As AI systems scale, the limitations of cloud-based architectures, including latency, bandwidth, and privacy concerns, demand decentralized alternatives. Federated learning (FL) and Edge AI provide a paradigm shift by...As AI systems scale, the limitations of cloud-based architectures, including latency, bandwidth, and privacy concerns, demand decentralized alternatives. Federated learning (FL) and Edge AI provide a paradigm shift by combining privacy preserving training with efficient, on device computation. This paper introduces a cutting-edge FL-edge integration framework, achieving a 10% to 15% increase in model accuracy and reducing communication costs by 25% in heterogeneous environments. Blockchain based secure aggregation ensures robust and tamper-proof model updates, while exploratory quantum AI techniques enhance computational efficiency. By addressing key challenges such as device variability and non-IID data, this work sets the stage for the next generation of adaptive, privacy-first AI systems, with applications in IoT, healthcare, and autonomous systems.展开更多
The fast increase of online communities has brought about an increase in cyber threats inclusive of cyberbullying, hate speech, misinformation, and online harassment, making content moderation a pressing necessity. Tr...The fast increase of online communities has brought about an increase in cyber threats inclusive of cyberbullying, hate speech, misinformation, and online harassment, making content moderation a pressing necessity. Traditional single-modal AI-based detection systems, which analyze both text, photos, or movies in isolation, have established useless at taking pictures multi-modal threats, in which malicious actors spread dangerous content throughout a couple of formats. To cope with these demanding situations, we advise a multi-modal deep mastering framework that integrates Natural Language Processing (NLP), Convolutional Neural Networks (CNNs), and Long Short-Term Memory (LSTM) networks to become aware of and mitigate online threats effectively. Our proposed model combines BERT for text class, ResNet50 for photograph processing, and a hybrid LSTM-3-d CNN community for video content material analysis. We constructed a large-scale dataset comprising 500,000 textual posts, 200,000 offensive images, and 50,000 annotated motion pictures from more than one platform, which includes Twitter, Reddit, YouTube, and online gaming forums. The system became carefully evaluated using trendy gadget mastering metrics which include accuracy, precision, remember, F1-score, and ROC-AUC curves. Experimental outcomes demonstrate that our multi-modal method extensively outperforms single-modal AI classifiers, achieving an accuracy of 92.3%, precision of 91.2%, do not forget of 90.1%, and an AUC rating of 0.95. The findings validate the necessity of integrating multi-modal AI for actual-time, high-accuracy online chance detection and moderation. Future paintings will have consciousness on improving hostile robustness, enhancing scalability for real-world deployment, and addressing ethical worries associated with AI-driven content moderation.展开更多
Membrane fouling is a persistent challenge in membrane-based technologies,significantly impacting efficiency,operational costs,and system lifespan in applications like water treatment,desalination,and industrial proce...Membrane fouling is a persistent challenge in membrane-based technologies,significantly impacting efficiency,operational costs,and system lifespan in applications like water treatment,desalination,and industrial processing.Foul-ing,caused by the accumulation of particulates,organic compounds,and microorganisms,leads to reduced permeability,increased energy demands,and frequent maintenance.Traditional fouling control approaches,relying on empirical models and reactive strategies,often fail to address these issues efficiently.In this context,artificial intelligence(AI)and machine learning(ML)have emerged as innovative tools offering predictive and proactive solutions for fouling man-agement.By utilizing historical and real-time data,AI/ML techniques such as artificial neural networks,support vector machines,and ensemble models enable accurate prediction of fouling onset,identification of fouling mechanisms,and optimization of control measures.This review provides a detailed examination of the integration of AI/ML in membrane fouling prediction and mitigation,discussing advanced algorithms,the role of sensor-based monitoring,and the importance of robust datasets in enhancing predictive accuracy.Case studies highlighting successful AI/ML applications across various membrane processes are presented,demonstrating their transformative potential in improving system performance.Emerging trends,such as hybrid modeling and IoT-enabled smart systems,are explored,alongside a criti-cal analysis of research gaps and opportunities.This review emphasizes AI/ML as a cornerstone for sustainable,cost-effective membrane operations.展开更多
Against the background of the continuous reform in medical education,biochemistry,as a fundamental medical course,maintains a close connection with clinical practice.However,under the traditional teaching model,the ef...Against the background of the continuous reform in medical education,biochemistry,as a fundamental medical course,maintains a close connection with clinical practice.However,under the traditional teaching model,the effectiveness of the“basic-clinical”connection is relatively poor,which hinders the improvement of educational outcomes.In the practical teaching of higher vocational medical education,the integration of the AI Case-Guided Learning System can enhance students’enthusiasm for knowledge exploration and effectively improve teaching quality.Starting from the perspective of the“basic-clinical”connection teaching in the biochemistry course,this paper analyzes the application value of the AI Case-Guided Learning System and proposes specific application strategies,aiming to accumulate experience for the innovation of biochemistry teaching.展开更多
This study explores a novel educational model of generative AI-empowered interdisciplinary project-based learning(PBL).By analyzing the current applications of generative AI technology in information technology curric...This study explores a novel educational model of generative AI-empowered interdisciplinary project-based learning(PBL).By analyzing the current applications of generative AI technology in information technology curricula,it elucidates its advantages and operational mechanisms in interdisciplinary PBL.Combining case studies and empirical research,the investigation proposes implementation pathways and strategies for the generative AI-enhanced interdisciplinary PBL model,detailing specific applications across three phases:project preparation,implementation,and evaluation.The research demonstrates that generative AI-enabled interdisciplinary project-based learning can effectively enhance students’learning motivation,interdisciplinary thinking capabilities,and innovative competencies,providing new conceptual frameworks and practical approaches for educational model innovation.展开更多
In the evolving landscape of cyber threats,phishing attacks pose significant challenges,particularly through deceptive webpages designed to extract sensitive information under the guise of legitimacy.Conventional and ...In the evolving landscape of cyber threats,phishing attacks pose significant challenges,particularly through deceptive webpages designed to extract sensitive information under the guise of legitimacy.Conventional and machine learning(ML)-based detection systems struggle to detect phishing websites owing to their constantly changing tactics.Furthermore,newer phishing websites exhibit subtle and expertly concealed indicators that are not readily detectable.Hence,effective detection depends on identifying the most critical features.Traditional feature selection(FS)methods often struggle to enhance ML model performance and instead decrease it.To combat these issues,we propose an innovative method using explainable AI(XAI)to enhance FS in ML models and improve the identification of phishing websites.Specifically,we employ SHapley Additive exPlanations(SHAP)for global perspective and aggregated local interpretable model-agnostic explanations(LIME)to deter-mine specific localized patterns.The proposed SHAP and LIME-aggregated FS(SLA-FS)framework pinpoints the most informative features,enabling more precise,swift,and adaptable phishing detection.Applying this approach to an up-to-date web phishing dataset,we evaluate the performance of three ML models before and after FS to assess their effectiveness.Our findings reveal that random forest(RF),with an accuracy of 97.41%and XGBoost(XGB)at 97.21%significantly benefit from the SLA-FS framework,while k-nearest neighbors lags.Our framework increases the accuracy of RF and XGB by 0.65%and 0.41%,respectively,outperforming traditional filter or wrapper methods and any prior methods evaluated on this dataset,showcasing its potential.展开更多
