Photoacoustic-computed tomography is a novel imaging technique that combines high absorption contrast and deep tissue penetration capability,enabling comprehensive three-dimensional imaging of biological targets.Howev...Photoacoustic-computed tomography is a novel imaging technique that combines high absorption contrast and deep tissue penetration capability,enabling comprehensive three-dimensional imaging of biological targets.However,the increasing demand for higher resolution and real-time imaging results in significant data volume,limiting data storage,transmission and processing efficiency of system.Therefore,there is an urgent need for an effective method to compress the raw data without compromising image quality.This paper presents a photoacoustic-computed tomography 3D data compression method and system based on Wavelet-Transformer.This method is based on the cooperative compression framework that integrates wavelet hard coding with deep learning-based soft decoding.It combines the multiscale analysis capability of wavelet transforms with the global feature modeling advantage of Transformers,achieving high-quality data compression and reconstruction.Experimental results using k-wave simulation suggest that the proposed compression system has advantages under extreme compression conditions,achieving a raw data compression ratio of up to 1:40.Furthermore,three-dimensional data compression experiment using in vivo mouse demonstrated that the maximum peak signal-to-noise ratio(PSNR)and structural similarity index(SSIM)values of reconstructed images reached 38.60 and 0.9583,effectively overcoming detail loss and artifacts introduced by raw data compression.All the results suggest that the proposed system can significantly reduce storage requirements and hardware cost,enhancing computational efficiency and image quality.These advantages support the development of photoacoustic-computed tomography toward higher efficiency,real-time performance and intelligent functionality.展开更多
Data compression plays a vital role in datamanagement and information theory by reducing redundancy.However,it lacks built-in security features such as secret keys or password-based access control,leaving sensitive da...Data compression plays a vital role in datamanagement and information theory by reducing redundancy.However,it lacks built-in security features such as secret keys or password-based access control,leaving sensitive data vulnerable to unauthorized access and misuse.With the exponential growth of digital data,robust security measures are essential.Data encryption,a widely used approach,ensures data confidentiality by making it unreadable and unalterable through secret key control.Despite their individual benefits,both require significant computational resources.Additionally,performing them separately for the same data increases complexity and processing time.Recognizing the need for integrated approaches that balance compression ratios and security levels,this research proposes an integrated data compression and encryption algorithm,named IDCE,for enhanced security and efficiency.Thealgorithmoperates on 128-bit block sizes and a 256-bit secret key length.It combines Huffman coding for compression and a Tent map for encryption.Additionally,an iterative Arnold cat map further enhances cryptographic confusion properties.Experimental analysis validates the effectiveness of the proposed algorithm,showcasing competitive performance in terms of compression ratio,security,and overall efficiency when compared to prior algorithms in the field.展开更多
The accurate prediction of battery pack capacity in electric vehicles(EVs)is crucial for ensuring safety and optimizing performance.Despite extensive research on predicting cell capacity using laboratory data,predicti...The accurate prediction of battery pack capacity in electric vehicles(EVs)is crucial for ensuring safety and optimizing performance.Despite extensive research on predicting cell capacity using laboratory data,predicting the capacity of onboard battery packs from field data remains challenging due to complex operating conditions and irregular EV usage in real-world settings.Most existing methods rely on extracting health feature parameters from raw data for capacity prediction of onboard battery packs,however,selecting specific parameters often results in a loss of critical information,which reduces prediction accuracy.To this end,this paper introduces a novel framework combining deep learning and data compression techniques to accurately predict battery pack capacity onboard.The proposed data compression method converts monthly EV charging data into feature maps,which preserve essential data characteristics while reducing the volume of raw data.To address missing capacity labels in field data,a capacity labeling method is proposed,which calculates monthly battery capacity by transforming the ampere-hour integration formula and applying linear regression.Subsequently,a deep learning model is proposed to build a capacity prediction model,using feature maps from historical months to predict the battery capacity of future months,thus facilitating accurate forecasts.The proposed framework,evaluated using field data from 20 EVs,achieves a mean absolute error of 0.79 Ah,a mean absolute percentage error of 0.65%,and a root mean square error of 1.02 Ah,highlighting its potential for real-world EV applications.展开更多
Three-dimensional(3D)single molecule localization microscopy(SMLM)plays an important role in biomedical applications,but its data processing is very complicated.Deep learning is a potential tool to solve this problem....Three-dimensional(3D)single molecule localization microscopy(SMLM)plays an important role in biomedical applications,but its data processing is very complicated.Deep learning is a potential tool to solve this problem.As the state of art 3D super-resolution localization algorithm based on deep learning,FD-DeepLoc algorithm reported recently still has a gap with the expected goal of online image processing,even though it has greatly improved the data processing throughput.In this paper,a new algorithm Lite-FD-DeepLoc is developed on the basis of FD-DeepLoc algorithm to meet the online image processing requirements of 3D SMLM.This new algorithm uses the feature compression method to reduce the parameters of the model,and combines it with pipeline programming to accelerate the inference process of the deep learning model.The simulated data processing results show that the image processing speed of Lite-FD-DeepLoc is about twice as fast as that of FD-DeepLoc with a slight decrease in localization accuracy,which can realize real-time processing of 256×256 pixels size images.The results of biological experimental data processing imply that Lite-FD-DeepLoc can successfully analyze the data based on astigmatism and saddle point engineering,and the global resolution of the reconstructed image is equivalent to or even better than FD-DeepLoc algorithm.展开更多
Deep neural networks have achieved excellent classification results on several computer vision benchmarks.This has led to the popularity of machine learning as a service,where trained algorithms are hosted on the clou...Deep neural networks have achieved excellent classification results on several computer vision benchmarks.This has led to the popularity of machine learning as a service,where trained algorithms are hosted on the cloud and inference can be obtained on real-world data.In most applications,it is important to compress the vision data due to the enormous bandwidth and memory requirements.Video codecs exploit spatial and temporal correlations to achieve high compression ratios,but they are computationally expensive.This work computes the motion fields between consecutive frames to facilitate the efficient classification of videos.However,contrary to the normal practice of reconstructing the full-resolution frames through motion compensation,this work proposes to infer the class label from the block-based computed motion fields directly.Motion fields are a richer and more complex representation of motion vectors,where each motion vector carries the magnitude and direction information.This approach has two advantages:the cost of motion compensation and video decoding is avoided,and the dimensions of the input signal are highly reduced.This results in a shallower network for classification.The neural network can be trained using motion vectors in two ways:complex representations and magnitude-direction pairs.The proposed work trains a convolutional neural network on the direction and magnitude tensors of the motion fields.Our experimental results show 20×faster convergence during training,reduced overfitting,and accelerated inference on a hand gesture recognition dataset compared to full-resolution and downsampled frames.We validate the proposed methodology on the HGds dataset,achieving a testing accuracy of 99.21%,on the HMDB51 dataset,achieving 82.54%accuracy,and on the UCF101 dataset,achieving 97.13%accuracy,outperforming state-of-the-art methods in computational efficiency.展开更多
