce computational complexity and storage requirements,facilitating deployment in resource-constrained environments.Thirdly,the 3D point cloud processing pipeline should be streamlined to improve the efficiency of data acquisition,filtering,registration,and segmentation,while promoting the integration of low-cost,high-resilience vision systems into practical farming scenarios.Fourthly,specific emphasis should be placed on improving the accuracy of curved-dimensional measurements,such as chest circumference,abdominal circumference,and shank circumference,through advances in pose standardization,refined 3D segmentation strategies,and multimodal data fusion.Finally,the cross-fertilization of sheep body size measurement technologies with analogous methods for other livestock species offers a promising pathway for mutual learning and collaborative innovation,accelerating the industrialization of automated sheep morphometric systems and supporting the development of intelligent,data-driven pasture management practices.
基金funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2025R442)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Accurate parameter extraction of photovoltaic(PV)models plays a critical role in enabling precise performance prediction,optimal system sizing,and effective operational control under diverse environmental conditions.While a wide range of metaheuristic optimisation techniques have been applied to this problem,many existing methods are hindered by slow convergence rates,susceptibility to premature stagnation,and reduced accuracy when applied to complex multi-diode PV configurations.These limitations can lead to suboptimal modelling,reducing the efficiency of PV system design and operation.In this work,we propose an enhanced hybrid optimisation approach,the modified Spider Wasp Optimization(mSWO)with Opposition-Based Learning algorithm,which integrates the exploration and exploitation capabilities of the Spider Wasp Optimization(SWO)metaheuristic with the diversityenhancing mechanism of Opposition-Based Learning(OBL).The hybridisation is designed to dynamically expand the search space coverage,avoid premature convergence,and improve both convergence speed and precision in highdimensional optimisation tasks.The mSWO algorithm is applied to three well-established PV configurations:the single diode model(SDM),the double diode model(DDM),and the triple diode model(TDM).Real experimental current-voltage(I-V)datasets from a commercial PV module under standard test conditions(STC)are used for evaluation.Comparative analysis is conducted against eighteen advanced metaheuristic algorithms,including BSDE,RLGBO,GWOCS,MFO,EO,TSA,and SCA.Performance metrics include minimum,mean,and maximum root mean square error(RMSE),standard deviation(SD),and convergence behaviour over 30 independent runs.The results reveal that mSWO consistently delivers superior accuracy and robustness across all PV models,achieving the lowest RMSE values of 0.000986022(SDM),0.000982884(DDM),and 0.000982529(TDM),with minimal SD values,indicating remarkable repeatability.Convergence analyses further show that mSWO reaches optimal solutions more rapidly and with fewer oscillations than all competing methods,with the performance gap widening as model complexity increases.These findings demonstrate that mSWO provides a scalable,computationally efficient,and highly reliable framework for PV parameter extraction.Its adaptability to models of growing complexity suggests strong potential for broader applications in renewable energy systems,including performance monitoring,fault detection,and intelligent control,thereby contributing to the optimisation of next-generation solar energy solutions.
基金supported by the National Natural Science Foundation of China(Nos.52394193 and U22A20569)the National Key R&D Program Projects(Nos.2023YFC3804200 and 2023YFC3804205).
文摘Mine surveying is an indispensable and crucial basic technical work in the process of mineral resource development.It plays an important role throughout the entire life cycle of a mine,from exploration,design,construction,and production to closure,and is known as the“eyes of the mine”.With the rapid development of satellite technology,computer science,artificial intelligence,robotics,and spatiotemporal big data,mine surveying science and technology supported by spatial information technology is increasingly playing the role of the“brain of the mine”.This paper systematically summarizes the characteristics of mining surveying science and technology in contemporary and future mining development.First,based on the requirements of safe,efficient,and green development in modern mining,an analysis is conducted on the innovative practices of intelligent mining methods;secondly,it explains the transformation of regional economic and mining economic integration towards lengthening the industrial chain and scientific and technological innovation.Regarding intelligent mining,this paper discusses three technical dimensions:(1)By establishing a spatiotemporal data model of the mine,real-time perception and remote intelligent control of the production system are realized;(2)Based on the transparent mine three-dimensional geological modelling technology,the accuracy of geological condition prediction and the scientific nature of mining decisions are significantly improved;(3)By integrating multi-source remote sensing data and deep learning algorithms,a high-precision coal and rock identification system is constructed.The study further revealed the innovative application value of mine surveying in the post-mining era,including:diversified utilization of underground space in mining areas(tourism development,geothermal energy storage,pumped storage,etc.),multi-platform remote sensing coordinated ecological restoration monitoring,and optimized land space planning in mining areas.Practice has proved that mine surveying technology is an important technical engine for promoting green transformation and high-quality development in resource-based regions,and has irreplaceable strategic significance for achieving coordinated development of energy,economy,and environment.