Predicting the health status of stroke patients at different stages of the disease is a critical clinical task.The onset and development of stroke are affected by an array of factors,encompassing genetic predispositio...Predicting the health status of stroke patients at different stages of the disease is a critical clinical task.The onset and development of stroke are affected by an array of factors,encompassing genetic predisposition,environmental exposure,unhealthy lifestyle habits,and existing medical conditions.Although existing machine learning-based methods for predicting stroke patients’health status have made significant progress,limitations remain in terms of prediction accuracy,model explainability,and system optimization.This paper proposes a multi-task learning approach based on Explainable Artificial Intelligence(XAI)for predicting the health status of stroke patients.First,we design a comprehensive multi-task learning framework that utilizes the task correlation of predicting various health status indicators in patients,enabling the parallel prediction of multiple health indicators.Second,we develop a multi-task Area Under Curve(AUC)optimization algorithm based on adaptive low-rank representation,which removes irrelevant information from the model structure to enhance the performance of multi-task AUC optimization.Additionally,the model’s explainability is analyzed through the stability analysis of SHAP values.Experimental results demonstrate that our approach outperforms comparison algorithms in key prognostic metrics F1 score and Efficiency.展开更多
Skin cancer is among the most common malignancies worldwide,but its mortality burden is largely driven by aggressive subtypes such as melanoma,with outcomes varying across regions and healthcare settings.These variati...Skin cancer is among the most common malignancies worldwide,but its mortality burden is largely driven by aggressive subtypes such as melanoma,with outcomes varying across regions and healthcare settings.These variations emphasize the importance of reliable diagnostic technologies that support clinicians in detecting skin malignancies with higher accuracy.Traditional diagnostic methods often rely on subjective visual assessments,which can lead to misdiagnosis.This study addresses these challenges by developing HybridFusionNet,a novel model that integrates Convolutional Neural Networks(CNN)with 1D feature extraction techniques to enhance diagnostic accuracy.Utilizing two extensive datasets,BCN20000 and HAM10000,the methodology includes data preprocessing,application of Synthetic Minority Oversampling Technique combined with Edited Nearest Neighbors(SMOTEENN)for data balancing,and optimization of feature selection using the Tree-based Pipeline Optimization Tool(TPOT).The results demonstrate significant performance improvements over traditional CNN models,achieving an accuracy of 0.9693 on the BCN20000 dataset and 0.9909 on the HAM10000 dataset.The HybridFusionNet model not only outperforms conventionalmethods but also effectively addresses class imbalance.To enhance transparency,it integrates post-hoc explanation techniques such as LIME,which highlight the features influencing predictions.These findings highlight the potential of HybridFusionNet to support real-world applications,including physician-assist systems,teledermatology,and large-scale skin cancer screening programs.By improving diagnostic efficiency and enabling access to expert-level analysis,the modelmay enhance patient outcomes and foster greater trust in artificial intelligence(AI)-assisted clinical decision-making.展开更多
Tungsten carbide-based(WC-based)cemented carbides are widely recognized as high-performance tool materials.Traditionally,single metals such as cobalt(Co)or nickel(Ni)serve as the binder phase,providing toughness and s...Tungsten carbide-based(WC-based)cemented carbides are widely recognized as high-performance tool materials.Traditionally,single metals such as cobalt(Co)or nickel(Ni)serve as the binder phase,providing toughness and structural integrity.Replacing this phase with high-entropy alloys(HEAs)offers a promising approach to enhancing mechanical properties and addressing sustainability challenges.However,the complex multi-element composition of HEAs complicates conventional experimental design,making it difficult to explore the vast compositional space efficiently.Traditional trial-and-error methods are time-consuming,resource-intensive,and often ineffective in identifying optimal compositions.In contrast,artificial intelligence(AI)-driven approaches enable rapid screening and optimization of alloy compositions,significantly improving predictive accuracy and interpretability.Feature selection techniques were employed to identify key alloying elements influencing hardness,toughness,and wear resistance.To enhance model interpretability,explainable artificial intelligence(XAI)techniques—SHapley Additive exPlanations(SHAP)and Local Interpretable Model-agnostic Explanations(LIME)—were applied to quantify the contributions of individual elements and uncover complex elemental interactions.Furthermore,a high-throughput machine learning(ML)–driven screening approach was implemented to optimize the binder phase composition,facilitating the discovery of HEAs with superiormechanical properties.Experimental validation demonstrated strong agreement between model predictions and measured performance,confirming the reliability of the ML framework.This study underscores the potential of integrating ML and XAI for data-driven materials design,providing a novel strategy for optimizing high-entropy cemented carbides.展开更多
With the rapid popularization of artificial intelligence technology in the field of higher education,college students are increasingly dependent on AI tools such as ChatGPT,automatic writing assistants,and intelligent...With the rapid popularization of artificial intelligence technology in the field of higher education,college students are increasingly dependent on AI tools such as ChatGPT,automatic writing assistants,and intelligent translators.Behind the convenience and efficiency,a decline trend in students’core learning abilities such as autonomous learning ability,critical thinking ability,and knowledge construction ability has gradually emerged.This study aims to explore the interactive logical mechanism between college students’reliance on AI tools and the weakening of their learning abilities,and on this basis,propose practical and feasible educational intervention strategies.Research has found that while AI tools lower the learning threshold,they also weaken students’cognitive investment and independent thinking abilities,further intensifying their reliance on technology.In this regard,this paper proposes a three-dimensional intervention path based on guided usage,ability compensation,and value reconstruction to achieve the collaborative improvement of students’technical usage ability and learning ability.This research has certain theoretical value and practical enlightenment significance for solving the structural predicament of higher education in the intelligent era.展开更多
Blood cell disorders are among the leading causes of serious diseases such as leukemia,anemia,blood clotting disorders,and immune-related conditions.The global incidence of hematological diseases is increasing,affecti...Blood cell disorders are among the leading causes of serious diseases such as leukemia,anemia,blood clotting disorders,and immune-related conditions.The global incidence of hematological diseases is increasing,affecting both children and adults.In clinical practice,blood smear analysis is still largely performed manually,relying heavily on the experience and expertise of laboratory technicians or hematologists.This manual process introduces risks of diagnostic errors,especially in cases with rare or morphologically ambiguous cells.The situation is more critical in developing countries,where there is a shortage of specialized medical personnel and limited access to modern diagnostic tools.High testing costs and delays in diagnosis hinder access to quality healthcare services.In this context,the integration of Artificial Intelligence(AI),particularly Explainable AI(XAI)based on deep learning,offers a promising solution for improving the accuracy,efficiency,and transparency of hematological diagnostics.In this study,we propose a Ghost Residual Network(GRsNet)integrated with XAI techniques such as Gradient-weighted Class Activation Mapping(Grad-CAM),Local Interpretable Model-Agnostic Explanations(LIME),and SHapley Additive exPlanations(SHAP)for automatic blood cell classification.These techniques provide visual explanations by highlighting important regions in the input images,thereby supporting clinical decision-making.The proposed model is evaluated on two public datasets:Naturalize 2K-PBC and Microscopic Blood Cell,achieving a classification accuracy of up to 95%.The results demonstrate the model’s strong potential for automated hematological diagnosis,particularly in resource-constrained settings.It not only enhances diagnostic reliability but also contributes to advancing digital transformation and equitable access to AI-driven healthcare in developing regions.展开更多