SiC/Al-based composite foams were prepared by a two-step foaming method.The influence of the SiC content and its distribution uniformity on the foaming stability,cell structure,and mechanical properties of the aluminu...SiC/Al-based composite foams were prepared by a two-step foaming method.The influence of the SiC content and its distribution uniformity on the foaming stability,cell structure,and mechanical properties of the aluminum foams was investigated.The macro/micro-features of the aluminum foams were characterized and analyzed.Results demonstrate that an appropriate increase in SiC content and the uniform distribution of SiC can improve the foaming stability,optimize the cell diameter and cell wall thickness,ameliorate the cell distribution,and enhance the hardness and compressive strength of the aluminum foams.However,either insufficient or excessive SiC leads to uneven distribution of SiC particles,which is unfavorable to foaming stability and good cell structure formation.With 6wt%SiC,both the foaming stability and cell structure of the aluminum foam reach the optimal state,resulting in the highest compressive strength and optimal energy absorption capacity.展开更多
Earthquakes are highly destructive spatio-temporal phenomena whose analysis is essential for disaster preparedness and risk mitigation.Modern seismological research produces vast volumes of heterogeneous data from sei...Earthquakes are highly destructive spatio-temporal phenomena whose analysis is essential for disaster preparedness and risk mitigation.Modern seismological research produces vast volumes of heterogeneous data from seismic networks,satellite observations,and geospatial repositories,creating the need for scalable infrastructures capable of integrating and analyzing such data to support intelligent decision-making.Data warehousing technologies provide a robust foundation for this purpose;however,existing earthquake-oriented data warehouses remain limited,often relying on simplified schemas,domain-specific analytics,or cataloguing efforts.This paper presents the design and implementation of a spatio-temporal data warehouse for seismic activity.The framework integrates spatial and temporal dimensions in a unified schema and introduces a novel array-based approach for managing many-to-many relationships between facts and dimensions without intermediate bridge tables.A comparative evaluation against a conventional bridge-table schema demonstrates that the array-based design improves fact-centric query performance,while the bridge-table schema remains advantageous for dimension-centric queries.To reconcile these trade-offs,a hybrid schema is proposed that retains both representations,ensuring balanced efficiency across heterogeneous workloads.The proposed framework demonstrates how spatio-temporal data warehousing can address schema complexity,improve query performance,and support multidimensional visualization.In doing so,it provides a foundation for integrating seismic analysis into broader big data-driven intelligent decision systems for disaster resilience,risk mitigation,and emergency management.展开更多
This work contributes to the theoretical foundation for pricing in data markets and offers practical insights for managing digital data exchanges in the era of big data.We propose a structured pricing model for data e...This work contributes to the theoretical foundation for pricing in data markets and offers practical insights for managing digital data exchanges in the era of big data.We propose a structured pricing model for data exchanges transitioning from quasi-public to marketoriented operations.To address the complex dynamics among data exchanges,suppliers,and consumers,the authors develop a threestage Stackelberg game framework.In this model,the data exchange acts as a leader setting transaction commission rates,suppliers are intermediate leaders determining unit prices,and consumers are followers making purchasing decisions.Two pricing strategies are examined:the Independent Pricing Approach(IPA)and the novel Perfectly Competitive Pricing Approach(PCPA),which accounts for competition among data providers.Using backward induction,the study derives subgame-perfect equilibria and proves the existence and uniqueness of Stackelberg equilibria under both approaches.Extensive numerical simulations are carried out in the model,demonstrating that PCPA enhances data demander utility,encourages supplier competition,increases transaction volume,and improves the overall profitability and sustainability of data exchanges.Social welfare analysis further confirms PCPA’s superiority in promoting efficient and fair data markets.展开更多
Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression...Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression is crucial for deploying deep neural network(DNN)models on resource-constrained embedded devices.展开更多
To achieve the potential performance gain of massive multiple-input multiple-output(MIMO)systems,base stations(BS)require downlink channel state information(CSI)fed back by users to execute beamforming design,especial...To achieve the potential performance gain of massive multiple-input multiple-output(MIMO)systems,base stations(BS)require downlink channel state information(CSI)fed back by users to execute beamforming design,especially in the frequency division duplex(FDD)systems.However,due to the enormous number of antennas in massive MIMO systems,the feedback overhead of downlink CSI acquisition is extremely large.To address this issue,deep learning(DL)techniques have been introduced to de velop high-accuracy feedback strategies under limited backhaul constraints.In this paper,we provide an overview of DL-based CSI compression and feedback approaches in massive MIMO systems.Specifically,we introduce the conventional CSI compression and feedback schemes and the existing problems.Besides,we elaborate on various DL techniques employed in CSI compression from the perspective of network architecture and analyze the advantages of different techniques.We also enumerate the applications of DL-based methods for solving practical challenges in CSI compression and feedback.In addition,we brief the remaining issues in deep CSI compression and indicate potential directions in future wireless networks.展开更多
