Background:While the treatment of metastatic renal cell carcinoma(mRCC)is evolving due to immune checkpoint inhibitors(ICIs),optimal strategies for later lines of therapy have yet to be defined.The combination of lenv...Background:While the treatment of metastatic renal cell carcinoma(mRCC)is evolving due to immune checkpoint inhibitors(ICIs),optimal strategies for later lines of therapy have yet to be defined.The combination of lenvatinib and everolimus represents a viable option,and the present review aimed to summarize its activity,effectiveness,and safety.Methods:A systematic review of the literature was conducted using PubMed,targeting studies published between 2018 and 2025.Eligible studies included English-language prospective and retrospective trials reporting survival outcomes in mRCC patients treated with lenvatinib and everolimus after at least one ICI-containing regimen.Results:Nine studies met the inclusion criteria,encompassing a total of 441 patients.The lenvatinib and everolimus combination was primarily used in the third and subsequent lines of therapy.Median overall survival ranged from 7.5 to 24.5 months,while median progression-free survival was more consistent,between 6.1 and 6.7 months,except for one study reporting 12.9 months.Objective response rates varied widely(14.0%–55.7%).Adverse events of grade≥3 did not exceed the expected rate,with diarrhoea and proteinuria as the most reported events.Dose reductions and treatment discontinuations due to toxicity occurred but were generally lower than in prior pivotal trials.Conclusions:Real-world evidence suggests that lenvatinib and everolimus represent an effective and safe option after ICI failure in mRCC patients.Nevertheless,the lack of randomized phase III trials and the heterogeneity of existing studies highlight the need for more robust prospective research to guide post-ICI therapeutic strategies.展开更多
This study demonstrates a novel integration of large language models,machine learning,and multicriteria decision-making to investigate self-moderation in small online communities,a topic under-explored compared to use...This study demonstrates a novel integration of large language models,machine learning,and multicriteria decision-making to investigate self-moderation in small online communities,a topic under-explored compared to user behavior and platform-driven moderation on social media.The proposed methodological framework(1)utilizes large language models for social media post analysis and categorization,(2)employs k-means clustering for content characterization,and(3)incorporates the TODIM(Tomada de Decisão Interativa Multicritério)method to determine moderation strategies based on expert judgments.In general,the fully integrated framework leverages the strengths of these intelligent systems in a more systematic evaluation of large-scale decision problems.When applied in social media moderation,this approach promotes nuanced and context-sensitive self-moderation by taking into account factors such as cultural background and geographic location.The application of this framework is demonstrated within Facebook groups.Eight distinct content clusters encompassing safety,harassment,diversity,and misinformation are identified.Analysis revealed a preference for content removal across all clusters,suggesting a cautious approach towards potentially harmful content.However,the framework also highlights the use of other moderation actions,like account suspension,depending on the content category.These findings contribute to the growing body of research on self-moderation and offer valuable insights for creating safer and more inclusive online spaces within smaller communities.展开更多
This study presents a comprehensive and secure architectural framework for the Internet of Medical Things(IoMT),integrating the foundational principles of the Confidentiality,Integrity,and Availability(CIA)triad along...This study presents a comprehensive and secure architectural framework for the Internet of Medical Things(IoMT),integrating the foundational principles of the Confidentiality,Integrity,and Availability(CIA)triad along with authentication mechanisms.Leveraging advanced Machine Learning(ML)and Deep Learning(DL)techniques,the proposed system is designed to safeguard Patient-Generated Health Data(PGHD)across interconnected medical devices.Given the increasing complexity and scale of cyber threats in IoMT environments,the integration of Intrusion Detection and Prevention Systems(IDPS)with intelligent analytics is critical.Our methodology employs both standalone and hybrid ML&DL models to automate threat detection and enable real-time analysis,while ensuring rapid and accurate responses to a diverse array of attacks.Emphasis is placed on systematic model evaluation using detection metrics such as accuracy,False Alarm Rate(FAR),and False Discovery Rate(FDR),with performance validation through cross-validation and statistical significance testing.Experimental results based on the Edge-IIoTset dataset demonstrate the superior performance of ensemble-based ML models such as Extreme Gradient Boosting(XGB)and hybrid DL models such as Convolutional Neural Networks with Autoencoders(CNN+AE),which achieved detection accuracies of 96%and 98%,respectively,with notably low FARs.These findings underscore the effectiveness of combining traditional security principles with advanced AI-driven methodologies to ensure secure,resilient,and trustworthy healthcare systems within the IoMT ecosystem.展开更多
This paper aims to provide a window opportunity to share a reflection and learning from different countries and from other disciplines with the focus on resilience.There is also an attempt to theorize the concept of l...This paper aims to provide a window opportunity to share a reflection and learning from different countries and from other disciplines with the focus on resilience.There is also an attempt to theorize the concept of learning from spurious success and failure in the context of COVID-19.The main emphasis is to provide understanding of the causal factors and the identification of improved measures and modelling approaches to prevent and mitigate against future pandemics.Proposed decision tools of resilience and bowtie modelling as enablers for decision makers to prevent hazards and protect against their consequences.展开更多
