Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subn...Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subnetworks in order to mitigate the exposure of sensitive data and reduce the overhead on client devices,thereby making SL particularly suitable for resource-constrained devices.Although SL prevents the direct transmission of raw data,it does not alleviate entirely the risk of privacy breaches.In fact,the data intermediately transmitted to the server sub-model may include patterns or information that could reveal sensitive data.Moreover,achieving a balance between model utility and data privacy has emerged as a challenging problem.In this article,we propose a novel defense approach that combines:(i)Adversarial learning,and(ii)Network channel pruning.In particular,the proposed adversarial learning approach is specifically designed to reduce the risk of private data exposure while maintaining high performance for the utility task.On the other hand,the suggested channel pruning enables the model to adaptively adjust and reactivate pruned channels while conducting adversarial training.The integration of these two techniques reduces the informativeness of the intermediate data transmitted by the client sub-model,thereby enhancing its robustness against attribute inference attacks without adding significant computational overhead,making it wellsuited for IoT devices,mobile platforms,and Internet of Vehicles(IoV)scenarios.The proposed defense approach was evaluated using EfficientNet-B0,a widely adopted compact model,along with three benchmark datasets.The obtained results showcased its superior defense capability against attribute inference attacks compared to existing state-of-the-art methods.This research’s findings demonstrated the effectiveness of the proposed channel pruning-based adversarial training approach in achieving the intended compromise between utility and privacy within SL frameworks.In fact,the classification accuracy attained by the attackers witnessed a drastic decrease of 70%.展开更多
Humans can learn complex and dexterous manipulation tasks by observing videos,imitating and exploring.Multiple endeffectors manipulation of free micron-sized deformable cells is one of the challenging tasks in robotic...Humans can learn complex and dexterous manipulation tasks by observing videos,imitating and exploring.Multiple endeffectors manipulation of free micron-sized deformable cells is one of the challenging tasks in robotic micromanipulation.We propose an imitation-enhanced reinforcement learning method inspired by the human learning process that enables robots to learn cell micromanipulation skills from videos.Firstly,for the microscopic robot micromanipulation videos,a multi-task observation(MTO)network is designed to identify the two end-effectors and the manipulated objects to obtain the spatiotemporal trajectories.The spatiotemporal constraints of the robot's actions are obtained by the task-parameterised hidden Markov model(THMM).To simultaneously address the safety and dexterity of robot micromanipulation,an imitation learning optimisation-based soft actor-critic(ILOSAC)algorithm is proposed in which the robot can perform skill learning by demonstration and exploration.The proposed method is capable of performing complex cell manipulation tasks in a realistic physical environment.Experiments indicated that compared with current methods and manual remote manipulation,the proposed framework achieved a shorter operation time and less deformation of cells,which is expected to facilitate the development of robot skill learning.展开更多
Unmanned aerial vehicles(UAVs)face the challenge of autonomous obstacle avoidance in complex,multi-obstacle environments.Behavior cloning offers a promising approach to rapidly acquire a learning policy from limited e...Unmanned aerial vehicles(UAVs)face the challenge of autonomous obstacle avoidance in complex,multi-obstacle environments.Behavior cloning offers a promising approach to rapidly acquire a learning policy from limited expert demonstrations.However,pure imitation learning inherently suffers from poor exploration and limited generalization,typically necessitating extensive datasets to train competent student policies.We utilize a cross-modal variational autoencoder(CM-VAE)to extract compact features from raw visual inputs and UAV states,which then feed into a policy network.We evaluated our approach in a simulated environment featuring a challenging circular trajectory with eight gate obstacles.The results demonstrate that the policy trained with pure behavior cloning consistently failed.In stark contrast,our DAgger-augmented behavior cloning method successfully traversed all gates without collision.Our findings confirm that DAgger effectively mitigates the shortcomings of behavior cloning,enabling the creation of reliable and sample-efficient navigation policies for UAVs.展开更多
The generation of synthetic trajectories has become essential in various fields for analyzing complex movement patterns.However,the use of real-world trajectory data poses significant privacy risks,such as location re...The generation of synthetic trajectories has become essential in various fields for analyzing complex movement patterns.However,the use of real-world trajectory data poses significant privacy risks,such as location reidentification and correlation attacks.To address these challenges,privacy-preserving trajectory generation methods are critical for applications relying on sensitive location data.This paper introduces DPIL-Traj,an advanced framework designed to generate synthetic trajectories while achieving a superior balance between data utility and privacy preservation.Firstly,the framework incorporates Differential Privacy Clustering,which anonymizes trajectory data by applying differential privacy techniques that add noise,ensuring the protection of sensitive user information.Secondly,Imitation Learning is used to replicate decision-making behaviors observed in real-world trajectories.By learning from expert trajectories,this component generates synthetic data that closely mimics real-world decision-making processes while optimizing the quality of the generated trajectories.Finally,Markov-based Trajectory Generation is employed to capture and maintain the inherent temporal dynamics of movement patterns.Extensive experiments conducted on the GeoLife trajectory dataset show that DPIL-Traj improves utility performance by an average of 19.85%,and in terms of privacy performance by an average of 12.51%,compared to state-of-the-art approaches.Ablation studies further reveal that DP clustering effectively safeguards privacy,imitation learning enhances utility under noise,and the Markov module strengthens temporal coherence.展开更多
Scene recognition is a critical component of computer vision,powering applications from autonomous vehicles to surveillance systems.However,its development is often constrained by a heavy reliance on large,expensively...Scene recognition is a critical component of computer vision,powering applications from autonomous vehicles to surveillance systems.However,its development is often constrained by a heavy reliance on large,expensively annotated datasets.This research presents a novel,efficient approach that leveragesmulti-model transfer learning from pre-trained deep neural networks—specifically DenseNet201 and Visual Geometry Group(VGG)—to overcome this limitation.Ourmethod significantly reduces dependency on vast labeled data while achieving high accuracy.Evaluated on the Aerial Image Dataset(AID)dataset,the model attained a validation accuracy of 93.6%with a loss of 0.35,demonstrating robust performance with minimal training data.These results underscore the viability of our approach for real-time,data-efficient scene recognition,offering a practical and cost-effective advancement for the field.展开更多