Voice, motion, and mimicry are naturalistic control modalities that have replaced text or display-driven control in human-computer communication (HCC). Specifically, the vocals contain a lot of knowledge, revealing de...Voice, motion, and mimicry are naturalistic control modalities that have replaced text or display-driven control in human-computer communication (HCC). Specifically, the vocals contain a lot of knowledge, revealing details about the speaker’s goals and desires, as well as their internal condition. Certain vocal characteristics reveal the speaker’s mood, intention, and motivation, while word study assists the speaker’s demand to be understood. Voice emotion recognition has become an essential component of modern HCC networks. Integrating findings from the various disciplines involved in identifying vocal emotions is also challenging. Many sound analysis techniques were developed in the past. Learning about the development of artificial intelligence (AI), and especially Deep Learning (DL) technology, research incorporating real data is becoming increasingly common these days. Thus, this research presents a novel selfish herd optimization-tuned long/short-term memory (SHO-LSTM) strategy to identify vocal emotions in human communication. The RAVDESS public dataset is used to train the suggested SHO-LSTM technique. Mel-frequency cepstral coefficient (MFCC) and wiener filter (WF) techniques are used, respectively, to remove noise and extract features from the data. LSTM and SHO are applied to the extracted data to optimize the LSTM network’s parameters for effective emotion recognition. Python Software was used to execute our proposed framework. In the finding assessment phase, Numerous metrics are used to evaluate the proposed model’s detection capability, Such as F1-score (95%), precision (95%), recall (96%), and accuracy (97%). The suggested approach is tested on a Python platform, and the SHO-LSTM’s outcomes are contrasted with those of other previously conducted research. Based on comparative assessments, our suggested approach outperforms the current approaches in vocal emotion recognition.展开更多
In this study,twelve machine learning(ML)techniques are used to accurately estimate the safety factor of rock slopes(SFRS).The dataset used for developing these models consists of 344 rock slopes from various open-pit...In this study,twelve machine learning(ML)techniques are used to accurately estimate the safety factor of rock slopes(SFRS).The dataset used for developing these models consists of 344 rock slopes from various open-pit mines around Iran,evenly distributed between the training(80%)and testing(20%)datasets.The models are evaluated for accuracy using Janbu's limit equilibrium method(LEM)and commercial tool GeoStudio methods.Statistical assessment metrics show that the random forest model is the most accurate in estimating the SFRS(MSE=0.0182,R2=0.8319)and shows high agreement with the results from the LEM method.The results from the long-short-term memory(LSTM)model are the least accurate(MSE=0.037,R2=0.6618)of all the models tested.However,only the null space support vector regression(NuSVR)model performs accurately compared to the practice mode by altering the value of one parameter while maintaining the other parameters constant.It is suggested that this model would be the best one to use to calculate the SFRS.A graphical user interface for the proposed models is developed to further assist in the calculation of the SFRS for engineering difficulties.In this study,we attempt to bridge the gap between modern slope stability evaluation techniques and more conventional analysis methods.展开更多
Sentiment Analysis,a significant domain within Natural Language Processing(NLP),focuses on extracting and interpreting subjective information-such as emotions,opinions,and attitudes-from textual data.With the increasi...Sentiment Analysis,a significant domain within Natural Language Processing(NLP),focuses on extracting and interpreting subjective information-such as emotions,opinions,and attitudes-from textual data.With the increasing volume of user-generated content on social media and digital platforms,sentiment analysis has become essential for deriving actionable insights across various sectors.This study presents a systematic literature review of sentiment analysis methodologies,encompassing traditional machine learning algorithms,lexicon-based approaches,and recent advancements in deep learning techniques.The review follows a structured protocol comprising three phases:planning,execution,and analysis/reporting.During the execution phase,67 peer-reviewed articles were initially retrieved,with 25 meeting predefined inclusion and exclusion criteria.The analysis phase involved a detailed examination of each study’s methodology,experimental setup,and key contributions.Among the deep learning models evaluated,Long Short-Term Memory(LSTM)networks were identified as the most frequently adopted architecture for sentiment classification tasks.This review highlights current trends,technical challenges,and emerging opportunities in the field,providing valuable guidance for future research and development in applications such as market analysis,public health monitoring,financial forecasting,and crisis management.展开更多
The rapid evolution of smart cities through IoT,cloud computing,and connected infrastructures has significantly enhanced sectors such as transportation,healthcare,energy,and public safety,but also increased exposure t...The rapid evolution of smart cities through IoT,cloud computing,and connected infrastructures has significantly enhanced sectors such as transportation,healthcare,energy,and public safety,but also increased exposure to sophisticated cyber threats.The diversity of devices,high data volumes,and real-time operational demands complicate security,requiring not just robust intrusion detection but also effective feature selection for relevance and scalability.Traditional Machine Learning(ML)based Intrusion Detection System(IDS)improves detection but often lacks interpretability,limiting stakeholder trust and timely responses.Moreover,centralized feature selection in conventional IDS compromises data privacy and fails to accommodate the decentralized nature of smart city infrastructures.To address these limitations,this research introduces an Interpretable Federated Learning(FL)based Cyber Intrusion Detection model tailored for smart city applications.The proposed system leverages privacy-preserving feature selection,where each client node independently identifies top-ranked features using ML models integrated with SHAP-based explainability.These local feature subsets are then aggregated at a central server to construct a global model without compromising sensitive data.Furthermore,the global model is enhanced with Explainable AI(XAI)techniques such as SHAP and LIME,offering both global interpretability and instance-level transparency for cyber threat decisions.Experimental results demonstrate that the proposed global model achieves a high detection accuracy of 98.51%,with a significantly low miss rate of 1.49%,outperforming existing models while ensuring explainability,privacy,and scalability across smart city infrastructures.展开更多