Advanced Persistent Threats(APTs)represent one of the most complex and dangerous categories of cyber-attacks characterised by their stealthy behaviour,long-term persistence,and ability to bypass traditional detection ...Advanced Persistent Threats(APTs)represent one of the most complex and dangerous categories of cyber-attacks characterised by their stealthy behaviour,long-term persistence,and ability to bypass traditional detection systems.The complexity of real-world network data poses significant challenges in detection.Machine learning models have shown promise in detecting APTs;however,their performance often suffers when trained on large datasets with redundant or irrelevant features.This study presents a novel,hybrid feature selection method designed to improve APT detection by reducing dimensionality while preserving the informative characteristics of the data.It combines Mutual Information(MI),Symmetric Uncertainty(SU)and Minimum Redundancy Maximum Relevance(mRMR)to enhance feature selection.MI and SU assess feature relevance,while mRMR maximises relevance and minimises redundancy,ensuring that the most impactful features are prioritised.This method addresses redundancy among selected features,improving the overall efficiency and effectiveness of the detection model.Experiments on a real-world APT datasets were conducted to evaluate the proposed method.Multiple classifiers including,Random Forest,Support Vector Machine(SVM),Gradient Boosting,and Neural Networks were used to assess classification performance.The results demonstrate that the proposed feature selection method significantly enhances detection accuracy compared to baseline models trained on the full feature set.The Random Forest algorithm achieved the highest performance,with near-perfect accuracy,precision,recall,and F1 scores(99.97%).The proposed adaptive thresholding algorithm within the selection method allows each classifier to benefit from a reduced and optimised feature space,resulting in improved training and predictive performance.This research offers a scalable and classifier-agnostic solution for dimensionality reduction in cybersecurity applications.展开更多
Image processing plays a vital role in various fields such as autonomous systems,healthcare,and cataloging,especially when integrated with deep learning(DL).It is crucial in medical diagnostics,including the early det...Image processing plays a vital role in various fields such as autonomous systems,healthcare,and cataloging,especially when integrated with deep learning(DL).It is crucial in medical diagnostics,including the early detection of diseases like chronic obstructive pulmonary disease(COPD),which claimed 3.2 million lives in 2015.COPD,a life-threatening condition often caused by prolonged exposure to lung irritants and smoking,progresses through stages.Early diagnosis through image processing can significantly improve survival rates.COPD encompasses chronic bronchitis(CB)and emphysema;CB particularly increases in smokers and generally affects individuals between 50 and 70 years old.It damages the lungs’air sacs,reducing oxygen transport and causing symptoms like coughing and shortness of breath.Treatments such as beta-agonists and inhaled steroids are used to manage symptoms and prolong lung function.Moreover,COVID-19 poses an additional risk to individuals with CB due to its impact on the respiratory system.The proposed system utilizes convolutional neural networks(CNN)to diagnose CB.In this system,CNN extracts essential and significant features from X-ray modalities,which are then fed into the neural network.The network undergoes training to recognize patterns and make accurate predictions based on the learned features.By leveraging DL techniques,the system aims to enhance the precision and reliability of CB detection.Our research specifically focuses on a subset of 189 lung disease images,carefully selected for model evaluation.To further refine the training process,various data augmentation and noise removal techniques are implemented.These techniques significantly enhance the quality of the training data,improving the model’s robustness and generalizability.As a result,the diagnostic accuracy has improved from 98.6%to 99.2%.This advancement not only validates the efficacy of our proposed model but also represents a significant improvement over existing literature.It highlights the potential of CNN-based approaches in transforming medical diagnostics through refined image analysis,learning capabilities,and automated feature extraction.展开更多
Let X be a real uniformly convex and uniformly smooth Banach space and C a nonempty closed and convex subset of X.Let Π_(C):X→C denote the generalized metric projection operator introduced by Alber in[1].In this pap...Let X be a real uniformly convex and uniformly smooth Banach space and C a nonempty closed and convex subset of X.Let Π_(C):X→C denote the generalized metric projection operator introduced by Alber in[1].In this paper,we define the Gâteaux directional differentiability of Π_(C).We investigate some properties of the Gâteaux directional differentiability of Π_(C).In particular,if C is a closed ball,or a closed and convex cone(including proper closed subspaces),or a closed and convex cylinder,then,we give the exact representations of the directional derivatives of Π_(C).By comparing the results in[12]and this paper,we see the significant difference between the directional derivatives of the generalized metric projection operator Π_(C) and the Gâteaux directional derivatives of the standard metric projection operator PC.展开更多