Ovarian cancer(OC)is one of the leading causes of death related to gynecological cancer,with the main difficulty of its early diagnosis and a heterogeneous nature of tumor biomarkers.Machine learning(ML)has the potent...Ovarian cancer(OC)is one of the leading causes of death related to gynecological cancer,with the main difficulty of its early diagnosis and a heterogeneous nature of tumor biomarkers.Machine learning(ML)has the potential to process complex datasets and support decision-making in OC diagnosis.Nevertheless,traditional ML models tend to be biased,overfitting,noisy,and less generalized.Moreover,their black-box nature reduces interpretability and limits their practical clinical applicability.In this study,we introduce an explainable ensemble learning(EL)model,TreeX-Stack,based on a stacking architecture that employs tree-based learners such as Decision Tree(DT),Random Forest(RF),Gradient Boosting(GB),and Extreme Gradient Boosting(XGBoost)as base learners,and Logistic Regression(LR)as the meta-learner to enhance ovarian cancer(OC)diagnosis.Local Interpretable ModelAgnostic Explanations(LIME)are used to explain individual predictions,making the model outputs more clinically interpretable and applicable.The model is trained on the dataset that includes demographic information,blood test,general chemistry,and tumor markers.Extensive preprocessing includes handling missing data using iterative imputation with Bayesian Ridge and addressing multicollinearity by removing features with correlation coefficients above 0.7.Relevant features are then selected using the Boruta feature selection method.To obtain robust and unbiased performance estimates during hyperparameter tuning,nested cross-validation(CV)with grid search is employed,and all experiments are repeated five times to ensure statistical reliability.TreeX-Stack demonstrates excellent diagnostic performance,achieving an accuracy of 0.9027,a precision of 0.8673,a recall of 0.9391,and an F1-score of 0.9012.Feature-importance analyses using LIME and permutation importance highlight Human Epididymis Protein 4(HE4)as the most significant biomarker for OC.The combination of high predictive performance and interpretability makes TreeX-Stack a reliable tool for clinical decision support in OC diagnosis.展开更多
Accurately assessing the relationship between tree growth and climatic factors is of great importance in dendrochronology.This study evaluated the consistency between alternative climate datasets(including station and...Accurately assessing the relationship between tree growth and climatic factors is of great importance in dendrochronology.This study evaluated the consistency between alternative climate datasets(including station and gridded data)and actual climate data(fixed-point observations near the sampling sites),in northeastern China’s warm temperate zone and analyzed differences in their correlations with tree-ring width index.The results were:(1)Gridded temperature data,as well as precipitation and relative humidity data from the Huailai meteorological station,was more consistent with the actual climate data;in contrast,gridded soil moisture content data showed significant discrepancies.(2)Horizontal distance had a greater impact on the representativeness of actual climate conditions than vertical elevation differences.(3)Differences in consistency between alternative and actual climate data also affected their correlations with tree-ring width indices.In some growing season months,correlation coefficients,both in magnitude and sign,differed significantly from those based on actual data.The selection of different alternative climate datasets can lead to biased results in assessing forest responses to climate change,which is detrimental to the management of forest ecosystems in harsh environments.Therefore,the scientific and rational selection of alternative climate data is essential for dendroecological and climatological research.展开更多
To address the severe challenges of PM_(2.5) and ozone co-control during the"14^(th) Five-Year Plan"period and to enhance the precision and intelligence level of air environment governance,it is imperative t...To address the severe challenges of PM_(2.5) and ozone co-control during the"14^(th) Five-Year Plan"period and to enhance the precision and intelligence level of air environment governance,it is imperative to build an efficient comprehensive management platform for regional air quality.In this paper,the specific practice in Zibo City,Shandong Province is as an example to systematically analyze the top-level design,technical implementation,and innovative application of a comprehensive management platform for regional air quality integrating"perception monitoring,data fusion,research judgment of early warnings,analysis of sources,collaborative dispatching,and evaluation assessment".Through the construction of an"sky-air-ground"integrated three-dimensional monitoring network,the platform integrates multi-source heterogeneous environmental data,and employs big data,cloud computing,artificial intelligence,CALPUFF/CMAQ,and other numerical model technologies to achieve comprehensive perception,precise prediction,intelligent source tracing,and closed-loop management of air pollution.The platform innovatively establishes a full-process closed-loop management mechanism of"data-early warning-disposition-evaluation",and achieves a fundamental transformation from passive response to active anticipation and from experience-based judgment to data driving in environmental supervision.The application results show that this platform significantly improves the scientific decision-making ability and collaborative execution efficiency of air pollution governance in Zibo City,providing a replicable and scalable comprehensive solution for similar industrial cities to achieve the continuous improvement of air quality.展开更多
Rock brittleness is a critical property in geotechnical and energy engineering,as it directly influences the prediction of rock failure and stability assessment.Although numerous methods have been developed to evaluat...Rock brittleness is a critical property in geotechnical and energy engineering,as it directly influences the prediction of rock failure and stability assessment.Although numerous methods have been developed to evaluate brittleness,many fail to comprehensively account for the impacts of microstructural changes,mineralogical characteristics,and stress conditions on energy evolution during failure.This study proposes a novel approach for brittleness evaluation based on the energy evolution throughout the post-peak failure process,integrating two micromechanical mechanisms:crack propagation and frictional sliding.A new brittleness index is defined as the ratio of generated surface energy to released elastic energy,providing a unified framework for assessing both Class I and Class II mechanical behaviors.The brittleness of cyan,white,and gray sandstones was investigated under various confining pressures and moisture conditions using X-ray diffraction(XRD),scanning electron microscopy(SEM),and conventional triaxial compression(CTC)tests.The results demonstrate that brittleness decreases with increasing confining pressure,due to suppressed crack propagation,and increases under saturated conditions,as moisture enhances crack propagation.By establishing connections between mineral composition,microstructural features,and stress-induced responses,the proposed method overcame limitations of previous approaches and offered a more precise tool for evaluating rock brittleness under diverse environmental scenarios.展开更多
tRNA-derived small RNAs(tsRNAs),as a class of regulatory small noncoding RNA,have been implicated in a wide variety of human diseases.Large amounts of tsRNA–disease associations have been identified in recent years f...tRNA-derived small RNAs(tsRNAs),as a class of regulatory small noncoding RNA,have been implicated in a wide variety of human diseases.Large amounts of tsRNA–disease associations have been identified in recent years from accumulating studies.However,repositories for cataloging the detailed information on tsRNA–disease associations are scarce.In this study,we provide a tsRNADisease database by integrating experimentally and computationally supported tsRNA–disease associations from manual curation of literatures and other related resources.tsRNADisease contains 5571 manually curated associations between 4759 tsRNAs and 166 diseases with experimental evidence from 346 studies.In addition,it also contains 5013 predicted associations between 1297 tsRNAs and 111 diseases.tsRNADisease provides a user-friendly interface to browse,retrieve,and download data conveniently.This database can improve our understanding of tsRNA deregulation in diseases and serve as a valuable resource for investigating the mechanism of disease-related tsRNAs.tsRNADisease is freely available at http://www.compgenelab.info/tsRNADisease.展开更多