Waves are important physical phenomena in an ocean,and their accurate prediction is essential for ocean engineering,maritime traffic,and marine early warning systems.This study focuses on the Qinhuangdao Sea area loca...Waves are important physical phenomena in an ocean,and their accurate prediction is essential for ocean engineering,maritime traffic,and marine early warning systems.This study focuses on the Qinhuangdao Sea area located in the Bohai Sea,China.Herein,we use on-site wind data to correct the reanalysis wind data obtained from the European Centre for Medium-Range Weather Forecasts(ECMWF),improving the accuracy of boundary conditions.Then,we use the Simulating WAves Nearshore(SWAN)model to simulate the regional wave field over time.A regional wave-parameter prediction model is then developed using a limited number of sampled data(covering only 2 years,2020–2021);the model is based on the Whale Optimization Algorithm(WOA),convolutional neural networks(CNNs),and long short-term memory(LSTM)neural networks.WOA is used to optimize the CNN and LSTM framework;in this framework,CNN extracts spatial features,and the LSTM network captures temporal features,enabling accurate short and long-term predictions of wave height,period,and direction.The experimental results showed that despite the small sample size,the model achieves a goodness of fit of 0.9957 for wave height prediction,0.9973 for period,and 0.9749 for wave direction in short-term forecasting.As the prediction step size increases,the accuracy of the model decreases.When the prediction step size reaches 9 h,the root mean square error for the prediction of wave height,period,and direction increases to 0.2060 m,0.4582 s,and32.5358°,respectively.The reliability and applicability of the model are further validated by the experimental results.Our findings highlighted the potential of the developed model in operational wave forecasting,even with a limited number of sampled data.展开更多
The rapid proliferation of commercial unmanned aerial vehicles(UAVs)has revolutionized fields such as precision agriculture and disaster response.However,their heavy reliance on GPS navigation leaves them highly vulne...The rapid proliferation of commercial unmanned aerial vehicles(UAVs)has revolutionized fields such as precision agriculture and disaster response.However,their heavy reliance on GPS navigation leaves them highly vulnerable to spoofing attacks,with potentially severe consequences.To mitigate this threat,we present a machine learning-driven framework for real-time GPS spoofing detection,designed with a balance of detection accuracy and computational efficiency.Our work is distinguished by the creation of a comprehensive dataset of 10,000 instances that integrates both simulated and real-world data,enabling robust and generalizable model development.A comprehensive evaluation ofmultiple classification algorithms identifies XGBoost as the superior performer,achieving 93.07% accuracy alongside outstanding precision,recall,and F1-scores.Beyond standard classification metrics,our assessment encompasses ROC-AUC,detection latency,and false positive rate,providing a comprehensive assessment of performance.This work contributes to UAV security by providing a robust and reproducible solution for detecting GPS spoofing attacks,supported by a detailed methodology,a comprehensive evaluation including inference-time latency,and a publicly available dataset.展开更多
Intermittent rivers and ephemeral streams(IRES),also known as non-perennial river segments(NPRs),have garnered attention due to their significant roles in watershed hydrology and ecosystem services,especially in the c...Intermittent rivers and ephemeral streams(IRES),also known as non-perennial river segments(NPRs),have garnered attention due to their significant roles in watershed hydrology and ecosystem services,especially in the context of climate change and escalating human activities.Recent advances in machine learning(ML)techniques have significantly improved the analysis of dynamic changes in IRES.Various ML models,including random forest(RF),long short-term memory(LSTM),and U-Net,demonstrate clear advantages in processing complex hydrological data,enhancing the efficiency and accuracy of IRES extraction from remote sensing data.Furthermore,hybrid ML approaches enhance predictive performance in complex hydrological scenarios by integrating multiple algorithms.However,ML methods still face challenges,including high data dependence,computational complexity,and scalability issues with models.This review proposes an IRES monitoring framework that combines satellite data with ML algorithms,integrating remote sensing technologies such as optical imaging and synthetic aperture radar,and evaluates the advantages and limitations of different ML methods.It further highlights the potential of integrating multiple ML techniques and high-resolution remote sensing data to monitor IRES dynamics,conduct ecological assessments,and support sustainable water management,offering a scientific foundation for addressing environmental and anthropogenic pressures.展开更多
With an increase in internet-connected devices and a dependency on online services,the threat of Distributed Denial of Service(DDoS)attacks has become a significant concern in cybersecurity.The proposed system follows...With an increase in internet-connected devices and a dependency on online services,the threat of Distributed Denial of Service(DDoS)attacks has become a significant concern in cybersecurity.The proposed system follows a multi-step process,beginning with the collection of datasets from different edge devices and network nodes.To verify its effectiveness,experiments were conducted using the CICDoS2017,NSL-KDD,and CICIDS benchmark datasets alongside other existing models.Recursive feature elimination(RFE)with random forest is used to select features from the CICDDoS2019 dataset,on which a BiLSTM model is trained on local nodes.Local models are trained until convergence or stability criteria are met while simultaneously sharing the updates globally for collaborative learning.A centralised server evaluates real-time traffic using the global BiLSTM model,which triggers alerts for potential DDoS attacks.Furthermore,blockchain technology is employed to secure model updates and to provide an immutable audit trail,thereby ensuring trust and accountability among network nodes.This research introduces a novel decentralized method called Federated Random Forest Bidirectional Long Short-Term Memory(FRF-BiLSTM)for detecting DDoS attacks,utilizing the advanced Bidirectional Long Short-Term Memory Networks(BiLSTMs)to analyze sequences in both forward and backward directions.The outcome shows the proposed model achieves a mean accuracy of 97.1%with an average training delay of 88.7 s and testing delay of 21.4 s.The model demonstrates scalability and the best detection performance in large-scale attack scenarios.展开更多