Drone photography is an essential building block of intelligent transportation,enabling wide-ranging monitoring,precise positioning,and rapid transmission.However,the high computational cost of transformer-based metho...Drone photography is an essential building block of intelligent transportation,enabling wide-ranging monitoring,precise positioning,and rapid transmission.However,the high computational cost of transformer-based methods in object detection tasks hinders real-time result transmission in drone target detection applications.Therefore,we propose mask adaptive transformer (MAT) tailored for such scenarios.Specifically,we introduce a structure that supports collaborative token sparsification in support windows,enhancing fault tolerance and reducing computational overhead.This structure comprises two modules:a binary mask strategy and adaptive window self-attention (A-WSA).The binary mask strategy focuses on significant objects in various complex scenes.The A-WSA mechanism is employed to self-attend for balance perfomance and computational cost to select objects and isolate all contextual leakage.Extensive experiments on the challenging CarPK and VisDrone datasets demonstrate the effectiveness and superiority of the proposed method.Specifically,it achieves a mean average precision (mAP@0.5) improvement of 1.25%over car detector based on you only look once version 5 (CD-YOLOv5) on the CarPK dataset and a 3.75%average precision(AP@0.5) improvement over cascaded zoom-in detector (CZ Det) on the VisDrone dataset.展开更多
0 INTRODUCTION Orogenic belts are commonly built by multiple-stage processes involving oceanic subduction and continental collisions that result in the generation of magma with distinct geochemical compositions,as exe...0 INTRODUCTION Orogenic belts are commonly built by multiple-stage processes involving oceanic subduction and continental collisions that result in the generation of magma with distinct geochemical compositions,as exemplified by Central Asian Orogenic Belts(e.g.,Wang et al.,2024;Yin et al.,2024;Xiao et al.,2005)and the Tethyan tectonic domains(e.g.,Chen et al.,2024;Li et al.,2024;Tao et al.,2024a;Gehrels et al.,2011;Yin and Harrison,2000).展开更多
The soil packing,influenced by variations in grain size and the gradation pattern within the soil matrix,plays a crucial role in constituting the mechanical properties of sandy soils.However,previous modeling approach...The soil packing,influenced by variations in grain size and the gradation pattern within the soil matrix,plays a crucial role in constituting the mechanical properties of sandy soils.However,previous modeling approaches have overlooked incorporating the full range of representative parameters to accurately predict the soaked California bearing ratio(CBR_(s))of sandy soils by precisely articulating soil packing in the modeling framework.This study presents an innovative artificial intelligence(AI)-based approach for modeling the CBR_(s)of sandy soils,considering grain size variability meticulously.By synthesizing extensive data from multiple sources,i.e.extensive tailored testing program undertaking multiple tests and extant literature,various modeling techniques including genetic expression programming(GEP),multi-expression programming(MEP),support vector machine(SVM),and multi-linear regression(MLR)are utilized to develop models.The research explores two modeling strategies,namely simplified and composite,with the former incorporating only sieve analysis test parameters,while the latter includes compaction test parameters alongside sieve analysis data.The models'performance is assessed using statistical key performance indicators(KPIs).Results indicate that genetic AI-based algorithms,particularly GEP,outperform SVM and conventional regression techniques,effectively capturing complex relationships between input parameters and CBR_(s).Additionally,the study reveals insights into model performance concerning the number of input parameters,with GEP consistently outperforming other models.External validation and Taylor diagram analysis demonstrate the GEP models'superiority over existing literature models on an independent dataset from the literature.Parametric and sensitivity analyses highlight the intricate relationships between grain sizes and CBR_(s),further emphasizing GEP's efficacy in modeling such complexities.This study contributes to enhancing CBR_(s)modeling accuracy for sandy soils,crucial for pertinent infrastructure design and construction rapidly and cost-effectively.展开更多
This paper proposes a lightweight reinforcement network (LRN) and auxiliary label distribution learning (ALDL)based robust facial expression recognition (FER) method.Our designed representation reinforcement (RR) netw...This paper proposes a lightweight reinforcement network (LRN) and auxiliary label distribution learning (ALDL)based robust facial expression recognition (FER) method.Our designed representation reinforcement (RR) network mainly comprises two modules,i.e.,the RR module and the auxiliary label space construction (ALSC) module.The RR module highlights key feature messaging nodes in feature maps,and ALSC allows multiple labels with different intensities to be linked to one expression.Therefore,LRN has a more robust feature extraction capability when model parameters are greatly reduced,and ALDL is proposed to contribute to the training effect of LRN in the condition of ambiguous training data.We tested our method on FER-Plus and RAF-DB datasets,and the experiment demonstrates the feasibility of our method in practice during rehabilitation robots.展开更多