Human Activity Recognition(HAR)represents a rapidly advancing research domain,propelled by continuous developments in sensor technologies and the Internet of Things(IoT).Deep learning has become the dominant paradigm ...Human Activity Recognition(HAR)represents a rapidly advancing research domain,propelled by continuous developments in sensor technologies and the Internet of Things(IoT).Deep learning has become the dominant paradigm in sensor-based HAR systems,offering significant advantages over traditional machine learning methods by eliminating manual feature extraction,enhancing recognition accuracy for complex activities,and enabling the exploitation of unlabeled data through generative models.This paper provides a comprehensive review of recent advancements and emerging trends in deep learning models developed for sensor-based human activity recognition(HAR)systems.We begin with an overview of fundamental HAR concepts in sensor-driven contexts,followed by a systematic categorization and summary of existing research.Our survey encompasses a wide range of deep learning approaches,including Multi-Layer Perceptrons(MLP),Convolutional Neural Networks(CNN),Recurrent Neural Networks(RNN),Long Short-Term Memory networks(LSTM),Gated Recurrent Units(GRU),Transformers,Deep Belief Networks(DBN),and hybrid architectures.A comparative evaluation of these models is provided,highlighting their performance,architectural complexity,and contributions to the field.Beyond Centralized deep learning models,we examine the role of Federated Learning(FL)in HAR,highlighting current applications and research directions.Finally,we discuss the growing importance of Explainable Artificial Intelligence(XAI)in sensor-based HAR,reviewing recent studies that integrate interpretability methods to enhance transparency and trustworthiness in deep learning-based HAR systems.展开更多
Crystal structure prediction(CSP)is a foundational computational technique for determining the atomic arrangements of crystalline materials,especially under high-pressure conditions.While CSP plays a critical role in ...Crystal structure prediction(CSP)is a foundational computational technique for determining the atomic arrangements of crystalline materials,especially under high-pressure conditions.While CSP plays a critical role in materials science,traditional approaches often encounter significant challenges related to computational efficiency and scalability,particularly when applied to complex systems.Recent advances in machine learning(ML)have shown tremendous promise in addressing these limitations,enabling the rapid and accurate prediction of crystal structures across a wide range of chemical compositions and external conditions.This review provides a concise overview of recent progress in ML-assisted CSP methodologies,with a particular focus on machine learning potentials and generative models.By critically analyzing these advances,we highlight the transformative impact of ML in accelerating materials discovery,enhancing computational efficiency,and broadening the applicability of CSP.Additionally,we discuss emerging opportunities and challenges in this rapidly evolving field.展开更多
The detection of multiple trace analytes using single sensors is often impeded by the limited sensitivity of material and the interference form overlapping signals in complex mixtures.Here,we introduce an efficient an...The detection of multiple trace analytes using single sensors is often impeded by the limited sensitivity of material and the interference form overlapping signals in complex mixtures.Here,we introduce an efficient and durable heterostructured high-entropy alloy(HEA)material,where non-noble HEA nanoparticles are used to disperse and stabilize Pt clusters(denoted as HEA@Pt).The HEA@Pt exhibits high sensitivity to three trace analytes(dopamine,uric acid,and paracetamol)during the electrochemical detection process,leveraging its multifunctional catalytic sensing capabilities for diverse mixtures.Additionally,to address the challenge of signal overlap,we integrate a recurrent neural network into multimodal sensing,combined with machine learning(ML)algorithms to accurately identify multiple analytes in mixtures.After five-fold cross-validation,the prediction accuracy deviations for dopamine,uric acid,and paracetamol were 3.20,9.18 and 3.84,respectively,with goodness-of-fit values of 0.984,0.992 and 0.990.The model achieved a prediction accuracy of 96.67%for unknown mixture samples,demonstrating robust generalization performance.This approach of multifunctional HEA combined with ML algorithms effectively overcomes detection errors caused by the complex detection of multiple chemical substances and the overlap of multiple response signals,thereby enabling accurate and reliable identification of multi-target analytes.展开更多
Federated Learning(FL)has recently emerged as a promising paradigm that enables medical institutions to collaboratively train robust models without centralizing sensitive patient information.Data collected from differ...Federated Learning(FL)has recently emerged as a promising paradigm that enables medical institutions to collaboratively train robust models without centralizing sensitive patient information.Data collected from different institutions represent distinct source domains.Consequently,discrepancies in feature distributions can significantly hinder a model’s generalization to unseen domains.While domain generalization(DG)methods have been proposed to address this challenge,many may compromise data privacy in FL by requiring clients to transmit their local feature representations to the server.Furthermore,existing adversarial training methods,commonly used to align marginal feature distributions,fail to ensure the consistency of conditional distributions.This consistency is often critical for accurate predictions in unseen domains.To address these limitations,we propose GPAF,a privacy-preserving federated learning(FL)framework that mitigates both domain and label shifts in healthcare applications.GPAF aligns conditional distributions across clients in the latent space and restricts communication to model parameters.This design preserves class semantics,enhances privacy,and improves communication efficiency.At the server,a global generator learns a conditional feature distribution from clients’feedback.During local training,each client minimizes an adversarial loss to align its local conditional distribution with the global distribution,enabling the FL model to learn robust,domain-invariant representations across all source domains.To evaluate the effectiveness of our approach,experiments on a medical imaging benchmark demonstrate that GPAF outperforms four FL baselines,achieving up to 17%higher classification accuracy and 25%faster convergence in non-IID scenarios.These results highlight GPAF’s capability to generalize across domains while maintaining strict privacy,offering a robust solution for decentralized healthcare challenges.展开更多
Introduction Deep learning(DL),as one of the most transformative technologies in artificial intelligence(AI),is undergoing a pivotal transition from laboratory research to industrial deployment.Advancing at an unprece...Introduction Deep learning(DL),as one of the most transformative technologies in artificial intelligence(AI),is undergoing a pivotal transition from laboratory research to industrial deployment.Advancing at an unprecedented pace,DL is transcending theoretical and application boundaries to penetrate emerging realworld scenarios such as industrial automation,urban management,and health monitoring,thereby driving a new wave of intelligent transformation.In August 2023,Goldman Sachs estimated that global AI investment will reach US$200 billion by 2025[1].However,the increasing complexity and dynamic nature of application scenarios expose critical challenges in traditional deep learning,including data heterogeneity,insufficient model generalization,computational resource constraints,and privacy-security trade-offs.The next generation of deep learning methodologies needs to achieve breakthroughs in multimodal fusion,lightweight design,interpretability enhancement,and cross-disciplinary collaborative optimization,in order to develop more efficient,robust,and practically valuable intelligent systems.展开更多
文摘As AI systems scale, the limitations of cloud-based architectures, including latency, bandwidth, and privacy concerns, demand decentralized alternatives. Federated learning (FL) and Edge AI provide a paradigm shift by combining privacy preserving training with efficient, on device computation. This paper introduces a cutting-edge FL-edge integration framework, achieving a 10% to 15% increase in model accuracy and reducing communication costs by 25% in heterogeneous environments. Blockchain based secure aggregation ensures robust and tamper-proof model updates, while exploratory quantum AI techniques enhance computational efficiency. By addressing key challenges such as device variability and non-IID data, this work sets the stage for the next generation of adaptive, privacy-first AI systems, with applications in IoT, healthcare, and autonomous systems.