An in-depth understanding of the behaviours of solid propellants under low-velocity impact loads is crucial for enhancing their safety in applications such as aerospace propulsion.This study investigated the dynamic r...An in-depth understanding of the behaviours of solid propellants under low-velocity impact loads is crucial for enhancing their safety in applications such as aerospace propulsion.This study investigated the dynamic responses of single ammonium perchlorate(AP)/octogen(HMX)particles embedded in a hydroxyl-terminated polybutadiene(HTPB)binder under dynamic compression loading via real-time synchrotron-based X-ray phase contrast imaging and a modified split Hopkinson pressure bar(SHPB)system.The compression of the viscoelastic binder and subsequent dynamic fracturing of the AP/HMX particles were captured.During compression,transverse cracks developed within the AP particles,and their propagation led to particle fracturing,resulting in ductile fracturing.Unlike AP,HMX generated numerous short cracks within the internal and edge regions simultaneously,leading to fragmentation and brittle fracturing.Moreover,particle damage reduced the modulus of the sample,shifting its dynamic stress response from nonlinear elasticity to strain softening and further strain hardening as the binder exhibited plastic deformation.A compression simulation incorporating a real particle microscopic structure was established to study the mechanical response of the interface and particles.The simulation results agreed with the experimental observations.These results indicate that the shear stress at the HTPB-AP interface is greater than that at the HTPB-HMX interface,which is a factor influencing the differences in the mesoscale damage mechanisms of the particles.展开更多
0 INTRODUCTION Earth science is a natural science concerned with the composition,dynamics,spatiotemporal evolution,and formation mechanisms of Earth materials(Chen and Yang,2023).Traditional Earth science research has...0 INTRODUCTION Earth science is a natural science concerned with the composition,dynamics,spatiotemporal evolution,and formation mechanisms of Earth materials(Chen and Yang,2023).Traditional Earth science research has largely been discipline-based,relying on field investigations,data collection,experimental analyses,and data interpretation to study individual components of the Earth system.展开更多
Amid the increasing demand for data sharing,the need for flexible,secure,and auditable access control mechanisms has garnered significant attention in the academic community.However,blockchain-based ciphertextpolicy a...Amid the increasing demand for data sharing,the need for flexible,secure,and auditable access control mechanisms has garnered significant attention in the academic community.However,blockchain-based ciphertextpolicy attribute-based encryption(CP-ABE)schemes still face cumbersome ciphertext re-encryption and insufficient oversight when handling dynamic attribute changes and cross-chain collaboration.To address these issues,we propose a dynamic permission attribute-encryption scheme for multi-chain collaboration.This scheme incorporates a multiauthority architecture for distributed attribute management and integrates an attribute revocation and granting mechanism that eliminates the need for ciphertext re-encryption,effectively reducing both computational and communication overhead.It leverages the InterPlanetary File System(IPFS)for off-chain data storage and constructs a cross-chain regulatory framework—comprising a Hyperledger Fabric business chain and a FISCO BCOS regulatory chain—to record changes in decryption privileges and access behaviors in an auditable manner.Security analysis shows selective indistinguishability under chosen-plaintext attack(sIND-CPA)security under the decisional q-Parallel Bilinear Diffie-Hellman Exponent Assumption(q-PBDHE).In the performance and experimental evaluations,we compared the proposed scheme with several advanced schemes.The results show that,while preserving security,the proposed scheme achieves higher encryption/decryption efficiency and lower storage overhead for ciphertexts and keys.展开更多
With the popularization of new technologies,telephone fraud has become the main means of stealing money and personal identity information.Taking inspiration from the website authentication mechanism,we propose an end-...With the popularization of new technologies,telephone fraud has become the main means of stealing money and personal identity information.Taking inspiration from the website authentication mechanism,we propose an end-to-end datamodem scheme that transmits the caller’s digital certificates through a voice channel for the recipient to verify the caller’s identity.Encoding useful information through voice channels is very difficult without the assistance of telecommunications providers.For example,speech activity detection may quickly classify encoded signals as nonspeech signals and reject input waveforms.To address this issue,we propose a novel modulation method based on linear frequency modulation that encodes 3 bits per symbol by varying its frequency,shape,and phase,alongside a lightweightMobileNetV3-Small-based demodulator for efficient and accurate signal decoding on resource-constrained devices.This method leverages the unique characteristics of linear frequency modulation signals,making them more easily transmitted and decoded in speech channels.To ensure reliable data delivery over unstable voice links,we further introduce a robust framing scheme with delimiter-based synchronization,a sample-level position remedying algorithm,and a feedback-driven retransmission mechanism.We have validated the feasibility and performance of our system through expanded real-world evaluations,demonstrating that it outperforms existing advanced methods in terms of robustness and data transfer rate.This technology establishes the foundational infrastructure for reliable certificate delivery over voice channels,which is crucial for achieving strong caller authentication and preventing telephone fraud at its root cause.展开更多
Artificial Intelligence(AI)in healthcare enables predicting diabetes using data-driven methods instead of the traditional ways of screening the disease,which include hemoglobin A1c(HbA1c),oral glucose tolerance test(O...Artificial Intelligence(AI)in healthcare enables predicting diabetes using data-driven methods instead of the traditional ways of screening the disease,which include hemoglobin A1c(HbA1c),oral glucose tolerance test(OGTT),and fasting plasma glucose(FPG)screening techniques,which are invasive and limited in scale.Machine learning(ML)and deep neural network(DNN)models that use large datasets to learn the complex,nonlinear feature interactions,but the conventional ML algorithms are data sensitive and often show unstable predictive accuracy.Conversely,DNN models are more robust,though the ability to reach a high accuracy rate consistently on heterogeneous datasets is still an open challenge.For predicting diabetes,this work proposed a hybrid DNN approach by integrating a bidirectional long short-term memory(BiLSTM)network with a bidirectional gated recurrent unit(BiGRU).A robust DL model,developed by combining various datasets with weighted coefficients,dense operations in the connection of deep layers,and the output aggregation using batch normalization and dropout functions to avoid overfitting.The goal of this hybrid model is better generalization and consistency among various datasets,which facilitates the effective management and early intervention.The proposed DNN model exhibits an excellent predictive performance as compared to the state-of-the-art and baseline ML and DNN models for diabetes prediction tasks.The robust performance indicates the possible usefulness of DL-based models in the development of disease prediction in healthcare and other areas that demand high-quality analytics.展开更多
基金supported by the National Key R&D Program of China[Grant No.2023YFF0713600]the National Natural Science Foundation of China[Grant No.62275062]+3 种基金Project of Shandong Innovation and Startup Community of High-end Medical Apparatus and Instruments[Grant No.2023-SGTTXM-002 and 2024-SGTTXM-005]the Shandong Province Technology Innovation Guidance Plan(Central Leading Local Science and Technology Development Fund)[Grant No.YDZX2023115]the Taishan Scholar Special Funding Project of Shandong Provincethe Shandong Laboratory of Advanced Biomaterials and Medical Devices in Weihai[Grant No.ZL202402].