The current deep learning models for braced excavation cannot predict deformation from the beginning of excavation due to the need for a substantial corpus of sufficient historical data for training purposes.To addres...The current deep learning models for braced excavation cannot predict deformation from the beginning of excavation due to the need for a substantial corpus of sufficient historical data for training purposes.To address this issue,this study proposes a transfer learning model based on a sequence-to-sequence twodimensional(2D)convolutional long short-term memory neural network(S2SCL2D).The model can use the existing data from other adjacent similar excavations to achieve wall deflection prediction once a limited amount of monitoring data from the target excavation has been recorded.In the absence of adjacent excavation data,numerical simulation data from the target project can be employed instead.A weight update strategy is proposed to improve the prediction accuracy by integrating the stochastic gradient masking with an early stopping mechanism.To illustrate the proposed methodology,an excavation project in Hangzhou,China is adopted.The proposed deep transfer learning model,which uses either adjacent excavation data or numerical simulation data as the source domain,shows a significant improvement in performance when compared to the non-transfer learning model.Using the simulation data from the target project even leads to better prediction performance than using the actual monitoring data from other adjacent excavations.The results demonstrate that the proposed model can reasonably predict the deformation with limited data from the target project.展开更多
As an effective strategy to address urban traffic congestion,traffic flow prediction has gained attention from Federated-Learning(FL)researchers due FL’s ability to preserving data privacy.However,existing methods fa...As an effective strategy to address urban traffic congestion,traffic flow prediction has gained attention from Federated-Learning(FL)researchers due FL’s ability to preserving data privacy.However,existing methods face challenges:some are too simplistic to capture complex traffic patterns effectively,and others are overly complex,leading to excessive communication overhead between cloud and edge devices.Moreover,the problem of single point failure limits their robustness and reliability in real-world applications.To tackle these challenges,this paper proposes a new method,CMBA-FL,a Communication-Mitigated and Blockchain-Assisted Federated Learning model.First,CMBA-FL improves the client model’s ability to capture temporal traffic patterns by employing the Encoder-Decoder framework for each edge device.Second,to reduce the communication overhead during federated learning,we introduce a verification method based on parameter update consistency,avoiding unnecessary parameter updates.Third,to mitigate the risk of a single point of failure,we integrate consensus mechanisms from blockchain technology.To validate the effectiveness of CMBA-FL,we assess its performance on two widely used traffic datasets.Our experimental results show that CMBA-FL reduces prediction error by 11.46%,significantly lowers communication overhead,and improves security.展开更多
The uplift resistance of the soil overlying shield tunnels significantly impacts their anti-floating stability.However,research on uplift resistance concerning special-shaped shield tunnels is limited.This study combi...The uplift resistance of the soil overlying shield tunnels significantly impacts their anti-floating stability.However,research on uplift resistance concerning special-shaped shield tunnels is limited.This study combines numerical simulation with machine learning techniques to explore this issue.It presents a summary of special-shaped tunnel geometries and introduces a shape coefficient.Through the finite element software,Plaxis3D,the study simulates six key parameters—shape coefficient,burial depth ratio,tunnel’s longest horizontal length,internal friction angle,cohesion,and soil submerged bulk density—that impact uplift resistance across different conditions.Employing XGBoost and ANN methods,the feature importance of each parameter was analyzed based on the numerical simulation results.The findings demonstrate that a tunnel shape more closely resembling a circle leads to reduced uplift resistance in the overlying soil,whereas other parameters exhibit the contrary effects.Furthermore,the study reveals a diminishing trend in the feature importance of buried depth ratio,internal friction angle,tunnel longest horizontal length,cohesion,soil submerged bulk density,and shape coefficient in influencing uplift resistance.展开更多
Interlayer heat accumulation(IHA)is major challenge in the laser powder bed fusion(LPBF)process,as it exacerbates the instability of melt pools,and compromises the quality of the as-built samples.Infrared radiation mo...Interlayer heat accumulation(IHA)is major challenge in the laser powder bed fusion(LPBF)process,as it exacerbates the instability of melt pools,and compromises the quality of the as-built samples.Infrared radiation monitoring is an effective method for exploring IHA.Based on the defined sequence features of interlayer infrared radiation intensity(IIRI),this study established a gated recurrent unit(GRU)neural network model for predicting IIRI in formed samples using machine learning to mitigate the IHA.The model trained on 316 L alloys achieved precise prediction results when transferred to the DZ125 superalloy,effectively managing various emergencies in the LPBF process.The truncated pyramid components were fabricated through parameter optimization based on IIRI prediction results.Compared with the non-optimized components,the CT results demonstrated a significant reduction in internal voids,with the relative density increasing from 91.6% to 98.5%.Additionally,surface roughness(Ra)decreased from 32.58μm to 19.91μm,while residual stress on the top surface was reduced from 169.21 MPa to 102.37 MPa.展开更多
Robots are key to expanding the scope of space applications.The end-to-end training for robot vision-based detection and precision operations is challenging owing to constraints such as extreme environments and high c...Robots are key to expanding the scope of space applications.The end-to-end training for robot vision-based detection and precision operations is challenging owing to constraints such as extreme environments and high computational overhead.This study proposes a lightweight integrated framework for grasp detection and imitation learning,named GD-IL;it comprises a grasp detection algorithm based on manipulability and Gaussian mixture model(manipulability-GMM),and a grasp trajectory generation algorithm based on a two-stage robot imitation learning algorithm(TS-RIL).In the manipulability-GMM algorithm,we apply GMM clustering and ellipse regression to the object point cloud,propose two judgment criteria to generate multiple candidate grasp bounding boxes for the robot,and use manipulability as a metric for selecting the optimal grasp bounding box.The stages of the TS-RIL algorithm are grasp trajectory learning and robot pose optimization.In the first stage,the robot grasp trajectory is characterized using a second-order dynamic movement primitive model and Gaussian mixture regression(GMM).By adjusting the function form of the forcing term,the robot closely approximates the target-grasping trajectory.In the second stage,a robot pose optimization model is built based on the derived pose error formula and manipulability metric.This model allows the robot to adjust its configuration in real time while grasping,thereby effectively avoiding singularities.Finally,an algorithm verification platform is developed based on a Robot Operating System and a series of comparative experiments are conducted in real-world scenarios.The experimental results demonstrate that GD-IL significantly improves the effectiveness and robustness of grasp detection and trajectory imitation learning,outperforming existing state-of-the-art methods in execution efficiency,manipulability,and success rate.展开更多