文摘The fast increase of online communities has brought about an increase in cyber threats inclusive of cyberbullying, hate speech, misinformation, and online harassment, making content moderation a pressing necessity. Traditional single-modal AI-based detection systems, which analyze both text, photos, or movies in isolation, have established useless at taking pictures multi-modal threats, in which malicious actors spread dangerous content throughout a couple of formats. To cope with these demanding situations, we advise a multi-modal deep mastering framework that integrates Natural Language Processing (NLP), Convolutional Neural Networks (CNNs), and Long Short-Term Memory (LSTM) networks to become aware of and mitigate online threats effectively. Our proposed model combines BERT for text class, ResNet50 for photograph processing, and a hybrid LSTM-3-d CNN community for video content material analysis. We constructed a large-scale dataset comprising 500,000 textual posts, 200,000 offensive images, and 50,000 annotated motion pictures from more than one platform, which includes Twitter, Reddit, YouTube, and online gaming forums. The system became carefully evaluated using trendy gadget mastering metrics which include accuracy, precision, remember, F1-score, and ROC-AUC curves. Experimental outcomes demonstrate that our multi-modal method extensively outperforms single-modal AI classifiers, achieving an accuracy of 92.3%, precision of 91.2%, do not forget of 90.1%, and an AUC rating of 0.95. The findings validate the necessity of integrating multi-modal AI for actual-time, high-accuracy online chance detection and moderation. Future paintings will have consciousness on improving hostile robustness, enhancing scalability for real-world deployment, and addressing ethical worries associated with AI-driven content moderation.
文摘Membrane fouling is a persistent challenge in membrane-based technologies,significantly impacting efficiency,operational costs,and system lifespan in applications like water treatment,desalination,and industrial processing.Foul-ing,caused by the accumulation of particulates,organic compounds,and microorganisms,leads to reduced permeability,increased energy demands,and frequent maintenance.Traditional fouling control approaches,relying on empirical models and reactive strategies,often fail to address these issues efficiently.In this context,artificial intelligence(AI)and machine learning(ML)have emerged as innovative tools offering predictive and proactive solutions for fouling man-agement.By utilizing historical and real-time data,AI/ML techniques such as artificial neural networks,support vector machines,and ensemble models enable accurate prediction of fouling onset,identification of fouling mechanisms,and optimization of control measures.This review provides a detailed examination of the integration of AI/ML in membrane fouling prediction and mitigation,discussing advanced algorithms,the role of sensor-based monitoring,and the importance of robust datasets in enhancing predictive accuracy.Case studies highlighting successful AI/ML applications across various membrane processes are presented,demonstrating their transformative potential in improving system performance.Emerging trends,such as hybrid modeling and IoT-enabled smart systems,are explored,alongside a criti-cal analysis of research gaps and opportunities.This review emphasizes AI/ML as a cornerstone for sustainable,cost-effective membrane operations.
文摘Against the background of the continuous reform in medical education,biochemistry,as a fundamental medical course,maintains a close connection with clinical practice.However,under the traditional teaching model,the effectiveness of the“basic-clinical”connection is relatively poor,which hinders the improvement of educational outcomes.In the practical teaching of higher vocational medical education,the integration of the AI Case-Guided Learning System can enhance students’enthusiasm for knowledge exploration and effectively improve teaching quality.Starting from the perspective of the“basic-clinical”connection teaching in the biochemistry course,this paper analyzes the application value of the AI Case-Guided Learning System and proposes specific application strategies,aiming to accumulate experience for the innovation of biochemistry teaching.
文摘This study explores a novel educational model of generative AI-empowered interdisciplinary project-based learning(PBL).By analyzing the current applications of generative AI technology in information technology curricula,it elucidates its advantages and operational mechanisms in interdisciplinary PBL.Combining case studies and empirical research,the investigation proposes implementation pathways and strategies for the generative AI-enhanced interdisciplinary PBL model,detailing specific applications across three phases:project preparation,implementation,and evaluation.The research demonstrates that generative AI-enabled interdisciplinary project-based learning can effectively enhance students’learning motivation,interdisciplinary thinking capabilities,and innovative competencies,providing new conceptual frameworks and practical approaches for educational model innovation.
文摘In the evolving landscape of cyber threats,phishing attacks pose significant challenges,particularly through deceptive webpages designed to extract sensitive information under the guise of legitimacy.Conventional and machine learning(ML)-based detection systems struggle to detect phishing websites owing to their constantly changing tactics.Furthermore,newer phishing websites exhibit subtle and expertly concealed indicators that are not readily detectable.Hence,effective detection depends on identifying the most critical features.Traditional feature selection(FS)methods often struggle to enhance ML model performance and instead decrease it.To combat these issues,we propose an innovative method using explainable AI(XAI)to enhance FS in ML models and improve the identification of phishing websites.Specifically,we employ SHapley Additive exPlanations(SHAP)for global perspective and aggregated local interpretable model-agnostic explanations(LIME)to deter-mine specific localized patterns.The proposed SHAP and LIME-aggregated FS(SLA-FS)framework pinpoints the most informative features,enabling more precise,swift,and adaptable phishing detection.Applying this approach to an up-to-date web phishing dataset,we evaluate the performance of three ML models before and after FS to assess their effectiveness.Our findings reveal that random forest(RF),with an accuracy of 97.41%and XGBoost(XGB)at 97.21%significantly benefit from the SLA-FS framework,while k-nearest neighbors lags.Our framework increases the accuracy of RF and XGB by 0.65%and 0.41%,respectively,outperforming traditional filter or wrapper methods and any prior methods evaluated on this dataset,showcasing its potential.
