The ability to accurately predict urban traffic flows is crucial for optimising city operations.Consequently,various methods for forecasting urban traffic have been developed,focusing on analysing historical data to u...The ability to accurately predict urban traffic flows is crucial for optimising city operations.Consequently,various methods for forecasting urban traffic have been developed,focusing on analysing historical data to understand complex mobility patterns.Deep learning techniques,such as graph neural networks(GNNs),are popular for their ability to capture spatio-temporal dependencies.However,these models often become overly complex due to the large number of hyper-parameters involved.In this study,we introduce Dynamic Multi-Graph Spatial-Temporal Graph Neural Ordinary Differential Equation Networks(DMST-GNODE),a framework based on ordinary differential equations(ODEs)that autonomously discovers effective spatial-temporal graph neural network(STGNN)architectures for traffic prediction tasks.The comparative analysis of DMST-GNODE and baseline models indicates that DMST-GNODE model demonstrates superior performance across multiple datasets,consistently achieving the lowest Root Mean Square Error(RMSE)and Mean Absolute Error(MAE)values,alongside the highest accuracy.On the BKK(Bangkok)dataset,it outperformed other models with an RMSE of 3.3165 and an accuracy of 0.9367 for a 20-min interval,maintaining this trend across 40 and 60 min.Similarly,on the PeMS08 dataset,DMST-GNODE achieved the best performance with an RMSE of 19.4863 and an accuracy of 0.9377 at 20 min,demonstrating its effectiveness over longer periods.The Los_Loop dataset results further emphasise this model’s advantage,with an RMSE of 3.3422 and an accuracy of 0.7643 at 20 min,consistently maintaining superiority across all time intervals.These numerical highlights indicate that DMST-GNODE not only outperforms baseline models but also achieves higher accuracy and lower errors across different time intervals and datasets.展开更多
Electrocardiogram (ECG) analysis is critical for detecting arrhythmias, but traditional methods struggle with large-scale Electrocardiogram data and rare arrhythmia events in imbalanced datasets. These methods fail to...Electrocardiogram (ECG) analysis is critical for detecting arrhythmias, but traditional methods struggle with large-scale Electrocardiogram data and rare arrhythmia events in imbalanced datasets. These methods fail to perform multi-perspective learning of temporal signals and Electrocardiogram images, nor can they fully extract the latent information within the data, falling short of the accuracy required by clinicians. Therefore, this paper proposes an innovative hybrid multimodal spatiotemporal neural network to address these challenges. The model employs a multimodal data augmentation framework integrating visual and signal-based features to enhance the classification performance of rare arrhythmias in imbalanced datasets. Additionally, the spatiotemporal fusion module incorporates a spatiotemporal graph convolutional network to jointly model temporal and spatial features, uncovering complex dependencies within the Electrocardiogram data and improving the model’s ability to represent complex patterns. In experiments conducted on the MIT-BIH arrhythmia dataset, the model achieved 99.95% accuracy, 99.80% recall, and a 99.78% F1 score. The model was further validated for generalization using the clinical INCART arrhythmia dataset, and the results demonstrated its effectiveness in terms of both generalization and robustness.展开更多
Discrete fracture network(DFN)commonly existing in natural rock masses plays an important role in geological complexity which can influence rock fracturing behaviour during fluid injection.This paper simulated the hyd...Discrete fracture network(DFN)commonly existing in natural rock masses plays an important role in geological complexity which can influence rock fracturing behaviour during fluid injection.This paper simulated the hydraulic fracturing process in lab-scale coal samples with DFNs and the induced seismic activities by the discrete element method(DEM).The effects of DFNs on hydraulic fracturing,induced seismicity and elastic property changes have been concluded.Denser DFNs can comprehensively decrease the peak injection pressure and injection duration.The proportion of strong seismic events increases first and then decreases with increasing DFN density.In addition,the relative modulus of the rock mass is derived innovatively from breakdown pressure,breakdown fracture length and the related initiation time.Increasing DFN densities among large(35–60 degrees)and small(0–30 degrees)fracture dip angles show opposite evolution trends in relative modulus.The transitional point(dip angle)for the opposite trends is also proportionally affected by the friction angle of the rock mass.The modelling results have much practical meaning to infer the density and geometry of pre-existing fractures and the elastic property of rock mass in the field,simply based on the hydraulic fracturing and induced seismicity monitoring data.展开更多