文摘Automated grading of dandruff severity is a clinically significant but challenging task due to the inherent ordinal nature of severity levels and the high prevalence of label noise from subjective expert annotations.Standard classification methods fail to address these dual challenges,limiting their real-world performance.In this paper,a novel,three-phase training framework is proposed that learns a robust ordinal classifier directly from noisy labels.The approach synergistically combines a rank-based ordinal regression backbone with a cooperative,semi-supervised learning strategy to dynamically partition the data into clean and noisy subsets.A hybrid training objective is then employed,applying a supervised ordinal loss to the clean set.The noisy set is simultaneously trained using a dualobjective that combines a semi-supervised ordinal loss with a parallel,label-agnostic contrastive loss.This design allows themodel to learn fromthe entire noisy subset while using contrastive learning to mitigate the risk of error propagation frompotentially corrupt supervision.Extensive experiments on a new,large-scale,multi-site clinical dataset validate our approach.Themethod achieves state-of-the-art performance with 80.71%accuracy and a 76.86%F1-score,significantly outperforming existing approaches,including a 2.26%improvement over the strongest baseline method.This work provides not only a robust solution for a practical medical imaging problem but also a generalizable framework for other tasks plagued by noisy ordinal labels.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting ProjectNumber(PNURSP2025R97),PrincessNourah bint AbdulrahmanUniversity,Riyadh,Saudi Arabia.
文摘The rapid expansion of the Internet of Things(IoT)and Edge Artificial Intelligence(AI)has redefined automation and connectivity acrossmodern networks.However,the heterogeneity and limited resources of IoT devices expose them to increasingly sophisticated and persistentmalware attacks.These adaptive and stealthy threats can evade conventional detection,establish remote control,propagate across devices,exfiltrate sensitive data,and compromise network integrity.This study presents a Software-Defined Internet of Things(SD-IoT)control-plane-based,AI-driven framework that integrates Gated Recurrent Units(GRU)and Long Short-TermMemory(LSTM)networks for efficient detection of evolving multi-vector,malware-driven botnet attacks.The proposed CUDA-enabled hybrid deep learning(DL)framework performs centralized real-time detection without adding computational overhead to IoT nodes.A feature selection strategy combining variable clustering,attribute evaluation,one-R attribute evaluation,correlation analysis,and principal component analysis(PCA)enhances detection accuracy and reduces complexity.The framework is rigorously evaluated using the N_BaIoT dataset under k-fold cross-validation.Experimental results achieve 99.96%detection accuracy,a false positive rate(FPR)of 0.0035%,and a detection latency of 0.18 ms,confirming its high efficiency and scalability.The findings demonstrate the framework’s potential as a robust and intelligent security solution for next-generation IoT ecosystems.
基金funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2026R234)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Spam emails remain one of the most persistent threats to digital communication,necessitating effective detection solutions that safeguard both individuals and organisations.We propose a spam email classification frame-work that uses Bidirectional Encoder Representations from Transformers(BERT)for contextual feature extraction and a multiple-window Convolutional Neural Network(CNN)for classification.To identify semantic nuances in email content,BERT embeddings are used,and CNN filters extract discriminative n-gram patterns at various levels of detail,enabling accurate spam identification.The proposed model outperformed Word2Vec-based baselines on a sample of 5728 labelled emails,achieving an accuracy of 98.69%,AUC of 0.9981,F1 Score of 0.9724,and MCC of 0.9639.With a medium kernel size of(6,9)and compact multi-window CNN architectures,it improves performance.Cross-validation illustrates stability and generalization across folds.By balancing high recall with minimal false positives,our method provides a reliable and scalable solution for current spam detection in advanced deep learning.By combining contextual embedding and a neural architecture,this study develops a security analysis method.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project Number(PNURSP2025R97)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)and Deep Learning(DL)techniques have demonstrated promising early detection capabilities.However,their effectiveness is limited when handling the vast volumes of IoT-generated data due to scalability constraints,high computational costs,and the costly time-intensive process of data labeling.To address these challenges,this study proposes a Federated Learning(FL)framework that leverages collaborative and hybrid supervised learning to enhance cyber threat detection in IoT networks.By employing Deep Neural Networks(DNNs)and decentralized model training,the approach reduces computational complexity while improving detection accuracy.The proposed model demonstrates robust performance,achieving accuracies of 94.34%,99.95%,and 87.94%on the publicly available kitsune,Bot-IoT,and UNSW-NB15 datasets,respectively.Furthermore,its ability to detect zero-day attacks is validated through evaluations on two additional benchmark datasets,TON-IoT and IoT-23,using a Deep Federated Learning(DFL)framework,underscoring the generalization and effectiveness of the model in heterogeneous and decentralized IoT environments.Experimental results demonstrate superior performance over existing methods,establishing the proposed framework as an efficient and scalable solution for IoT security.