基金funded by the Excellent Talent Training Funding Project in Dongcheng District,Beijing,with project number 2024-dchrcpyzz-9.
文摘Predicting the health status of stroke patients at different stages of the disease is a critical clinical task.The onset and development of stroke are affected by an array of factors,encompassing genetic predisposition,environmental exposure,unhealthy lifestyle habits,and existing medical conditions.Although existing machine learning-based methods for predicting stroke patients’health status have made significant progress,limitations remain in terms of prediction accuracy,model explainability,and system optimization.This paper proposes a multi-task learning approach based on Explainable Artificial Intelligence(XAI)for predicting the health status of stroke patients.First,we design a comprehensive multi-task learning framework that utilizes the task correlation of predicting various health status indicators in patients,enabling the parallel prediction of multiple health indicators.Second,we develop a multi-task Area Under Curve(AUC)optimization algorithm based on adaptive low-rank representation,which removes irrelevant information from the model structure to enhance the performance of multi-task AUC optimization.Additionally,the model’s explainability is analyzed through the stability analysis of SHAP values.Experimental results demonstrate that our approach outperforms comparison algorithms in key prognostic metrics F1 score and Efficiency.
文摘Skin cancer is among the most common malignancies worldwide,but its mortality burden is largely driven by aggressive subtypes such as melanoma,with outcomes varying across regions and healthcare settings.These variations emphasize the importance of reliable diagnostic technologies that support clinicians in detecting skin malignancies with higher accuracy.Traditional diagnostic methods often rely on subjective visual assessments,which can lead to misdiagnosis.This study addresses these challenges by developing HybridFusionNet,a novel model that integrates Convolutional Neural Networks(CNN)with 1D feature extraction techniques to enhance diagnostic accuracy.Utilizing two extensive datasets,BCN20000 and HAM10000,the methodology includes data preprocessing,application of Synthetic Minority Oversampling Technique combined with Edited Nearest Neighbors(SMOTEENN)for data balancing,and optimization of feature selection using the Tree-based Pipeline Optimization Tool(TPOT).The results demonstrate significant performance improvements over traditional CNN models,achieving an accuracy of 0.9693 on the BCN20000 dataset and 0.9909 on the HAM10000 dataset.The HybridFusionNet model not only outperforms conventionalmethods but also effectively addresses class imbalance.To enhance transparency,it integrates post-hoc explanation techniques such as LIME,which highlight the features influencing predictions.These findings highlight the potential of HybridFusionNet to support real-world applications,including physician-assist systems,teledermatology,and large-scale skin cancer screening programs.By improving diagnostic efficiency and enabling access to expert-level analysis,the modelmay enhance patient outcomes and foster greater trust in artificial intelligence(AI)-assisted clinical decision-making.
文摘Tungsten carbide-based(WC-based)cemented carbides are widely recognized as high-performance tool materials.Traditionally,single metals such as cobalt(Co)or nickel(Ni)serve as the binder phase,providing toughness and structural integrity.Replacing this phase with high-entropy alloys(HEAs)offers a promising approach to enhancing mechanical properties and addressing sustainability challenges.However,the complex multi-element composition of HEAs complicates conventional experimental design,making it difficult to explore the vast compositional space efficiently.Traditional trial-and-error methods are time-consuming,resource-intensive,and often ineffective in identifying optimal compositions.In contrast,artificial intelligence(AI)-driven approaches enable rapid screening and optimization of alloy compositions,significantly improving predictive accuracy and interpretability.Feature selection techniques were employed to identify key alloying elements influencing hardness,toughness,and wear resistance.To enhance model interpretability,explainable artificial intelligence(XAI)techniques—SHapley Additive exPlanations(SHAP)and Local Interpretable Model-agnostic Explanations(LIME)—were applied to quantify the contributions of individual elements and uncover complex elemental interactions.Furthermore,a high-throughput machine learning(ML)–driven screening approach was implemented to optimize the binder phase composition,facilitating the discovery of HEAs with superiormechanical properties.Experimental validation demonstrated strong agreement between model predictions and measured performance,confirming the reliability of the ML framework.This study underscores the potential of integrating ML and XAI for data-driven materials design,providing a novel strategy for optimizing high-entropy cemented carbides.
基金The 2024 Higher Education Teaching Reform Project of Guangdong University of Science and Technology,“Teaching Practice of Human Resource Management Course Based on SPOC+FC Hybrid Teaching Mode”(GKZLGC2024024)。
文摘With the rapid popularization of artificial intelligence technology in the field of higher education,college students are increasingly dependent on AI tools such as ChatGPT,automatic writing assistants,and intelligent translators.Behind the convenience and efficiency,a decline trend in students’core learning abilities such as autonomous learning ability,critical thinking ability,and knowledge construction ability has gradually emerged.This study aims to explore the interactive logical mechanism between college students’reliance on AI tools and the weakening of their learning abilities,and on this basis,propose practical and feasible educational intervention strategies.Research has found that while AI tools lower the learning threshold,they also weaken students’cognitive investment and independent thinking abilities,further intensifying their reliance on technology.In this regard,this paper proposes a three-dimensional intervention path based on guided usage,ability compensation,and value reconstruction to achieve the collaborative improvement of students’technical usage ability and learning ability.This research has certain theoretical value and practical enlightenment significance for solving the structural predicament of higher education in the intelligent era.
文摘Blood cell disorders are among the leading causes of serious diseases such as leukemia,anemia,blood clotting disorders,and immune-related conditions.The global incidence of hematological diseases is increasing,affecting both children and adults.In clinical practice,blood smear analysis is still largely performed manually,relying heavily on the experience and expertise of laboratory technicians or hematologists.This manual process introduces risks of diagnostic errors,especially in cases with rare or morphologically ambiguous cells.The situation is more critical in developing countries,where there is a shortage of specialized medical personnel and limited access to modern diagnostic tools.High testing costs and delays in diagnosis hinder access to quality healthcare services.In this context,the integration of Artificial Intelligence(AI),particularly Explainable AI(XAI)based on deep learning,offers a promising solution for improving the accuracy,efficiency,and transparency of hematological diagnostics.In this study,we propose a Ghost Residual Network(GRsNet)integrated with XAI techniques such as Gradient-weighted Class Activation Mapping(Grad-CAM),Local Interpretable Model-Agnostic Explanations(LIME),and SHapley Additive exPlanations(SHAP)for automatic blood cell classification.These techniques provide visual explanations by highlighting important regions in the input images,thereby supporting clinical decision-making.The proposed model is evaluated on two public datasets:Naturalize 2K-PBC and Microscopic Blood Cell,achieving a classification accuracy of up to 95%.The results demonstrate the model’s strong potential for automated hematological diagnosis,particularly in resource-constrained settings.It not only enhances diagnostic reliability but also contributes to advancing digital transformation and equitable access to AI-driven healthcare in developing regions.