Wellbore breakout is one of the critical issues in drilling due to the fact that the related problems result in additional costs and impact the drilling scheme severely.However,the majority of such wellbore breakout a...Wellbore breakout is one of the critical issues in drilling due to the fact that the related problems result in additional costs and impact the drilling scheme severely.However,the majority of such wellbore breakout analyses were based on continuum mechanics.In addition to failure in intact rocks,wellbore breakouts can also be initiated along natural discontinuities,e.g.weak planes and fractures.Furthermore,the conventional models in wellbore breakouts with uniform distribution fractures could not reflect the real drilling situation.This paper presents a fully coupled hydro-mechanical model of the SB-X well in the Tarim Basin,China for evaluating wellbore breakouts in heavily fractured rocks under anisotropic stress states using the distinct element method(DEM)and the discrete fracture network(DFN).The developed model was validated against caliper log measurement,and its stability study was carried out by stress and displacement analyses.A parametric study was performed to investigate the effects of the characteristics of fracture distribution(orientation and length)on borehole stability by sensitivity studies.Simulation results demonstrate that the increase of the standard deviation of orientation when the fracture direction aligns parallel or perpendicular to the principal stress direction aggravates borehole instability.Moreover,an elevation in the average fracture length causes the borehole failure to change from the direction of the minimum in-situ horizontal principal stress(i.e.the direction of wellbore breakouts)towards alternative directions,ultimately leading to the whole wellbore failure.These findings provide theoretical insights for predicting wellbore breakouts in heavily fractured rocks.展开更多
The technique of incremental updating,which can better guarantee the real-time situation of navigational map,is the developing orientation of navigational road network updating.The data center of vehicle navigation sy...The technique of incremental updating,which can better guarantee the real-time situation of navigational map,is the developing orientation of navigational road network updating.The data center of vehicle navigation system is in charge of storing incremental data,and the spatio-temporal data model for storing incremental data does affect the efficiency of the response of the data center to the requirements of incremental data from the vehicle terminal.According to the analysis on the shortcomings of several typical spatio-temporal data models used in the data center and based on the base map with overlay model,the reverse map with overlay model (RMOM) was put forward for the data center to make rapid response to incremental data request.RMOM supports the data center to store not only the current complete road network data,but also the overlays of incremental data from the time when each road network changed to the current moment.Moreover,the storage mechanism and index structure of the incremental data were designed,and the implementation algorithm of RMOM was developed.Taking navigational road network in Guangzhou City as an example,the simulation test was conducted to validate the efficiency of RMOM.Results show that the navigation database in the data center can response to the requirements of incremental data by only one query with RMOM,and costs less time.Compared with the base map with overlay model,the data center does not need to temporarily overlay incremental data with RMOM,so time-consuming of response is significantly reduced.RMOM greatly improves the efficiency of response and provides strong support for the real-time situation of navigational road network.展开更多
Spatio-temporal cellular network traffic prediction at wide-area level plays an important role in resource reconfiguration,traffic scheduling and intrusion detection,thus potentially supporting connected intelligence ...Spatio-temporal cellular network traffic prediction at wide-area level plays an important role in resource reconfiguration,traffic scheduling and intrusion detection,thus potentially supporting connected intelligence of the sixth generation of mobile communications technology(6G).However,the existing studies just focus on the spatio-temporal modeling of traffic data of single network service,such as short message,call,or Internet.It is not conducive to accurate prediction of traffic data,characterised by diverse network service,spatio-temporality and supersize volume.To address this issue,a novel multi-task deep learning framework is developed for citywide cellular network traffic prediction.Functionally,this framework mainly consists of a dual modular feature sharing layer and a multi-task learning layer(DMFS-MT).The former aims at mining long-term spatio-temporal dependencies and local spatio-temporal fluctuation trends in data,respectively,via a new combination of convolutional gated recurrent unit(ConvGRU)and 3-dimensional convolutional neural network(3D-CNN).For the latter,each task is performed for predicting service-specific traffic data based on a fully connected network.On the real-world Telecom Italia dataset,simulation results demonstrate the effectiveness of our proposal through prediction performance measure,spatial pattern comparison and statistical distribution verification.展开更多