文摘Photoacoustic-computed tomography is a novel imaging technique that combines high absorption contrast and deep tissue penetration capability,enabling comprehensive three-dimensional imaging of biological targets.However,the increasing demand for higher resolution and real-time imaging results in significant data volume,limiting data storage,transmission and processing efficiency of system.Therefore,there is an urgent need for an effective method to compress the raw data without compromising image quality.This paper presents a photoacoustic-computed tomography 3D data compression method and system based on Wavelet-Transformer.This method is based on the cooperative compression framework that integrates wavelet hard coding with deep learning-based soft decoding.It combines the multiscale analysis capability of wavelet transforms with the global feature modeling advantage of Transformers,achieving high-quality data compression and reconstruction.Experimental results using k-wave simulation suggest that the proposed compression system has advantages under extreme compression conditions,achieving a raw data compression ratio of up to 1:40.Furthermore,three-dimensional data compression experiment using in vivo mouse demonstrated that the maximum peak signal-to-noise ratio(PSNR)and structural similarity index(SSIM)values of reconstructed images reached 38.60 and 0.9583,effectively overcoming detail loss and artifacts introduced by raw data compression.All the results suggest that the proposed system can significantly reduce storage requirements and hardware cost,enhancing computational efficiency and image quality.These advantages support the development of photoacoustic-computed tomography toward higher efficiency,real-time performance and intelligent functionality.
基金the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support(QU-APC-2025).
文摘Data compression plays a vital role in datamanagement and information theory by reducing redundancy.However,it lacks built-in security features such as secret keys or password-based access control,leaving sensitive data vulnerable to unauthorized access and misuse.With the exponential growth of digital data,robust security measures are essential.Data encryption,a widely used approach,ensures data confidentiality by making it unreadable and unalterable through secret key control.Despite their individual benefits,both require significant computational resources.Additionally,performing them separately for the same data increases complexity and processing time.Recognizing the need for integrated approaches that balance compression ratios and security levels,this research proposes an integrated data compression and encryption algorithm,named IDCE,for enhanced security and efficiency.Thealgorithmoperates on 128-bit block sizes and a 256-bit secret key length.It combines Huffman coding for compression and a Tent map for encryption.Additionally,an iterative Arnold cat map further enhances cryptographic confusion properties.Experimental analysis validates the effectiveness of the proposed algorithm,showcasing competitive performance in terms of compression ratio,security,and overall efficiency when compared to prior algorithms in the field.
基金supported in part by the Science and Technology Department of Sichuan Province(No.2025ZNSFSC0427,No.2024ZDZX0035)the Open Project Fund of Vehicle Measurement,Control and Safety Key Laboratory of Sichuan Province(No.QCCK2024-004)the Industrial and Educational Integration Project of Yibin(No.YB-XHU-20240001)。
文摘The accurate prediction of battery pack capacity in electric vehicles(EVs)is crucial for ensuring safety and optimizing performance.Despite extensive research on predicting cell capacity using laboratory data,predicting the capacity of onboard battery packs from field data remains challenging due to complex operating conditions and irregular EV usage in real-world settings.Most existing methods rely on extracting health feature parameters from raw data for capacity prediction of onboard battery packs,however,selecting specific parameters often results in a loss of critical information,which reduces prediction accuracy.To this end,this paper introduces a novel framework combining deep learning and data compression techniques to accurately predict battery pack capacity onboard.The proposed data compression method converts monthly EV charging data into feature maps,which preserve essential data characteristics while reducing the volume of raw data.To address missing capacity labels in field data,a capacity labeling method is proposed,which calculates monthly battery capacity by transforming the ampere-hour integration formula and applying linear regression.Subsequently,a deep learning model is proposed to build a capacity prediction model,using feature maps from historical months to predict the battery capacity of future months,thus facilitating accurate forecasts.The proposed framework,evaluated using field data from 20 EVs,achieves a mean absolute error of 0.79 Ah,a mean absolute percentage error of 0.65%,and a root mean square error of 1.02 Ah,highlighting its potential for real-world EV applications.
基金supported by the Start-up Fund from Hainan University(No.KYQD(ZR)-20077)。
文摘Three-dimensional(3D)single molecule localization microscopy(SMLM)plays an important role in biomedical applications,but its data processing is very complicated.Deep learning is a potential tool to solve this problem.As the state of art 3D super-resolution localization algorithm based on deep learning,FD-DeepLoc algorithm reported recently still has a gap with the expected goal of online image processing,even though it has greatly improved the data processing throughput.In this paper,a new algorithm Lite-FD-DeepLoc is developed on the basis of FD-DeepLoc algorithm to meet the online image processing requirements of 3D SMLM.This new algorithm uses the feature compression method to reduce the parameters of the model,and combines it with pipeline programming to accelerate the inference process of the deep learning model.The simulated data processing results show that the image processing speed of Lite-FD-DeepLoc is about twice as fast as that of FD-DeepLoc with a slight decrease in localization accuracy,which can realize real-time processing of 256×256 pixels size images.The results of biological experimental data processing imply that Lite-FD-DeepLoc can successfully analyze the data based on astigmatism and saddle point engineering,and the global resolution of the reconstructed image is equivalent to or even better than FD-DeepLoc algorithm.
基金Supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2025R896).