Hydrogen energy is a crucial support for China’s low-carbon energy transition.With the large-scale integration of renewable energy,the combination of hydrogen and integrated energy systems has become one of the most ...Hydrogen energy is a crucial support for China’s low-carbon energy transition.With the large-scale integration of renewable energy,the combination of hydrogen and integrated energy systems has become one of the most promising directions of development.This paper proposes an optimized schedulingmodel for a hydrogen-coupled electro-heat-gas integrated energy system(HCEHG-IES)using generative adversarial imitation learning(GAIL).The model aims to enhance renewable-energy absorption,reduce carbon emissions,and improve grid-regulation flexibility.First,the optimal scheduling problem of HCEHG-IES under uncertainty is modeled as a Markov decision process(MDP).To overcome the limitations of conventional deep reinforcement learning algorithms—including long optimization time,slow convergence,and subjective reward design—this study augments the PPO algorithm by incorporating a discriminator network and expert data.The newly developed algorithm,termed GAIL,enables the agent to perform imitation learning from expert data.Based on this model,dynamic scheduling decisions are made in continuous state and action spaces,generating optimal energy-allocation and management schemes.Simulation results indicate that,compared with traditional reinforcement-learning algorithms,the proposed algorithmoffers better economic performance.Guided by expert data,the agent avoids blind optimization,shortens the offline training time,and improves convergence performance.In the online phase,the algorithm enables flexible energy utilization,thereby promoting renewable-energy absorption and reducing carbon emissions.展开更多
Countries around the world have been making efforts to reduce pollutant emissions. However, the response of global black carbon(BC) aging to emission changes remains unclear. Using the Community Atmosphere Model versi...Countries around the world have been making efforts to reduce pollutant emissions. However, the response of global black carbon(BC) aging to emission changes remains unclear. Using the Community Atmosphere Model version 6 with a machine-learning-integrated four-mode version of the Modal Aerosol Module, we quantify global BC aging responses to emission reductions for 2011–2018 and for 2050 and 2100 under carbon neutrality. During 2011–18, global trends in BC aging degree(mass ratio of coatings to BC, R_(BC)) exhibited marked regional disparities, with a significant increase in China(5.4% yr^(-1)), which contrasts with minimal changes in the USA, Europe, and India. The divergence is attributed to opposing trends in secondary organic aerosol(SOA) and sulfate coatings, driven by regional changes in the emission ratios of corresponding coating precursors to BC(volatile organic compounds-VOCs/BC and SO_(2)/BC). Projections under carbon neutrality reveal that R_(BC) will increase globally by 47%(118%) in 2050(2100), with strong convergent increases expected across major source regions. The R_(BC) increase, primarily driven by enhanced SOA coatings due to sharper BC reductions relative to VOCs, will enhance the global BC mass absorption cross-section(MAC) by 11%(17%) in 2050(2100).Consequently, although the global BC burden will decline sharply by 60%(76%), the enhanced MAC partially offsets the magnitude of the decline in the BC direct radiative effect, resulting in the moderation of global BC DRE decreases to 88%(92%) of the BC burden reductions in 2050(2100). This study highlights the globally enhanced BC aging and light absorption capacity under carbon neutrality, thereby partly offsetting the impact of BC direct emission reductions on future changes in BC radiative effects globally.展开更多
Autonomous driving technology is constantly developing to a higher level of complex scenes,and there is a growing demand for the utilization of end-to-end data-driven control.However,the end-to-end path tracking proce...Autonomous driving technology is constantly developing to a higher level of complex scenes,and there is a growing demand for the utilization of end-to-end data-driven control.However,the end-to-end path tracking process often encounters challenges in learning efficiency and generalization.To address this issue,this paper designs a deep deterministic policy gradient(DDPG)-based reinforcement learning strategy that integrates imitation learning and feedforward exploration in the path following process.In imitation learning,the path tracking control data generated by the model predictive control(MPC)method is used to train an end-to-end steering control model of a deep neural network.Another feedforward exploration behavior is predicted by road curvature and vehicle speed,and adds it and imitation learning to the DDPG reinforcement learning to obtain decision-making experience and action prediction behavior of the path tracking process.In the reinforcement learning process,imitation learning is used to update the pre-training parameters of the actor network,and a feedforward steering technique with random noise is adopted for strategy exploration.In the reward function,a hierarchical progressive reward form and a constrained objective reward function referring to MPC are designed,and the actor-critic network architecture is determined.Finally,the path tracking performance of the designed method is verified by comparing various training results,simulations,and HIL tests.The results show that the designed method can effectively utilize pre-training and feedforward prior experience to obtain optimal path tracking performance of an autonomous vehicle,and has better generalization ability than other methods.This study provides an efficient control scheme for improving the end-to-end control performance of autonomous vehicles.展开更多
BACKGROUND:Rapid and accurate identification of high-risk patients in the emergency departments(EDs)is crucial for optimizing resource allocation and improving patient outcomes.This study aimed to develop an early pre...BACKGROUND:Rapid and accurate identification of high-risk patients in the emergency departments(EDs)is crucial for optimizing resource allocation and improving patient outcomes.This study aimed to develop an early prediction model for identifying high-risk patients in EDs using initial vital sign measurements.METHODS:This retrospective cohort study analyzed initial vital signs from the Chinese Emergency Triage,Assessment,and Treatment(CETAT)database,which was collected between January 1^(st),2020,and June 25^(th),2023.The primary outcome was the identification of high-risk patients needing immediate treatment.Various machine learning methods,including a deep-learningbased multilayer perceptron(MLP)classifier were evaluated.Model performance was assessed using the area under the receiver operating characteristic curve(AUC-ROC).AUC-ROC values were reported for three scenarios:a default case,a scenario requiring sensitivity greater than 0.8(Scenario I),and a scenario requiring specificity greater than 0.8(Scenario II).SHAP values were calculated to determine the importance of each predictor within the MLP model.RESULTS:A total of 38,797 patients were analyzed,of whom 18.2%were identified as high-risk.Comparative analysis of the predictive models for high-risk patients showed AUC-ROC values ranging from 0.717 to 0.738,with the MLP model outperforming logistic regression(LR),Gaussian Naive Bayes(GNB),and the National Early Warning Score(NEWS).SHAP value analysis identified coma state,peripheral capillary oxygen saturation(SpO_(2)),and systolic blood pressure as the top three predictive factors in the MLP model,with coma state exerting the most contribution.CONCLUSION:Compared with other methods,the MLP model with initial vital signs demonstrated optimal prediction accuracy,highlighting its potential to enhance clinical decision-making in triage in the EDs.展开更多
基金supported by a grant(No.CRPG-25-2054)under the Cybersecurity Research and Innovation Pioneers Initiative,provided by the National Cybersecurity Authority(NCA)in the Kingdom of Saudi Arabia.