Accurate prediction of hydraulic fracture propagation is vital for Enhanced Geothermal System(EGS)design.We study the first hydraulic fracturing job at the GR1 well in the Gonghe Basin using field data,where the overa...Accurate prediction of hydraulic fracture propagation is vital for Enhanced Geothermal System(EGS)design.We study the first hydraulic fracturing job at the GR1 well in the Gonghe Basin using field data,where the overall direction of hydraulic fractures does not show a delineated shape parallel to the maximum principal stress orientation.A field-scale numerical model based on the distinct element method is set up to carry out a fully coupled hydromechanical simulation,with the explicit representation of natural fractures via the discrete fracture network(DFN)approach.The effects of injection parameters and in situ stress on hydraulic fracture patterns are then quantitatively assessed.The study reveals that shear-induced deformation primarily governs the fracturing morphology in the GR1 well,driven by smaller injection rates and viscosities that promote massive activation of natural fractures,ultimately dominating the direction of hydraulic fracturing.Furthermore,the increase of in situ differential stress may promote shear damage of natural fracture surfaces,with the exact influence pattern depending on the combination of specific discontinuity properties and in situ stress state.Finally,we provide recommendations for EGS fracturing based on the influence characteristics of multiple parameters.This study can serve as an effective basis and reference for the design and optimization of EGS in the Gonghe basin and other sites.展开更多
In this paper,we propose a symmetric difference data enhancement physics-informed neural network(SDE-PINN)to study soliton solutions for discrete nonlinear lattice equations(NLEs).By considering known and unknown symm...In this paper,we propose a symmetric difference data enhancement physics-informed neural network(SDE-PINN)to study soliton solutions for discrete nonlinear lattice equations(NLEs).By considering known and unknown symmetric points,numerical simulations are conducted to one-soliton and two-soliton solutions of a discrete KdV equation,as well as a one-soliton solution of a discrete Toda lattice equation.Compared with the existing discrete deep learning approach,the numerical results reveal that within the specified spatiotemporal domain,the prediction accuracy by SDE-PINN is excellent regardless of the interior or extrapolation prediction,with a significant reduction in training time.The proposed data enhancement technique and symmetric structure development provides a new perspective for the deep learning approach to solve discrete NLEs.The newly proposed SDE-PINN can also be applied to solve continuous nonlinear equations and other discrete NLEs numerically.展开更多
The goal of this research is to develop mine-scale discrete fracture network(DFN)models in which the influence of the spatial heterogeneity of fracture distributions may be investigated on the rock wedge stability of ...The goal of this research is to develop mine-scale discrete fracture network(DFN)models in which the influence of the spatial heterogeneity of fracture distributions may be investigated on the rock wedge stability of an open pit slope.For this purpose,spatially conditioned DFN models were developed for the pit walls at Tasiast mine using comprehensive structural data from the mine.Using Sequential Gaussian Simulation(SGS),volumetric fracture intensities(P32)were modeled across the entire mine site in the form of 3D block models.The simulated P32 block models were used as the input constraints for conditional DFN fracture generation,where the DFN grid dimension is the same as the SGS 3D blocks.The spatially constrained DFN models were further calibrated using aerial fracture intensities(P21)data from the pit walls,obtained by a survey of the pit walls using an unmanned aerial vehicle(UAV)and measured traces of joints from 3D point cloud data.The final DFN model is expected to honor the fracture intensities gathered through different means with optimal model accuracy.Finally,bench-scale and interramp scale rock wedge slope stability analyses were conducted using the calibrated conditional DFN models.This work proves the significance of conditioned DFN models in rock wedge stability analysis.Such models provide detailed information regarding rock wedge stability so that site monitoring and prevention plans can be conducted with higher efficiency.展开更多