The integration of endoscopy has significantly propelled the diagnosis and treatment of gastrointestinal diseases,with colonoscopy establishing itself as the primary method for early diagnosis and preventive care in c...The integration of endoscopy has significantly propelled the diagnosis and treatment of gastrointestinal diseases,with colonoscopy establishing itself as the primary method for early diagnosis and preventive care in colorectal cancer(CRC).Although deep learning holds promise in mitigating missed polyp rates,modern endoscopy examinations pose additional challenges,such as image blurring and atomizing.This study explores lightweight yet powerful attention mechanisms,introducing the spatial-channel transformer(SCT),an innovative approach that leverages spatial channel relationships for attention weight calculation.The method utilizes rotation operations for inter-dimensional dependencies,followed by residual transformation,encoding inter-channel and spatial information with minimal computational overhead.Extensive experiments on the CVC-Clinic DB polyp detection dataset,addressing endoscopy pitfalls,underscore the superiority of our SCT over other state-of-the-art methods.The proposed model maintains high performance,even in challenging scenarios.展开更多
In this work,we investigate a joint fitting approach based on theoretical models of power spectra associated with density-field reconstruction.Specifically,we consider the matter auto-power spectra before and after ba...In this work,we investigate a joint fitting approach based on theoretical models of power spectra associated with density-field reconstruction.Specifically,we consider the matter auto-power spectra before and after baryon acoustic oscillations(BAO)reconstruction,as well as the cross-power spectrum between the pre-and post-reconstructed density fields.We present redshift-space models for these three power spectra at the one-loop level within the framework of standard perturbation theory,and perform a joint analysis using three types of power spectra,and quantify their impact on parameter constraints.When restricting the analysis to wavenumbers k≤0.2 h Mpc^(−1)and adopting a smoothing scale of R_(s)=15 h^(−1)Mpc,we find that incorporating all three power spectra improves parameter constraints by approximately 11%–16%compared to using only the post-reconstruction power spectrum,with the Figure of Merit increasing by 10.5%.These results highlight the advantages of leveraging multiple power spectra in BAO reconstruction,ultimately enabling more precise cosmological parameter estimation.展开更多
High-plastic clays with significant volume change due to moisture variations present critical challenges to civil engineering structures.Limestone calcined clay cement(LC3),an innovative and sustainable hydraulic bind...High-plastic clays with significant volume change due to moisture variations present critical challenges to civil engineering structures.Limestone calcined clay cement(LC3),an innovative and sustainable hydraulic binder,demonstrates significant potential for improving the engineering characteristics of such soils.Nevertheless,the impact of LC3 on the physico-mechanical characteristics of treated soil under a cyclic wet-dry environment remains unclear.This study for the first time investigates LC3's impact on the long-term durability of treated high-plastic clays through comprehensive macro-micro testing including physical,mechanical,mineralogical,and microstructural investigations with an emphasis on wet-dry cycles.The results revealed that LC3 treatment exhibits significant resistance to wet-dry cycles by completely mitigating the swelling potential,and a considerable reduction in plasticity resulting in enhanced workability.The compressibility and shear strength parameters have been significantly improved to several orders of magnitude.However,after six wet-dry cycles,a slight to modest reduction is observed,but overall durability remains superior to untreated soil.Cohesive and structural bonding ratios quantitatively assessed the impact of wet-dry cycles emphasizing the advantage of LC3 treatment.According to mineralogical and microstructural evaluation,the mechanism behind the adverse effects of wet-dry cycles on the compressibility and strength behavior of LC3-treated soil is mainly attributed to:(1)weakening of CSH/C(A)SH and ettringite(AFt)phases by exhibiting lower peak intensities;and(2)larger pore spaces due to repeated wet-dry cycles.These findings highlight LC3's performance in enhancing the long-term behavior and resilience of treated soils in real-world scenarios,providing durable solutions for infrastructure challenges.展开更多
文摘Background:While the treatment of metastatic renal cell carcinoma(mRCC)is evolving due to immune checkpoint inhibitors(ICIs),optimal strategies for later lines of therapy have yet to be defined.The combination of lenvatinib and everolimus represents a viable option,and the present review aimed to summarize its activity,effectiveness,and safety.Methods:A systematic review of the literature was conducted using PubMed,targeting studies published between 2018 and 2025.Eligible studies included English-language prospective and retrospective trials reporting survival outcomes in mRCC patients treated with lenvatinib and everolimus after at least one ICI-containing regimen.Results:Nine studies met the inclusion criteria,encompassing a total of 441 patients.The lenvatinib and everolimus combination was primarily used in the third and subsequent lines of therapy.Median overall survival ranged from 7.5 to 24.5 months,while median progression-free survival was more consistent,between 6.1 and 6.7 months,except for one study reporting 12.9 months.Objective response rates varied widely(14.0%–55.7%).Adverse events of grade≥3 did not exceed the expected rate,with diarrhoea and proteinuria as the most reported events.Dose reductions and treatment discontinuations due to toxicity occurred but were generally lower than in prior pivotal trials.Conclusions:Real-world evidence suggests that lenvatinib and everolimus represent an effective and safe option after ICI failure in mRCC patients.Nevertheless,the lack of randomized phase III trials and the heterogeneity of existing studies highlight the need for more robust prospective research to guide post-ICI therapeutic strategies.