基金The author Dr.Arshiya S.Ansari extends the appreciation to the Deanship of Postgraduate Studies and Scientific Research at Majmaah University for funding this research work through the project number(R-2025-1538).
文摘Voice, motion, and mimicry are naturalistic control modalities that have replaced text or display-driven control in human-computer communication (HCC). Specifically, the vocals contain a lot of knowledge, revealing details about the speaker’s goals and desires, as well as their internal condition. Certain vocal characteristics reveal the speaker’s mood, intention, and motivation, while word study assists the speaker’s demand to be understood. Voice emotion recognition has become an essential component of modern HCC networks. Integrating findings from the various disciplines involved in identifying vocal emotions is also challenging. Many sound analysis techniques were developed in the past. Learning about the development of artificial intelligence (AI), and especially Deep Learning (DL) technology, research incorporating real data is becoming increasingly common these days. Thus, this research presents a novel selfish herd optimization-tuned long/short-term memory (SHO-LSTM) strategy to identify vocal emotions in human communication. The RAVDESS public dataset is used to train the suggested SHO-LSTM technique. Mel-frequency cepstral coefficient (MFCC) and wiener filter (WF) techniques are used, respectively, to remove noise and extract features from the data. LSTM and SHO are applied to the extracted data to optimize the LSTM network’s parameters for effective emotion recognition. Python Software was used to execute our proposed framework. In the finding assessment phase, Numerous metrics are used to evaluate the proposed model’s detection capability, Such as F1-score (95%), precision (95%), recall (96%), and accuracy (97%). The suggested approach is tested on a Python platform, and the SHO-LSTM’s outcomes are contrasted with those of other previously conducted research. Based on comparative assessments, our suggested approach outperforms the current approaches in vocal emotion recognition.
基金supported via funding from Prince Satam bin Abdulaziz University project number (PSAU/2024/R/1445)The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University for funding this work through large Group Research Project (Grant No.RGP.2/357/44).
文摘In this study,twelve machine learning(ML)techniques are used to accurately estimate the safety factor of rock slopes(SFRS).The dataset used for developing these models consists of 344 rock slopes from various open-pit mines around Iran,evenly distributed between the training(80%)and testing(20%)datasets.The models are evaluated for accuracy using Janbu's limit equilibrium method(LEM)and commercial tool GeoStudio methods.Statistical assessment metrics show that the random forest model is the most accurate in estimating the SFRS(MSE=0.0182,R2=0.8319)and shows high agreement with the results from the LEM method.The results from the long-short-term memory(LSTM)model are the least accurate(MSE=0.037,R2=0.6618)of all the models tested.However,only the null space support vector regression(NuSVR)model performs accurately compared to the practice mode by altering the value of one parameter while maintaining the other parameters constant.It is suggested that this model would be the best one to use to calculate the SFRS.A graphical user interface for the proposed models is developed to further assist in the calculation of the SFRS for engineering difficulties.In this study,we attempt to bridge the gap between modern slope stability evaluation techniques and more conventional analysis methods.
基金supported by the“Technology Commercialization Collaboration Platform Construction”project of the Innopolis Foundation(Project Number:2710033536)the Competitive Research Fund of The University of Aizu,Japan.
文摘Sentiment Analysis,a significant domain within Natural Language Processing(NLP),focuses on extracting and interpreting subjective information-such as emotions,opinions,and attitudes-from textual data.With the increasing volume of user-generated content on social media and digital platforms,sentiment analysis has become essential for deriving actionable insights across various sectors.This study presents a systematic literature review of sentiment analysis methodologies,encompassing traditional machine learning algorithms,lexicon-based approaches,and recent advancements in deep learning techniques.The review follows a structured protocol comprising three phases:planning,execution,and analysis/reporting.During the execution phase,67 peer-reviewed articles were initially retrieved,with 25 meeting predefined inclusion and exclusion criteria.The analysis phase involved a detailed examination of each study’s methodology,experimental setup,and key contributions.Among the deep learning models evaluated,Long Short-Term Memory(LSTM)networks were identified as the most frequently adopted architecture for sentiment classification tasks.This review highlights current trends,technical challenges,and emerging opportunities in the field,providing valuable guidance for future research and development in applications such as market analysis,public health monitoring,financial forecasting,and crisis management.
文摘The rapid evolution of smart cities through IoT,cloud computing,and connected infrastructures has significantly enhanced sectors such as transportation,healthcare,energy,and public safety,but also increased exposure to sophisticated cyber threats.The diversity of devices,high data volumes,and real-time operational demands complicate security,requiring not just robust intrusion detection but also effective feature selection for relevance and scalability.Traditional Machine Learning(ML)based Intrusion Detection System(IDS)improves detection but often lacks interpretability,limiting stakeholder trust and timely responses.Moreover,centralized feature selection in conventional IDS compromises data privacy and fails to accommodate the decentralized nature of smart city infrastructures.To address these limitations,this research introduces an Interpretable Federated Learning(FL)based Cyber Intrusion Detection model tailored for smart city applications.The proposed system leverages privacy-preserving feature selection,where each client node independently identifies top-ranked features using ML models integrated with SHAP-based explainability.These local feature subsets are then aggregated at a central server to construct a global model without compromising sensitive data.Furthermore,the global model is enhanced with Explainable AI(XAI)techniques such as SHAP and LIME,offering both global interpretability and instance-level transparency for cyber threat decisions.Experimental results demonstrate that the proposed global model achieves a high detection accuracy of 98.51%,with a significantly low miss rate of 1.49%,outperforming existing models while ensuring explainability,privacy,and scalability across smart city infrastructures.