文摘Deep neural networks have achieved excellent classification results on several computer vision benchmarks.This has led to the popularity of machine learning as a service,where trained algorithms are hosted on the cloud and inference can be obtained on real-world data.In most applications,it is important to compress the vision data due to the enormous bandwidth and memory requirements.Video codecs exploit spatial and temporal correlations to achieve high compression ratios,but they are computationally expensive.This work computes the motion fields between consecutive frames to facilitate the efficient classification of videos.However,contrary to the normal practice of reconstructing the full-resolution frames through motion compensation,this work proposes to infer the class label from the block-based computed motion fields directly.Motion fields are a richer and more complex representation of motion vectors,where each motion vector carries the magnitude and direction information.This approach has two advantages:the cost of motion compensation and video decoding is avoided,and the dimensions of the input signal are highly reduced.This results in a shallower network for classification.The neural network can be trained using motion vectors in two ways:complex representations and magnitude-direction pairs.The proposed work trains a convolutional neural network on the direction and magnitude tensors of the motion fields.Our experimental results show 20×faster convergence during training,reduced overfitting,and accelerated inference on a hand gesture recognition dataset compared to full-resolution and downsampled frames.We validate the proposed methodology on the HGds dataset,achieving a testing accuracy of 99.21%,on the HMDB51 dataset,achieving 82.54%accuracy,and on the UCF101 dataset,achieving 97.13%accuracy,outperforming state-of-the-art methods in computational efficiency.
基金Doctoral Startup Fund(20192066,20212028)Laijin Excellent Doctoral Fund(20202021)+1 种基金Scientific and Technological Innovation of Colleges and Universities in Shanxi Province(2020L0342)Fundamental Research Program of Shanxi Province(202303021222178)。
文摘SiC/Al-based composite foams were prepared by a two-step foaming method.The influence of the SiC content and its distribution uniformity on the foaming stability,cell structure,and mechanical properties of the aluminum foams was investigated.The macro/micro-features of the aluminum foams were characterized and analyzed.Results demonstrate that an appropriate increase in SiC content and the uniform distribution of SiC can improve the foaming stability,optimize the cell diameter and cell wall thickness,ameliorate the cell distribution,and enhance the hardness and compressive strength of the aluminum foams.However,either insufficient or excessive SiC leads to uneven distribution of SiC particles,which is unfavorable to foaming stability and good cell structure formation.With 6wt%SiC,both the foaming stability and cell structure of the aluminum foam reach the optimal state,resulting in the highest compressive strength and optimal energy absorption capacity.
文摘Earthquakes are highly destructive spatio-temporal phenomena whose analysis is essential for disaster preparedness and risk mitigation.Modern seismological research produces vast volumes of heterogeneous data from seismic networks,satellite observations,and geospatial repositories,creating the need for scalable infrastructures capable of integrating and analyzing such data to support intelligent decision-making.Data warehousing technologies provide a robust foundation for this purpose;however,existing earthquake-oriented data warehouses remain limited,often relying on simplified schemas,domain-specific analytics,or cataloguing efforts.This paper presents the design and implementation of a spatio-temporal data warehouse for seismic activity.The framework integrates spatial and temporal dimensions in a unified schema and introduces a novel array-based approach for managing many-to-many relationships between facts and dimensions without intermediate bridge tables.A comparative evaluation against a conventional bridge-table schema demonstrates that the array-based design improves fact-centric query performance,while the bridge-table schema remains advantageous for dimension-centric queries.To reconcile these trade-offs,a hybrid schema is proposed that retains both representations,ensuring balanced efficiency across heterogeneous workloads.The proposed framework demonstrates how spatio-temporal data warehousing can address schema complexity,improve query performance,and support multidimensional visualization.In doing so,it provides a foundation for integrating seismic analysis into broader big data-driven intelligent decision systems for disaster resilience,risk mitigation,and emergency management.
基金supported by the National Natural Science Foundation of China[grant numbers 12171158,12371474 and 12571510]Fundamental Research Funds for the Central Universities[grant number 2025ECNU-WLJC006].
文摘This work contributes to the theoretical foundation for pricing in data markets and offers practical insights for managing digital data exchanges in the era of big data.We propose a structured pricing model for data exchanges transitioning from quasi-public to marketoriented operations.To address the complex dynamics among data exchanges,suppliers,and consumers,the authors develop a threestage Stackelberg game framework.In this model,the data exchange acts as a leader setting transaction commission rates,suppliers are intermediate leaders determining unit prices,and consumers are followers making purchasing decisions.Two pricing strategies are examined:the Independent Pricing Approach(IPA)and the novel Perfectly Competitive Pricing Approach(PCPA),which accounts for competition among data providers.Using backward induction,the study derives subgame-perfect equilibria and proves the existence and uniqueness of Stackelberg equilibria under both approaches.Extensive numerical simulations are carried out in the model,demonstrating that PCPA enhances data demander utility,encourages supplier competition,increases transaction volume,and improves the overall profitability and sustainability of data exchanges.Social welfare analysis further confirms PCPA’s superiority in promoting efficient and fair data markets.
基金supported by the Science and Technology Innovation Key R&D Program of Chongqing(CSTB2025TIAD-STX0032)National Key Research and Development Program of China(2024YFF0908200)+1 种基金the Chongqing Technology Innovation and Application Development Special Key Project(CSTB2024TIAD-KPX0018)the Southwest University Graduate Student Research Innovation(SWUB24051)。
文摘Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression is crucial for deploying deep neural network(DNN)models on resource-constrained embedded devices.
基金supported by ZTE Industry-University-Institute Cooperation Funds under Grant No.IA20240319003the NSFC under Grant No.62571112。
文摘To achieve the potential performance gain of massive multiple-input multiple-output(MIMO)systems,base stations(BS)require downlink channel state information(CSI)fed back by users to execute beamforming design,especially in the frequency division duplex(FDD)systems.However,due to the enormous number of antennas in massive MIMO systems,the feedback overhead of downlink CSI acquisition is extremely large.To address this issue,deep learning(DL)techniques have been introduced to de velop high-accuracy feedback strategies under limited backhaul constraints.In this paper,we provide an overview of DL-based CSI compression and feedback approaches in massive MIMO systems.Specifically,we introduce the conventional CSI compression and feedback schemes and the existing problems.Besides,we elaborate on various DL techniques employed in CSI compression from the perspective of network architecture and analyze the advantages of different techniques.We also enumerate the applications of DL-based methods for solving practical challenges in CSI compression and feedback.In addition,we brief the remaining issues in deep CSI compression and indicate potential directions in future wireless networks.
基金supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University(IMSIU)under the grant number IMSIU-DDRSP2601.