基金funded by the Office of the Vice-President for Research and Development of Cebu Technological University.
文摘This study demonstrates a novel integration of large language models,machine learning,and multicriteria decision-making to investigate self-moderation in small online communities,a topic under-explored compared to user behavior and platform-driven moderation on social media.The proposed methodological framework(1)utilizes large language models for social media post analysis and categorization,(2)employs k-means clustering for content characterization,and(3)incorporates the TODIM(Tomada de Decisão Interativa Multicritério)method to determine moderation strategies based on expert judgments.In general,the fully integrated framework leverages the strengths of these intelligent systems in a more systematic evaluation of large-scale decision problems.When applied in social media moderation,this approach promotes nuanced and context-sensitive self-moderation by taking into account factors such as cultural background and geographic location.The application of this framework is demonstrated within Facebook groups.Eight distinct content clusters encompassing safety,harassment,diversity,and misinformation are identified.Analysis revealed a preference for content removal across all clusters,suggesting a cautious approach towards potentially harmful content.However,the framework also highlights the use of other moderation actions,like account suspension,depending on the content category.These findings contribute to the growing body of research on self-moderation and offer valuable insights for creating safer and more inclusive online spaces within smaller communities.
基金funded by the Deanship of Graduate Studies and Scientific Research at Jouf University under Grant Number(DGSSR-2023-02-02516).
文摘This study presents a comprehensive and secure architectural framework for the Internet of Medical Things(IoMT),integrating the foundational principles of the Confidentiality,Integrity,and Availability(CIA)triad along with authentication mechanisms.Leveraging advanced Machine Learning(ML)and Deep Learning(DL)techniques,the proposed system is designed to safeguard Patient-Generated Health Data(PGHD)across interconnected medical devices.Given the increasing complexity and scale of cyber threats in IoMT environments,the integration of Intrusion Detection and Prevention Systems(IDPS)with intelligent analytics is critical.Our methodology employs both standalone and hybrid ML&DL models to automate threat detection and enable real-time analysis,while ensuring rapid and accurate responses to a diverse array of attacks.Emphasis is placed on systematic model evaluation using detection metrics such as accuracy,False Alarm Rate(FAR),and False Discovery Rate(FDR),with performance validation through cross-validation and statistical significance testing.Experimental results based on the Edge-IIoTset dataset demonstrate the superior performance of ensemble-based ML models such as Extreme Gradient Boosting(XGB)and hybrid DL models such as Convolutional Neural Networks with Autoencoders(CNN+AE),which achieved detection accuracies of 96%and 98%,respectively,with notably low FARs.These findings underscore the effectiveness of combining traditional security principles with advanced AI-driven methodologies to ensure secure,resilient,and trustworthy healthcare systems within the IoMT ecosystem.
基金supported by the project“Societal and Economic Resilience within multi-hazards environment in Romania”funded by European Union-Next generation EU and Romanian Government,under National Recovery and Resilience Plan for Romania,contract no.760050/23.05.2023,cod PNRR-C9-I8-CF 267/29.11.2022through the Romanian Ministry of Research,Innovation and Digitalization,within Component 9,Investment I8.