文摘Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subnetworks in order to mitigate the exposure of sensitive data and reduce the overhead on client devices,thereby making SL particularly suitable for resource-constrained devices.Although SL prevents the direct transmission of raw data,it does not alleviate entirely the risk of privacy breaches.In fact,the data intermediately transmitted to the server sub-model may include patterns or information that could reveal sensitive data.Moreover,achieving a balance between model utility and data privacy has emerged as a challenging problem.In this article,we propose a novel defense approach that combines:(i)Adversarial learning,and(ii)Network channel pruning.In particular,the proposed adversarial learning approach is specifically designed to reduce the risk of private data exposure while maintaining high performance for the utility task.On the other hand,the suggested channel pruning enables the model to adaptively adjust and reactivate pruned channels while conducting adversarial training.The integration of these two techniques reduces the informativeness of the intermediate data transmitted by the client sub-model,thereby enhancing its robustness against attribute inference attacks without adding significant computational overhead,making it wellsuited for IoT devices,mobile platforms,and Internet of Vehicles(IoV)scenarios.The proposed defense approach was evaluated using EfficientNet-B0,a widely adopted compact model,along with three benchmark datasets.The obtained results showcased its superior defense capability against attribute inference attacks compared to existing state-of-the-art methods.This research’s findings demonstrated the effectiveness of the proposed channel pruning-based adversarial training approach in achieving the intended compromise between utility and privacy within SL frameworks.In fact,the classification accuracy attained by the attackers witnessed a drastic decrease of 70%.
基金supported in part with the General Programme of the National Natural Science Foundation of China(Grant 62576312)the Key Research and Development Program of Zhejiang Province(Grant 2025C01132)the Shandong Province Key R&D Plan Project(Grant 2022LZGC020).
文摘Humans can learn complex and dexterous manipulation tasks by observing videos,imitating and exploring.Multiple endeffectors manipulation of free micron-sized deformable cells is one of the challenging tasks in robotic micromanipulation.We propose an imitation-enhanced reinforcement learning method inspired by the human learning process that enables robots to learn cell micromanipulation skills from videos.Firstly,for the microscopic robot micromanipulation videos,a multi-task observation(MTO)network is designed to identify the two end-effectors and the manipulated objects to obtain the spatiotemporal trajectories.The spatiotemporal constraints of the robot's actions are obtained by the task-parameterised hidden Markov model(THMM).To simultaneously address the safety and dexterity of robot micromanipulation,an imitation learning optimisation-based soft actor-critic(ILOSAC)algorithm is proposed in which the robot can perform skill learning by demonstration and exploration.The proposed method is capable of performing complex cell manipulation tasks in a realistic physical environment.Experiments indicated that compared with current methods and manual remote manipulation,the proposed framework achieved a shorter operation time and less deformation of cells,which is expected to facilitate the development of robot skill learning.
基金supported by the National Natural Science Foundation of China(No.62576349)。
文摘Unmanned aerial vehicles(UAVs)face the challenge of autonomous obstacle avoidance in complex,multi-obstacle environments.Behavior cloning offers a promising approach to rapidly acquire a learning policy from limited expert demonstrations.However,pure imitation learning inherently suffers from poor exploration and limited generalization,typically necessitating extensive datasets to train competent student policies.We utilize a cross-modal variational autoencoder(CM-VAE)to extract compact features from raw visual inputs and UAV states,which then feed into a policy network.We evaluated our approach in a simulated environment featuring a challenging circular trajectory with eight gate obstacles.The results demonstrate that the policy trained with pure behavior cloning consistently failed.In stark contrast,our DAgger-augmented behavior cloning method successfully traversed all gates without collision.Our findings confirm that DAgger effectively mitigates the shortcomings of behavior cloning,enabling the creation of reliable and sample-efficient navigation policies for UAVs.
基金supported by the Natural Science Foundation of Fujian Province of China(2025J01380)National Natural Science Foundation of China(No.62471139)+3 种基金the Major Health Research Project of Fujian Province(2021ZD01001)Fujian Provincial Units Special Funds for Education and Research(2022639)Fujian University of Technology Research Start-up Fund(GY-S24002)Fujian Research and Training Grants for Young and Middle-aged Leaders in Healthcare(GY-H-24179).
文摘The generation of synthetic trajectories has become essential in various fields for analyzing complex movement patterns.However,the use of real-world trajectory data poses significant privacy risks,such as location reidentification and correlation attacks.To address these challenges,privacy-preserving trajectory generation methods are critical for applications relying on sensitive location data.This paper introduces DPIL-Traj,an advanced framework designed to generate synthetic trajectories while achieving a superior balance between data utility and privacy preservation.Firstly,the framework incorporates Differential Privacy Clustering,which anonymizes trajectory data by applying differential privacy techniques that add noise,ensuring the protection of sensitive user information.Secondly,Imitation Learning is used to replicate decision-making behaviors observed in real-world trajectories.By learning from expert trajectories,this component generates synthetic data that closely mimics real-world decision-making processes while optimizing the quality of the generated trajectories.Finally,Markov-based Trajectory Generation is employed to capture and maintain the inherent temporal dynamics of movement patterns.Extensive experiments conducted on the GeoLife trajectory dataset show that DPIL-Traj improves utility performance by an average of 19.85%,and in terms of privacy performance by an average of 12.51%,compared to state-of-the-art approaches.Ablation studies further reveal that DP clustering effectively safeguards privacy,imitation learning enhances utility under noise,and the Markov module strengthens temporal coherence.
基金funded by the Deanship of Scientific Research and Libraries,Princess Nourah bint Abdulrahman University,through the Program of Research Project Funding After Publication,grant No.(RPFAP-23-1445).
文摘Scene recognition is a critical component of computer vision,powering applications from autonomous vehicles to surveillance systems.However,its development is often constrained by a heavy reliance on large,expensively annotated datasets.This research presents a novel,efficient approach that leveragesmulti-model transfer learning from pre-trained deep neural networks—specifically DenseNet201 and Visual Geometry Group(VGG)—to overcome this limitation.Ourmethod significantly reduces dependency on vast labeled data while achieving high accuracy.Evaluated on the Aerial Image Dataset(AID)dataset,the model attained a validation accuracy of 93.6%with a loss of 0.35,demonstrating robust performance with minimal training data.These results underscore the viability of our approach for real-time,data-efficient scene recognition,offering a practical and cost-effective advancement for the field.