文摘Ovarian cancer(OC)is one of the leading causes of death related to gynecological cancer,with the main difficulty of its early diagnosis and a heterogeneous nature of tumor biomarkers.Machine learning(ML)has the potential to process complex datasets and support decision-making in OC diagnosis.Nevertheless,traditional ML models tend to be biased,overfitting,noisy,and less generalized.Moreover,their black-box nature reduces interpretability and limits their practical clinical applicability.In this study,we introduce an explainable ensemble learning(EL)model,TreeX-Stack,based on a stacking architecture that employs tree-based learners such as Decision Tree(DT),Random Forest(RF),Gradient Boosting(GB),and Extreme Gradient Boosting(XGBoost)as base learners,and Logistic Regression(LR)as the meta-learner to enhance ovarian cancer(OC)diagnosis.Local Interpretable ModelAgnostic Explanations(LIME)are used to explain individual predictions,making the model outputs more clinically interpretable and applicable.The model is trained on the dataset that includes demographic information,blood test,general chemistry,and tumor markers.Extensive preprocessing includes handling missing data using iterative imputation with Bayesian Ridge and addressing multicollinearity by removing features with correlation coefficients above 0.7.Relevant features are then selected using the Boruta feature selection method.To obtain robust and unbiased performance estimates during hyperparameter tuning,nested cross-validation(CV)with grid search is employed,and all experiments are repeated five times to ensure statistical reliability.TreeX-Stack demonstrates excellent diagnostic performance,achieving an accuracy of 0.9027,a precision of 0.8673,a recall of 0.9391,and an F1-score of 0.9012.Feature-importance analyses using LIME and permutation importance highlight Human Epididymis Protein 4(HE4)as the most significant biomarker for OC.The combination of high predictive performance and interpretability makes TreeX-Stack a reliable tool for clinical decision support in OC diagnosis.
基金supported by the International Partnership program of the Chinese Academy of Sciences(170GJHZ2023074GC)National Natural Science Foundation of China(42425706 and 42488201)+1 种基金National Key Research and Development Program of China(2024YFF0807902)Beijing Natural Science Foundation(8242041),and China Postdoctoral Science Foundation(2025M770353).
文摘Accurately assessing the relationship between tree growth and climatic factors is of great importance in dendrochronology.This study evaluated the consistency between alternative climate datasets(including station and gridded data)and actual climate data(fixed-point observations near the sampling sites),in northeastern China’s warm temperate zone and analyzed differences in their correlations with tree-ring width index.The results were:(1)Gridded temperature data,as well as precipitation and relative humidity data from the Huailai meteorological station,was more consistent with the actual climate data;in contrast,gridded soil moisture content data showed significant discrepancies.(2)Horizontal distance had a greater impact on the representativeness of actual climate conditions than vertical elevation differences.(3)Differences in consistency between alternative and actual climate data also affected their correlations with tree-ring width indices.In some growing season months,correlation coefficients,both in magnitude and sign,differed significantly from those based on actual data.The selection of different alternative climate datasets can lead to biased results in assessing forest responses to climate change,which is detrimental to the management of forest ecosystems in harsh environments.Therefore,the scientific and rational selection of alternative climate data is essential for dendroecological and climatological research.
文摘To address the severe challenges of PM_(2.5) and ozone co-control during the"14^(th) Five-Year Plan"period and to enhance the precision and intelligence level of air environment governance,it is imperative to build an efficient comprehensive management platform for regional air quality.In this paper,the specific practice in Zibo City,Shandong Province is as an example to systematically analyze the top-level design,technical implementation,and innovative application of a comprehensive management platform for regional air quality integrating"perception monitoring,data fusion,research judgment of early warnings,analysis of sources,collaborative dispatching,and evaluation assessment".Through the construction of an"sky-air-ground"integrated three-dimensional monitoring network,the platform integrates multi-source heterogeneous environmental data,and employs big data,cloud computing,artificial intelligence,CALPUFF/CMAQ,and other numerical model technologies to achieve comprehensive perception,precise prediction,intelligent source tracing,and closed-loop management of air pollution.The platform innovatively establishes a full-process closed-loop management mechanism of"data-early warning-disposition-evaluation",and achieves a fundamental transformation from passive response to active anticipation and from experience-based judgment to data driving in environmental supervision.The application results show that this platform significantly improves the scientific decision-making ability and collaborative execution efficiency of air pollution governance in Zibo City,providing a replicable and scalable comprehensive solution for similar industrial cities to achieve the continuous improvement of air quality.
基金supported by the National Natural Science Foundation of China(Grant No.42277147)Ningbo Public Welfare Research Program(Grant No.2024S081)Ningbo Natural Science Foundation(Grant No.2024J186).
文摘Rock brittleness is a critical property in geotechnical and energy engineering,as it directly influences the prediction of rock failure and stability assessment.Although numerous methods have been developed to evaluate brittleness,many fail to comprehensively account for the impacts of microstructural changes,mineralogical characteristics,and stress conditions on energy evolution during failure.This study proposes a novel approach for brittleness evaluation based on the energy evolution throughout the post-peak failure process,integrating two micromechanical mechanisms:crack propagation and frictional sliding.A new brittleness index is defined as the ratio of generated surface energy to released elastic energy,providing a unified framework for assessing both Class I and Class II mechanical behaviors.The brittleness of cyan,white,and gray sandstones was investigated under various confining pressures and moisture conditions using X-ray diffraction(XRD),scanning electron microscopy(SEM),and conventional triaxial compression(CTC)tests.The results demonstrate that brittleness decreases with increasing confining pressure,due to suppressed crack propagation,and increases under saturated conditions,as moisture enhances crack propagation.By establishing connections between mineral composition,microstructural features,and stress-induced responses,the proposed method overcame limitations of previous approaches and offered a more precise tool for evaluating rock brittleness under diverse environmental scenarios.
基金supported by the National Natural Science Foundation of China(91959106)the Foundation of the Shanghai Municipal Education Commission(24RGZNC02)+4 种基金Shanghai Key Laboratory of Intelligent Information Processing,Fudan University(IIPL-2025-RD3-02)Key University Science Research Project of Anhui Province(2023AH030108)Climbing Peak Training Program for Innovative Technology team of Yijishan Hospital,Wannan Medical College(PF201904)Peak Training Program for Scientific Research of Yijishan Hospital,Wannan Medical College(GF2019G15)the talent project of the First Affiliated Hospital of Wannan Medical College(Yijishan Hospital of Wannan Medical College)(YR202422).