文摘This paper aims to provide a window opportunity to share a reflection and learning from different countries and from other disciplines with the focus on resilience.There is also an attempt to theorize the concept of learning from spurious success and failure in the context of COVID-19.The main emphasis is to provide understanding of the causal factors and the identification of improved measures and modelling approaches to prevent and mitigate against future pandemics.Proposed decision tools of resilience and bowtie modelling as enablers for decision makers to prevent hazards and protect against their consequences.
基金funded by Universiti Teknologi Malaysia under the UTM RA ICONIC Grant(Q.J130000.4351.09G61).
文摘Advanced Persistent Threats(APTs)represent one of the most complex and dangerous categories of cyber-attacks characterised by their stealthy behaviour,long-term persistence,and ability to bypass traditional detection systems.The complexity of real-world network data poses significant challenges in detection.Machine learning models have shown promise in detecting APTs;however,their performance often suffers when trained on large datasets with redundant or irrelevant features.This study presents a novel,hybrid feature selection method designed to improve APT detection by reducing dimensionality while preserving the informative characteristics of the data.It combines Mutual Information(MI),Symmetric Uncertainty(SU)and Minimum Redundancy Maximum Relevance(mRMR)to enhance feature selection.MI and SU assess feature relevance,while mRMR maximises relevance and minimises redundancy,ensuring that the most impactful features are prioritised.This method addresses redundancy among selected features,improving the overall efficiency and effectiveness of the detection model.Experiments on a real-world APT datasets were conducted to evaluate the proposed method.Multiple classifiers including,Random Forest,Support Vector Machine(SVM),Gradient Boosting,and Neural Networks were used to assess classification performance.The results demonstrate that the proposed feature selection method significantly enhances detection accuracy compared to baseline models trained on the full feature set.The Random Forest algorithm achieved the highest performance,with near-perfect accuracy,precision,recall,and F1 scores(99.97%).The proposed adaptive thresholding algorithm within the selection method allows each classifier to benefit from a reduced and optimised feature space,resulting in improved training and predictive performance.This research offers a scalable and classifier-agnostic solution for dimensionality reduction in cybersecurity applications.
文摘Image processing plays a vital role in various fields such as autonomous systems,healthcare,and cataloging,especially when integrated with deep learning(DL).It is crucial in medical diagnostics,including the early detection of diseases like chronic obstructive pulmonary disease(COPD),which claimed 3.2 million lives in 2015.COPD,a life-threatening condition often caused by prolonged exposure to lung irritants and smoking,progresses through stages.Early diagnosis through image processing can significantly improve survival rates.COPD encompasses chronic bronchitis(CB)and emphysema;CB particularly increases in smokers and generally affects individuals between 50 and 70 years old.It damages the lungs’air sacs,reducing oxygen transport and causing symptoms like coughing and shortness of breath.Treatments such as beta-agonists and inhaled steroids are used to manage symptoms and prolong lung function.Moreover,COVID-19 poses an additional risk to individuals with CB due to its impact on the respiratory system.The proposed system utilizes convolutional neural networks(CNN)to diagnose CB.In this system,CNN extracts essential and significant features from X-ray modalities,which are then fed into the neural network.The network undergoes training to recognize patterns and make accurate predictions based on the learned features.By leveraging DL techniques,the system aims to enhance the precision and reliability of CB detection.Our research specifically focuses on a subset of 189 lung disease images,carefully selected for model evaluation.To further refine the training process,various data augmentation and noise removal techniques are implemented.These techniques significantly enhance the quality of the training data,improving the model’s robustness and generalizability.As a result,the diagnostic accuracy has improved from 98.6%to 99.2%.This advancement not only validates the efficacy of our proposed model but also represents a significant improvement over existing literature.It highlights the potential of CNN-based approaches in transforming medical diagnostics through refined image analysis,learning capabilities,and automated feature extraction.
文摘Let X be a real uniformly convex and uniformly smooth Banach space and C a nonempty closed and convex subset of X.Let Π_(C):X→C denote the generalized metric projection operator introduced by Alber in[1].In this paper,we define the Gâteaux directional differentiability of Π_(C).We investigate some properties of the Gâteaux directional differentiability of Π_(C).In particular,if C is a closed ball,or a closed and convex cone(including proper closed subspaces),or a closed and convex cylinder,then,we give the exact representations of the directional derivatives of Π_(C).By comparing the results in[12]and this paper,we see the significant difference between the directional derivatives of the generalized metric projection operator Π_(C) and the Gâteaux directional derivatives of the standard metric projection operator PC.