基金supported by the National Natural Science Foundation of China(Nos.52071057,52171247)the Liaoning Youth Elite Talent Program(No.XLYC220309)。
文摘Waves are important physical phenomena in an ocean,and their accurate prediction is essential for ocean engineering,maritime traffic,and marine early warning systems.This study focuses on the Qinhuangdao Sea area located in the Bohai Sea,China.Herein,we use on-site wind data to correct the reanalysis wind data obtained from the European Centre for Medium-Range Weather Forecasts(ECMWF),improving the accuracy of boundary conditions.Then,we use the Simulating WAves Nearshore(SWAN)model to simulate the regional wave field over time.A regional wave-parameter prediction model is then developed using a limited number of sampled data(covering only 2 years,2020–2021);the model is based on the Whale Optimization Algorithm(WOA),convolutional neural networks(CNNs),and long short-term memory(LSTM)neural networks.WOA is used to optimize the CNN and LSTM framework;in this framework,CNN extracts spatial features,and the LSTM network captures temporal features,enabling accurate short and long-term predictions of wave height,period,and direction.The experimental results showed that despite the small sample size,the model achieves a goodness of fit of 0.9957 for wave height prediction,0.9973 for period,and 0.9749 for wave direction in short-term forecasting.As the prediction step size increases,the accuracy of the model decreases.When the prediction step size reaches 9 h,the root mean square error for the prediction of wave height,period,and direction increases to 0.2060 m,0.4582 s,and32.5358°,respectively.The reliability and applicability of the model are further validated by the experimental results.Our findings highlighted the potential of the developed model in operational wave forecasting,even with a limited number of sampled data.
文摘The rapid proliferation of commercial unmanned aerial vehicles(UAVs)has revolutionized fields such as precision agriculture and disaster response.However,their heavy reliance on GPS navigation leaves them highly vulnerable to spoofing attacks,with potentially severe consequences.To mitigate this threat,we present a machine learning-driven framework for real-time GPS spoofing detection,designed with a balance of detection accuracy and computational efficiency.Our work is distinguished by the creation of a comprehensive dataset of 10,000 instances that integrates both simulated and real-world data,enabling robust and generalizable model development.A comprehensive evaluation ofmultiple classification algorithms identifies XGBoost as the superior performer,achieving 93.07% accuracy alongside outstanding precision,recall,and F1-scores.Beyond standard classification metrics,our assessment encompasses ROC-AUC,detection latency,and false positive rate,providing a comprehensive assessment of performance.This work contributes to UAV security by providing a robust and reproducible solution for detecting GPS spoofing attacks,supported by a detailed methodology,a comprehensive evaluation including inference-time latency,and a publicly available dataset.
基金National Natural Science Foundation of China,No.41671026。
文摘Intermittent rivers and ephemeral streams(IRES),also known as non-perennial river segments(NPRs),have garnered attention due to their significant roles in watershed hydrology and ecosystem services,especially in the context of climate change and escalating human activities.Recent advances in machine learning(ML)techniques have significantly improved the analysis of dynamic changes in IRES.Various ML models,including random forest(RF),long short-term memory(LSTM),and U-Net,demonstrate clear advantages in processing complex hydrological data,enhancing the efficiency and accuracy of IRES extraction from remote sensing data.Furthermore,hybrid ML approaches enhance predictive performance in complex hydrological scenarios by integrating multiple algorithms.However,ML methods still face challenges,including high data dependence,computational complexity,and scalability issues with models.This review proposes an IRES monitoring framework that combines satellite data with ML algorithms,integrating remote sensing technologies such as optical imaging and synthetic aperture radar,and evaluates the advantages and limitations of different ML methods.It further highlights the potential of integrating multiple ML techniques and high-resolution remote sensing data to monitor IRES dynamics,conduct ecological assessments,and support sustainable water management,offering a scientific foundation for addressing environmental and anthropogenic pressures.
基金supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea(NRF-2025S1A5A2A01005171)by the BK21 programat Chungbuk National University(2025).
文摘With an increase in internet-connected devices and a dependency on online services,the threat of Distributed Denial of Service(DDoS)attacks has become a significant concern in cybersecurity.The proposed system follows a multi-step process,beginning with the collection of datasets from different edge devices and network nodes.To verify its effectiveness,experiments were conducted using the CICDoS2017,NSL-KDD,and CICIDS benchmark datasets alongside other existing models.Recursive feature elimination(RFE)with random forest is used to select features from the CICDDoS2019 dataset,on which a BiLSTM model is trained on local nodes.Local models are trained until convergence or stability criteria are met while simultaneously sharing the updates globally for collaborative learning.A centralised server evaluates real-time traffic using the global BiLSTM model,which triggers alerts for potential DDoS attacks.Furthermore,blockchain technology is employed to secure model updates and to provide an immutable audit trail,thereby ensuring trust and accountability among network nodes.This research introduces a novel decentralized method called Federated Random Forest Bidirectional Long Short-Term Memory(FRF-BiLSTM)for detecting DDoS attacks,utilizing the advanced Bidirectional Long Short-Term Memory Networks(BiLSTMs)to analyze sequences in both forward and backward directions.The outcome shows the proposed model achieves a mean accuracy of 97.1%with an average training delay of 88.7 s and testing delay of 21.4 s.The model demonstrates scalability and the best detection performance in large-scale attack scenarios.
基金supported by the National Key Research and Development Program of China(Grant No.2023YFC3009400)the National Natural Science Foundation of China(Grant Nos.42307218 and U2239251).
文摘The current deep learning models for braced excavation cannot predict deformation from the beginning of excavation due to the need for a substantial corpus of sufficient historical data for training purposes.To address this issue,this study proposes a transfer learning model based on a sequence-to-sequence twodimensional(2D)convolutional long short-term memory neural network(S2SCL2D).The model can use the existing data from other adjacent similar excavations to achieve wall deflection prediction once a limited amount of monitoring data from the target excavation has been recorded.In the absence of adjacent excavation data,numerical simulation data from the target project can be employed instead.A weight update strategy is proposed to improve the prediction accuracy by integrating the stochastic gradient masking with an early stopping mechanism.To illustrate the proposed methodology,an excavation project in Hangzhou,China is adopted.The proposed deep transfer learning model,which uses either adjacent excavation data or numerical simulation data as the source domain,shows a significant improvement in performance when compared to the non-transfer learning model.Using the simulation data from the target project even leads to better prediction performance than using the actual monitoring data from other adjacent excavations.The results demonstrate that the proposed model can reasonably predict the deformation with limited data from the target project.