文摘tRNA-derived small RNAs(tsRNAs),as a class of regulatory small noncoding RNA,have been implicated in a wide variety of human diseases.Large amounts of tsRNA–disease associations have been identified in recent years from accumulating studies.However,repositories for cataloging the detailed information on tsRNA–disease associations are scarce.In this study,we provide a tsRNADisease database by integrating experimentally and computationally supported tsRNA–disease associations from manual curation of literatures and other related resources.tsRNADisease contains 5571 manually curated associations between 4759 tsRNAs and 166 diseases with experimental evidence from 346 studies.In addition,it also contains 5013 predicted associations between 1297 tsRNAs and 111 diseases.tsRNADisease provides a user-friendly interface to browse,retrieve,and download data conveniently.This database can improve our understanding of tsRNA deregulation in diseases and serve as a valuable resource for investigating the mechanism of disease-related tsRNAs.tsRNADisease is freely available at http://www.compgenelab.info/tsRNADisease.
基金supported by the National Natural Science Foundation of China(U2341288 and 12302492)。
文摘An in-depth understanding of the behaviours of solid propellants under low-velocity impact loads is crucial for enhancing their safety in applications such as aerospace propulsion.This study investigated the dynamic responses of single ammonium perchlorate(AP)/octogen(HMX)particles embedded in a hydroxyl-terminated polybutadiene(HTPB)binder under dynamic compression loading via real-time synchrotron-based X-ray phase contrast imaging and a modified split Hopkinson pressure bar(SHPB)system.The compression of the viscoelastic binder and subsequent dynamic fracturing of the AP/HMX particles were captured.During compression,transverse cracks developed within the AP particles,and their propagation led to particle fracturing,resulting in ductile fracturing.Unlike AP,HMX generated numerous short cracks within the internal and edge regions simultaneously,leading to fragmentation and brittle fracturing.Moreover,particle damage reduced the modulus of the sample,shifting its dynamic stress response from nonlinear elasticity to strain softening and further strain hardening as the binder exhibited plastic deformation.A compression simulation incorporating a real particle microscopic structure was established to study the mechanical response of the interface and particles.The simulation results agreed with the experimental observations.These results indicate that the shear stress at the HTPB-AP interface is greater than that at the HTPB-HMX interface,which is a factor influencing the differences in the mesoscale damage mechanisms of the particles.
基金supported by National Key R&D Program of China(No.2021YFF0501301)the National Natural Science Foundation of China(No.42172231)。
文摘0 INTRODUCTION Earth science is a natural science concerned with the composition,dynamics,spatiotemporal evolution,and formation mechanisms of Earth materials(Chen and Yang,2023).Traditional Earth science research has largely been discipline-based,relying on field investigations,data collection,experimental analyses,and data interpretation to study individual components of the Earth system.
文摘Amid the increasing demand for data sharing,the need for flexible,secure,and auditable access control mechanisms has garnered significant attention in the academic community.However,blockchain-based ciphertextpolicy attribute-based encryption(CP-ABE)schemes still face cumbersome ciphertext re-encryption and insufficient oversight when handling dynamic attribute changes and cross-chain collaboration.To address these issues,we propose a dynamic permission attribute-encryption scheme for multi-chain collaboration.This scheme incorporates a multiauthority architecture for distributed attribute management and integrates an attribute revocation and granting mechanism that eliminates the need for ciphertext re-encryption,effectively reducing both computational and communication overhead.It leverages the InterPlanetary File System(IPFS)for off-chain data storage and constructs a cross-chain regulatory framework—comprising a Hyperledger Fabric business chain and a FISCO BCOS regulatory chain—to record changes in decryption privileges and access behaviors in an auditable manner.Security analysis shows selective indistinguishability under chosen-plaintext attack(sIND-CPA)security under the decisional q-Parallel Bilinear Diffie-Hellman Exponent Assumption(q-PBDHE).In the performance and experimental evaluations,we compared the proposed scheme with several advanced schemes.The results show that,while preserving security,the proposed scheme achieves higher encryption/decryption efficiency and lower storage overhead for ciphertexts and keys.
文摘With the popularization of new technologies,telephone fraud has become the main means of stealing money and personal identity information.Taking inspiration from the website authentication mechanism,we propose an end-to-end datamodem scheme that transmits the caller’s digital certificates through a voice channel for the recipient to verify the caller’s identity.Encoding useful information through voice channels is very difficult without the assistance of telecommunications providers.For example,speech activity detection may quickly classify encoded signals as nonspeech signals and reject input waveforms.To address this issue,we propose a novel modulation method based on linear frequency modulation that encodes 3 bits per symbol by varying its frequency,shape,and phase,alongside a lightweightMobileNetV3-Small-based demodulator for efficient and accurate signal decoding on resource-constrained devices.This method leverages the unique characteristics of linear frequency modulation signals,making them more easily transmitted and decoded in speech channels.To ensure reliable data delivery over unstable voice links,we further introduce a robust framing scheme with delimiter-based synchronization,a sample-level position remedying algorithm,and a feedback-driven retransmission mechanism.We have validated the feasibility and performance of our system through expanded real-world evaluations,demonstrating that it outperforms existing advanced methods in terms of robustness and data transfer rate.This technology establishes the foundational infrastructure for reliable certificate delivery over voice channels,which is crucial for achieving strong caller authentication and preventing telephone fraud at its root cause.
基金supported by the School of Digital Science,Universiti Brunei Darussalam,Brunei.
文摘Artificial Intelligence(AI)in healthcare enables predicting diabetes using data-driven methods instead of the traditional ways of screening the disease,which include hemoglobin A1c(HbA1c),oral glucose tolerance test(OGTT),and fasting plasma glucose(FPG)screening techniques,which are invasive and limited in scale.Machine learning(ML)and deep neural network(DNN)models that use large datasets to learn the complex,nonlinear feature interactions,but the conventional ML algorithms are data sensitive and often show unstable predictive accuracy.Conversely,DNN models are more robust,though the ability to reach a high accuracy rate consistently on heterogeneous datasets is still an open challenge.For predicting diabetes,this work proposed a hybrid DNN approach by integrating a bidirectional long short-term memory(BiLSTM)network with a bidirectional gated recurrent unit(BiGRU).A robust DL model,developed by combining various datasets with weighted coefficients,dense operations in the connection of deep layers,and the output aggregation using batch normalization and dropout functions to avoid overfitting.The goal of this hybrid model is better generalization and consistency among various datasets,which facilitates the effective management and early intervention.The proposed DNN model exhibits an excellent predictive performance as compared to the state-of-the-art and baseline ML and DNN models for diabetes prediction tasks.The robust performance indicates the possible usefulness of DL-based models in the development of disease prediction in healthcare and other areas that demand high-quality analytics.