文摘Drone photography is an essential building block of intelligent transportation,enabling wide-ranging monitoring,precise positioning,and rapid transmission.However,the high computational cost of transformer-based methods in object detection tasks hinders real-time result transmission in drone target detection applications.Therefore,we propose mask adaptive transformer (MAT) tailored for such scenarios.Specifically,we introduce a structure that supports collaborative token sparsification in support windows,enhancing fault tolerance and reducing computational overhead.This structure comprises two modules:a binary mask strategy and adaptive window self-attention (A-WSA).The binary mask strategy focuses on significant objects in various complex scenes.The A-WSA mechanism is employed to self-attend for balance perfomance and computational cost to select objects and isolate all contextual leakage.Extensive experiments on the challenging CarPK and VisDrone datasets demonstrate the effectiveness and superiority of the proposed method.Specifically,it achieves a mean average precision (mAP@0.5) improvement of 1.25%over car detector based on you only look once version 5 (CD-YOLOv5) on the CarPK dataset and a 3.75%average precision(AP@0.5) improvement over cascaded zoom-in detector (CZ Det) on the VisDrone dataset.
基金supported by the National Key Research and Development Project(No.2022YFC2903302)the Second Tibet Plateau Scientific Expedition and Research Program(STEP),(No.2019QZKK0802)+2 种基金the National Natural Science Foundation of China(No.42361144841)the Chinese Academy of Geological Sciences Basal Research Fund(No.JKYZD202402)the Scientific Research Fund Project of BGRIMM Technology Group(No.JTKY202427822)。
文摘0 INTRODUCTION Orogenic belts are commonly built by multiple-stage processes involving oceanic subduction and continental collisions that result in the generation of magma with distinct geochemical compositions,as exemplified by Central Asian Orogenic Belts(e.g.,Wang et al.,2024;Yin et al.,2024;Xiao et al.,2005)and the Tethyan tectonic domains(e.g.,Chen et al.,2024;Li et al.,2024;Tao et al.,2024a;Gehrels et al.,2011;Yin and Harrison,2000).
文摘The soil packing,influenced by variations in grain size and the gradation pattern within the soil matrix,plays a crucial role in constituting the mechanical properties of sandy soils.However,previous modeling approaches have overlooked incorporating the full range of representative parameters to accurately predict the soaked California bearing ratio(CBR_(s))of sandy soils by precisely articulating soil packing in the modeling framework.This study presents an innovative artificial intelligence(AI)-based approach for modeling the CBR_(s)of sandy soils,considering grain size variability meticulously.By synthesizing extensive data from multiple sources,i.e.extensive tailored testing program undertaking multiple tests and extant literature,various modeling techniques including genetic expression programming(GEP),multi-expression programming(MEP),support vector machine(SVM),and multi-linear regression(MLR)are utilized to develop models.The research explores two modeling strategies,namely simplified and composite,with the former incorporating only sieve analysis test parameters,while the latter includes compaction test parameters alongside sieve analysis data.The models'performance is assessed using statistical key performance indicators(KPIs).Results indicate that genetic AI-based algorithms,particularly GEP,outperform SVM and conventional regression techniques,effectively capturing complex relationships between input parameters and CBR_(s).Additionally,the study reveals insights into model performance concerning the number of input parameters,with GEP consistently outperforming other models.External validation and Taylor diagram analysis demonstrate the GEP models'superiority over existing literature models on an independent dataset from the literature.Parametric and sensitivity analyses highlight the intricate relationships between grain sizes and CBR_(s),further emphasizing GEP's efficacy in modeling such complexities.This study contributes to enhancing CBR_(s)modeling accuracy for sandy soils,crucial for pertinent infrastructure design and construction rapidly and cost-effectively.
基金supported by the National Natural Science Foundation of China (No.52075530)the AiBle Project Co-financed by the European Regional Development Fund+3 种基金Liaoning Province Higher Education Innovative Talents Program Support Project (No.LR2019058)the Scientific Research Project of Liaoning Education Department (No.LG201909)the Liaoning Province Joint Open Fund for Key Scientific and Technological Innovation Bases (No.2021-KF-12-05)the Zhejiang Provincial Natural Science Foundation of China (No.LQ23F030001)。
文摘This paper proposes a lightweight reinforcement network (LRN) and auxiliary label distribution learning (ALDL)based robust facial expression recognition (FER) method.Our designed representation reinforcement (RR) network mainly comprises two modules,i.e.,the RR module and the auxiliary label space construction (ALSC) module.The RR module highlights key feature messaging nodes in feature maps,and ALSC allows multiple labels with different intensities to be linked to one expression.Therefore,LRN has a more robust feature extraction capability when model parameters are greatly reduced,and ALDL is proposed to contribute to the training effect of LRN in the condition of ambiguous training data.We tested our method on FER-Plus and RAF-DB datasets,and the experiment demonstrates the feasibility of our method in practice during rehabilitation robots.