基金supported by the National Natural Science Foundation of China under Grant No.U20A20182.
文摘As an effective strategy to address urban traffic congestion,traffic flow prediction has gained attention from Federated-Learning(FL)researchers due FL’s ability to preserving data privacy.However,existing methods face challenges:some are too simplistic to capture complex traffic patterns effectively,and others are overly complex,leading to excessive communication overhead between cloud and edge devices.Moreover,the problem of single point failure limits their robustness and reliability in real-world applications.To tackle these challenges,this paper proposes a new method,CMBA-FL,a Communication-Mitigated and Blockchain-Assisted Federated Learning model.First,CMBA-FL improves the client model’s ability to capture temporal traffic patterns by employing the Encoder-Decoder framework for each edge device.Second,to reduce the communication overhead during federated learning,we introduce a verification method based on parameter update consistency,avoiding unnecessary parameter updates.Third,to mitigate the risk of a single point of failure,we integrate consensus mechanisms from blockchain technology.To validate the effectiveness of CMBA-FL,we assess its performance on two widely used traffic datasets.Our experimental results show that CMBA-FL reduces prediction error by 11.46%,significantly lowers communication overhead,and improves security.
基金Guangzhou Metro Scientific Research Project(No.JT204-100111-23001)Chongqing Municipal Special Project for Technological Innovation and Application Development(No.CSTB2022TIAD-KPX0101)Science and Technology Research and Development Program of China State Railway Group Co.,Ltd.(No.N2023G045)。
文摘The uplift resistance of the soil overlying shield tunnels significantly impacts their anti-floating stability.However,research on uplift resistance concerning special-shaped shield tunnels is limited.This study combines numerical simulation with machine learning techniques to explore this issue.It presents a summary of special-shaped tunnel geometries and introduces a shape coefficient.Through the finite element software,Plaxis3D,the study simulates six key parameters—shape coefficient,burial depth ratio,tunnel’s longest horizontal length,internal friction angle,cohesion,and soil submerged bulk density—that impact uplift resistance across different conditions.Employing XGBoost and ANN methods,the feature importance of each parameter was analyzed based on the numerical simulation results.The findings demonstrate that a tunnel shape more closely resembling a circle leads to reduced uplift resistance in the overlying soil,whereas other parameters exhibit the contrary effects.Furthermore,the study reveals a diminishing trend in the feature importance of buried depth ratio,internal friction angle,tunnel longest horizontal length,cohesion,soil submerged bulk density,and shape coefficient in influencing uplift resistance.
基金supported by National Natural Science Foundation of China(Grant No.52275333)the Key Research&Development Program of China Hubei Province(Grant No.2023BAB089).
文摘Interlayer heat accumulation(IHA)is major challenge in the laser powder bed fusion(LPBF)process,as it exacerbates the instability of melt pools,and compromises the quality of the as-built samples.Infrared radiation monitoring is an effective method for exploring IHA.Based on the defined sequence features of interlayer infrared radiation intensity(IIRI),this study established a gated recurrent unit(GRU)neural network model for predicting IIRI in formed samples using machine learning to mitigate the IHA.The model trained on 316 L alloys achieved precise prediction results when transferred to the DZ125 superalloy,effectively managing various emergencies in the LPBF process.The truncated pyramid components were fabricated through parameter optimization based on IIRI prediction results.Compared with the non-optimized components,the CT results demonstrated a significant reduction in internal voids,with the relative density increasing from 91.6% to 98.5%.Additionally,surface roughness(Ra)decreased from 32.58μm to 19.91μm,while residual stress on the top surface was reduced from 169.21 MPa to 102.37 MPa.
基金Supported by National Natural Science Foundation of China(Grant No.52475280)Shaanxi Provincial Natural Science Basic Research Program(Grant No.2025SYSSYSZD-105).
文摘Robots are key to expanding the scope of space applications.The end-to-end training for robot vision-based detection and precision operations is challenging owing to constraints such as extreme environments and high computational overhead.This study proposes a lightweight integrated framework for grasp detection and imitation learning,named GD-IL;it comprises a grasp detection algorithm based on manipulability and Gaussian mixture model(manipulability-GMM),and a grasp trajectory generation algorithm based on a two-stage robot imitation learning algorithm(TS-RIL).In the manipulability-GMM algorithm,we apply GMM clustering and ellipse regression to the object point cloud,propose two judgment criteria to generate multiple candidate grasp bounding boxes for the robot,and use manipulability as a metric for selecting the optimal grasp bounding box.The stages of the TS-RIL algorithm are grasp trajectory learning and robot pose optimization.In the first stage,the robot grasp trajectory is characterized using a second-order dynamic movement primitive model and Gaussian mixture regression(GMM).By adjusting the function form of the forcing term,the robot closely approximates the target-grasping trajectory.In the second stage,a robot pose optimization model is built based on the derived pose error formula and manipulability metric.This model allows the robot to adjust its configuration in real time while grasping,thereby effectively avoiding singularities.Finally,an algorithm verification platform is developed based on a Robot Operating System and a series of comparative experiments are conducted in real-world scenarios.The experimental results demonstrate that GD-IL significantly improves the effectiveness and robustness of grasp detection and trajectory imitation learning,outperforming existing state-of-the-art methods in execution efficiency,manipulability,and success rate.
基金supported by State Grid Corporation Technology Project(No.522437250003).