文摘Human Activity Recognition(HAR)represents a rapidly advancing research domain,propelled by continuous developments in sensor technologies and the Internet of Things(IoT).Deep learning has become the dominant paradigm in sensor-based HAR systems,offering significant advantages over traditional machine learning methods by eliminating manual feature extraction,enhancing recognition accuracy for complex activities,and enabling the exploitation of unlabeled data through generative models.This paper provides a comprehensive review of recent advancements and emerging trends in deep learning models developed for sensor-based human activity recognition(HAR)systems.We begin with an overview of fundamental HAR concepts in sensor-driven contexts,followed by a systematic categorization and summary of existing research.Our survey encompasses a wide range of deep learning approaches,including Multi-Layer Perceptrons(MLP),Convolutional Neural Networks(CNN),Recurrent Neural Networks(RNN),Long Short-Term Memory networks(LSTM),Gated Recurrent Units(GRU),Transformers,Deep Belief Networks(DBN),and hybrid architectures.A comparative evaluation of these models is provided,highlighting their performance,architectural complexity,and contributions to the field.Beyond Centralized deep learning models,we examine the role of Federated Learning(FL)in HAR,highlighting current applications and research directions.Finally,we discuss the growing importance of Explainable Artificial Intelligence(XAI)in sensor-based HAR,reviewing recent studies that integrate interpretability methods to enhance transparency and trustworthiness in deep learning-based HAR systems.
基金supported by the National Key Research and Development Program of China(Grant No.2022YFA1402304)the National Natural Science Foundation of China(Grant Nos.12034009,12374005,52288102,52090024,and T2225013)+1 种基金the Fundamental Research Funds for the Central Universitiesthe Program for JLU Science and Technology Innovative Research Team.
文摘Crystal structure prediction(CSP)is a foundational computational technique for determining the atomic arrangements of crystalline materials,especially under high-pressure conditions.While CSP plays a critical role in materials science,traditional approaches often encounter significant challenges related to computational efficiency and scalability,particularly when applied to complex systems.Recent advances in machine learning(ML)have shown tremendous promise in addressing these limitations,enabling the rapid and accurate prediction of crystal structures across a wide range of chemical compositions and external conditions.This review provides a concise overview of recent progress in ML-assisted CSP methodologies,with a particular focus on machine learning potentials and generative models.By critically analyzing these advances,we highlight the transformative impact of ML in accelerating materials discovery,enhancing computational efficiency,and broadening the applicability of CSP.Additionally,we discuss emerging opportunities and challenges in this rapidly evolving field.
基金supported by the National Natural Science Foundation of China(Nos.52422505,12274124,and 52401280)the National Key R&D Program of China(No.2021YFA1202300)+1 种基金the Shanghai Pilot Program for Basic Research(No.22TQ1400100-6)the Fundamental Research Funds for the Central Universities,and the Innovative Research Group Project of the National Natural Science Foundation of China(No.52321002).
文摘The detection of multiple trace analytes using single sensors is often impeded by the limited sensitivity of material and the interference form overlapping signals in complex mixtures.Here,we introduce an efficient and durable heterostructured high-entropy alloy(HEA)material,where non-noble HEA nanoparticles are used to disperse and stabilize Pt clusters(denoted as HEA@Pt).The HEA@Pt exhibits high sensitivity to three trace analytes(dopamine,uric acid,and paracetamol)during the electrochemical detection process,leveraging its multifunctional catalytic sensing capabilities for diverse mixtures.Additionally,to address the challenge of signal overlap,we integrate a recurrent neural network into multimodal sensing,combined with machine learning(ML)algorithms to accurately identify multiple analytes in mixtures.After five-fold cross-validation,the prediction accuracy deviations for dopamine,uric acid,and paracetamol were 3.20,9.18 and 3.84,respectively,with goodness-of-fit values of 0.984,0.992 and 0.990.The model achieved a prediction accuracy of 96.67%for unknown mixture samples,demonstrating robust generalization performance.This approach of multifunctional HEA combined with ML algorithms effectively overcomes detection errors caused by the complex detection of multiple chemical substances and the overlap of multiple response signals,thereby enabling accurate and reliable identification of multi-target analytes.
文摘Federated Learning(FL)has recently emerged as a promising paradigm that enables medical institutions to collaboratively train robust models without centralizing sensitive patient information.Data collected from different institutions represent distinct source domains.Consequently,discrepancies in feature distributions can significantly hinder a model’s generalization to unseen domains.While domain generalization(DG)methods have been proposed to address this challenge,many may compromise data privacy in FL by requiring clients to transmit their local feature representations to the server.Furthermore,existing adversarial training methods,commonly used to align marginal feature distributions,fail to ensure the consistency of conditional distributions.This consistency is often critical for accurate predictions in unseen domains.To address these limitations,we propose GPAF,a privacy-preserving federated learning(FL)framework that mitigates both domain and label shifts in healthcare applications.GPAF aligns conditional distributions across clients in the latent space and restricts communication to model parameters.This design preserves class semantics,enhances privacy,and improves communication efficiency.At the server,a global generator learns a conditional feature distribution from clients’feedback.During local training,each client minimizes an adversarial loss to align its local conditional distribution with the global distribution,enabling the FL model to learn robust,domain-invariant representations across all source domains.To evaluate the effectiveness of our approach,experiments on a medical imaging benchmark demonstrate that GPAF outperforms four FL baselines,achieving up to 17%higher classification accuracy and 25%faster convergence in non-IID scenarios.These results highlight GPAF’s capability to generalize across domains while maintaining strict privacy,offering a robust solution for decentralized healthcare challenges.
基金supported in part by Guangdong Basic and Applied Basic Research Foundation under Grant 2024A1515012485in part by Shenzhen Fundamental Research Program under Grant JCYJ20220810112354002+4 种基金in part by Shenzhen Science and Technology Program under Grant KJZD20230923114111021in part by the Fund for Academic Innovation Teams and Research Platform of South-Central Minzu University under Grant XTZ24003 and Grant PTZ24001in part by the Knowledge Innovation Program of Wuhan-Basic Research through Project 2023010201010151in part by the Research Start-up Funds of South-Central Minzu University under Grant YZZ18006in part by the Spring Sunshine Program of Ministry of Education of the People’s Republic of China under Grant HZKY20220331.
文摘Introduction Deep learning(DL),as one of the most transformative technologies in artificial intelligence(AI),is undergoing a pivotal transition from laboratory research to industrial deployment.Advancing at an unprecedented pace,DL is transcending theoretical and application boundaries to penetrate emerging realworld scenarios such as industrial automation,urban management,and health monitoring,thereby driving a new wave of intelligent transformation.In August 2023,Goldman Sachs estimated that global AI investment will reach US$200 billion by 2025[1].However,the increasing complexity and dynamic nature of application scenarios expose critical challenges in traditional deep learning,including data heterogeneity,insufficient model generalization,computational resource constraints,and privacy-security trade-offs.The next generation of deep learning methodologies needs to achieve breakthroughs in multimodal fusion,lightweight design,interpretability enhancement,and cross-disciplinary collaborative optimization,in order to develop more efficient,robust,and practically valuable intelligent systems.