基金supported by the Zhejiang Provincial Natural Science Foundation of China(No.LQ23F030001)Hangzhou Innovation Team(No.TD2022011)+1 种基金the Ai Ble Project co-financed by the European Regional Development Fundthe National Natural Science Foundation of China(No.52075530)。
文摘The integration of endoscopy has significantly propelled the diagnosis and treatment of gastrointestinal diseases,with colonoscopy establishing itself as the primary method for early diagnosis and preventive care in colorectal cancer(CRC).Although deep learning holds promise in mitigating missed polyp rates,modern endoscopy examinations pose additional challenges,such as image blurring and atomizing.This study explores lightweight yet powerful attention mechanisms,introducing the spatial-channel transformer(SCT),an innovative approach that leverages spatial channel relationships for attention weight calculation.The method utilizes rotation operations for inter-dimensional dependencies,followed by residual transformation,encoding inter-channel and spatial information with minimal computational overhead.Extensive experiments on the CVC-Clinic DB polyp detection dataset,addressing endoscopy pitfalls,underscore the superiority of our SCT over other state-of-the-art methods.The proposed model maintains high performance,even in challenging scenarios.
基金supported by the National Natural Science Foundation of China(NSFC,Grant No.12525301)supported by the Science and Technology Facilities Council(STFC)under Grant ST/W001225/1+6 种基金supported by JSPS KAKENHI grant Nos.JP22H00130 and JP20H05855further acknowledges support form the National Key R&D Program of China No.(2022YFF0503404,2023YFA1607800,2023YFA1607803)the National Natural Science Foundation of China(NSFC,Grant Nos.12273048 and 12422301)the CAS Project for Young Scientists in Basic Research(No.YSBR-092)support from the CAS Project for Young Scientists in Basic Research(No.YSBR092)the China Manned Space Projectthe New Cornerstone Science Foundation through the XPLORER Prize.
文摘In this work,we investigate a joint fitting approach based on theoretical models of power spectra associated with density-field reconstruction.Specifically,we consider the matter auto-power spectra before and after baryon acoustic oscillations(BAO)reconstruction,as well as the cross-power spectrum between the pre-and post-reconstructed density fields.We present redshift-space models for these three power spectra at the one-loop level within the framework of standard perturbation theory,and perform a joint analysis using three types of power spectra,and quantify their impact on parameter constraints.When restricting the analysis to wavenumbers k≤0.2 h Mpc^(−1)and adopting a smoothing scale of R_(s)=15 h^(−1)Mpc,we find that incorporating all three power spectra improves parameter constraints by approximately 11%–16%compared to using only the post-reconstruction power spectrum,with the Figure of Merit increasing by 10.5%.These results highlight the advantages of leveraging multiple power spectra in BAO reconstruction,ultimately enabling more precise cosmological parameter estimation.
基金The financial support of the National Natural Science Foundation of China(Grant No.42030714)the National Key R&D Program of China(Grant No.2019YFC1509900)is greatly acknowledged.
文摘High-plastic clays with significant volume change due to moisture variations present critical challenges to civil engineering structures.Limestone calcined clay cement(LC3),an innovative and sustainable hydraulic binder,demonstrates significant potential for improving the engineering characteristics of such soils.Nevertheless,the impact of LC3 on the physico-mechanical characteristics of treated soil under a cyclic wet-dry environment remains unclear.This study for the first time investigates LC3's impact on the long-term durability of treated high-plastic clays through comprehensive macro-micro testing including physical,mechanical,mineralogical,and microstructural investigations with an emphasis on wet-dry cycles.The results revealed that LC3 treatment exhibits significant resistance to wet-dry cycles by completely mitigating the swelling potential,and a considerable reduction in plasticity resulting in enhanced workability.The compressibility and shear strength parameters have been significantly improved to several orders of magnitude.However,after six wet-dry cycles,a slight to modest reduction is observed,but overall durability remains superior to untreated soil.Cohesive and structural bonding ratios quantitatively assessed the impact of wet-dry cycles emphasizing the advantage of LC3 treatment.According to mineralogical and microstructural evaluation,the mechanism behind the adverse effects of wet-dry cycles on the compressibility and strength behavior of LC3-treated soil is mainly attributed to:(1)weakening of CSH/C(A)SH and ettringite(AFt)phases by exhibiting lower peak intensities;and(2)larger pore spaces due to repeated wet-dry cycles.These findings highlight LC3's performance in enhancing the long-term behavior and resilience of treated soils in real-world scenarios,providing durable solutions for infrastructure challenges.