文摘Hydrogen energy is a crucial support for China’s low-carbon energy transition.With the large-scale integration of renewable energy,the combination of hydrogen and integrated energy systems has become one of the most promising directions of development.This paper proposes an optimized schedulingmodel for a hydrogen-coupled electro-heat-gas integrated energy system(HCEHG-IES)using generative adversarial imitation learning(GAIL).The model aims to enhance renewable-energy absorption,reduce carbon emissions,and improve grid-regulation flexibility.First,the optimal scheduling problem of HCEHG-IES under uncertainty is modeled as a Markov decision process(MDP).To overcome the limitations of conventional deep reinforcement learning algorithms—including long optimization time,slow convergence,and subjective reward design—this study augments the PPO algorithm by incorporating a discriminator network and expert data.The newly developed algorithm,termed GAIL,enables the agent to perform imitation learning from expert data.Based on this model,dynamic scheduling decisions are made in continuous state and action spaces,generating optimal energy-allocation and management schemes.Simulation results indicate that,compared with traditional reinforcement-learning algorithms,the proposed algorithmoffers better economic performance.Guided by expert data,the agent avoids blind optimization,shortens the offline training time,and improves convergence performance.In the online phase,the algorithm enables flexible energy utilization,thereby promoting renewable-energy absorption and reducing carbon emissions.
基金supported by the National Natural Science Foundation of China (42505149,41925023,U2342223,42105069,and 91744208)the China Postdoctoral Science Foundation (2025M770303)+1 种基金the Fundamental Research Funds for the Central Universities (14380230)the Jiangsu Funding Program for Excellent Postdoctoral Talent,and Jiangsu Collaborative Innovation Center of Climate Change。
文摘Countries around the world have been making efforts to reduce pollutant emissions. However, the response of global black carbon(BC) aging to emission changes remains unclear. Using the Community Atmosphere Model version 6 with a machine-learning-integrated four-mode version of the Modal Aerosol Module, we quantify global BC aging responses to emission reductions for 2011–2018 and for 2050 and 2100 under carbon neutrality. During 2011–18, global trends in BC aging degree(mass ratio of coatings to BC, R_(BC)) exhibited marked regional disparities, with a significant increase in China(5.4% yr^(-1)), which contrasts with minimal changes in the USA, Europe, and India. The divergence is attributed to opposing trends in secondary organic aerosol(SOA) and sulfate coatings, driven by regional changes in the emission ratios of corresponding coating precursors to BC(volatile organic compounds-VOCs/BC and SO_(2)/BC). Projections under carbon neutrality reveal that R_(BC) will increase globally by 47%(118%) in 2050(2100), with strong convergent increases expected across major source regions. The R_(BC) increase, primarily driven by enhanced SOA coatings due to sharper BC reductions relative to VOCs, will enhance the global BC mass absorption cross-section(MAC) by 11%(17%) in 2050(2100).Consequently, although the global BC burden will decline sharply by 60%(76%), the enhanced MAC partially offsets the magnitude of the decline in the BC direct radiative effect, resulting in the moderation of global BC DRE decreases to 88%(92%) of the BC burden reductions in 2050(2100). This study highlights the globally enhanced BC aging and light absorption capacity under carbon neutrality, thereby partly offsetting the impact of BC direct emission reductions on future changes in BC radiative effects globally.
基金Supported by National Natural Science Foundation of China(Grant No.52405104)Jiangxi Provincial Natural Science Foundation(Grant Nos.20242BAB20249 and 20232BAB204041)Science and Technology Project of Department of Transportation of Jiangxi Province(Grant No.2025QN003).
文摘Autonomous driving technology is constantly developing to a higher level of complex scenes,and there is a growing demand for the utilization of end-to-end data-driven control.However,the end-to-end path tracking process often encounters challenges in learning efficiency and generalization.To address this issue,this paper designs a deep deterministic policy gradient(DDPG)-based reinforcement learning strategy that integrates imitation learning and feedforward exploration in the path following process.In imitation learning,the path tracking control data generated by the model predictive control(MPC)method is used to train an end-to-end steering control model of a deep neural network.Another feedforward exploration behavior is predicted by road curvature and vehicle speed,and adds it and imitation learning to the DDPG reinforcement learning to obtain decision-making experience and action prediction behavior of the path tracking process.In the reinforcement learning process,imitation learning is used to update the pre-training parameters of the actor network,and a feedforward steering technique with random noise is adopted for strategy exploration.In the reward function,a hierarchical progressive reward form and a constrained objective reward function referring to MPC are designed,and the actor-critic network architecture is determined.Finally,the path tracking performance of the designed method is verified by comparing various training results,simulations,and HIL tests.The results show that the designed method can effectively utilize pre-training and feedforward prior experience to obtain optimal path tracking performance of an autonomous vehicle,and has better generalization ability than other methods.This study provides an efficient control scheme for improving the end-to-end control performance of autonomous vehicles.
基金Applicable Funding Source University of Science and Technology of China(to YLL)National Natural Science Foundation of China(12126604)(to MPZ)+1 种基金R&D project of Pazhou Lab(Huangpu)(2023K0609)(to MPZ)Anhui Provincial Natural Science(grant number 2208085MH235)(to KJ)。
文摘BACKGROUND:Rapid and accurate identification of high-risk patients in the emergency departments(EDs)is crucial for optimizing resource allocation and improving patient outcomes.This study aimed to develop an early prediction model for identifying high-risk patients in EDs using initial vital sign measurements.METHODS:This retrospective cohort study analyzed initial vital signs from the Chinese Emergency Triage,Assessment,and Treatment(CETAT)database,which was collected between January 1^(st),2020,and June 25^(th),2023.The primary outcome was the identification of high-risk patients needing immediate treatment.Various machine learning methods,including a deep-learningbased multilayer perceptron(MLP)classifier were evaluated.Model performance was assessed using the area under the receiver operating characteristic curve(AUC-ROC).AUC-ROC values were reported for three scenarios:a default case,a scenario requiring sensitivity greater than 0.8(Scenario I),and a scenario requiring specificity greater than 0.8(Scenario II).SHAP values were calculated to determine the importance of each predictor within the MLP model.RESULTS:A total of 38,797 patients were analyzed,of whom 18.2%were identified as high-risk.Comparative analysis of the predictive models for high-risk patients showed AUC-ROC values ranging from 0.717 to 0.738,with the MLP model outperforming logistic regression(LR),Gaussian Naive Bayes(GNB),and the National Early Warning Score(NEWS).SHAP value analysis identified coma state,peripheral capillary oxygen saturation(SpO_(2)),and systolic blood pressure as the top three predictive factors in the MLP model,with coma state exerting the most contribution.CONCLUSION:Compared with other methods,the MLP model with initial vital signs demonstrated optimal prediction accuracy,highlighting its potential to enhance clinical decision-making in triage in the EDs.