Appropriately characterising the mixed space-time relations of the contagion process caused by hybrid space and time factors remains the primary challenge in COVID-19 forecasting.However,in previous deep learning mode...Appropriately characterising the mixed space-time relations of the contagion process caused by hybrid space and time factors remains the primary challenge in COVID-19 forecasting.However,in previous deep learning models for epidemic forecasting,spatial and temporal variations are captured separately.A unified model is developed to cover all spatio-temporal relations.However,this measure is insufficient for modelling the complex spatio-temporal relations of infectious disease transmission.A dynamic adaptive spatio-temporal graph network(DASTGN)is proposed based on attention mechanisms to improve prediction accuracy.In DASTGN,complex spatio-temporal relations are depicted by adaptively fusing the mixed space-time effects and dynamic space-time dependency structure.This dual-scale model considers the time-specific,space-specific,and direct effects of the propagation process at the fine-grained level.Furthermore,the model characterises impacts from various space-time neighbour blocks under time-varying interventions at the coarse-grained level.The performance comparisons on the three COVID-19 datasets reveal that DASTGN achieves state-of-the-art results with a maximum improvement of 17.092%in the root mean-square error and 11.563%in the mean absolute error.Experimental results indicate that the mechanisms of designing DASTGN can effectively detect some spreading characteristics of COVID-19.The spatio-temporal weight matrices learned in each proposed module reveal diffusion patterns in various scenarios.In conclusion,DASTGN has successfully captured the dynamic spatio-temporal variations of COVID-19,and considering multiple dynamic space-time relationships is essential in epidemic forecasting.展开更多
Owing to the wide range of applications in various fields,generative models have become increasingly popular.However,they do not handle spatio-temporal features well.Inspired by the recent advances in these models,thi...Owing to the wide range of applications in various fields,generative models have become increasingly popular.However,they do not handle spatio-temporal features well.Inspired by the recent advances in these models,this paper designs a distributed spatio-temporal generative adversarial network(STGAN-D)that,given some initial data and random noise,generates a consecutive sequence of spatio-temporal samples which have a logical relationship.This paper builds a spatio-temporal discriminator to distinguish whether the samples generated by the generator meet the requirements for time and space coherence,and builds a controller for distributed training of the network gradient updated to separate the model training and parameter updating,to improve the network training rate.The model is trained on the skeletal dataset and the traffic dataset.In contrast to traditional generative adversarial networks(GANs),the proposed STGAN-D can generate logically coherent samples with the corresponding spatial and temporal features while avoiding mode collapse.In addition,this paper shows that the proposed model can generate different styles of spatio-temporal samples given different random noise inputs,and the controller can improve the network training rate.This model will extend the potential range of applications of GANs to areas such as traffic information simulation and multiagent adversarial simulation.展开更多
The manner and conditions of running the decision-making system with self-defense electronic jamming are given. After proposing the scenario of applying discrete dynamic Bayesian network to the decision making with se...The manner and conditions of running the decision-making system with self-defense electronic jamming are given. After proposing the scenario of applying discrete dynamic Bayesian network to the decision making with self-defense electronic jamming, a decision-making model with self-defense electronic jamming based on the discrete dynamic Bayesian network is established. Then jamming decision inferences by the aid of the algorithm of discrete dynamic Bayesian network are carried on. The simulating result shows that this method is able to synthesize different targets which are not predominant. In this way, various features at the same time, as well as the same feature appearing at different time complement mutually; in addition, the accuracy and reliability of electronic jamming decision making are enhanced significantly.展开更多
Terrestrial laser scanning(TLS) is a useful technology for rock mass characterization. A laser scanner produces a massive point cloud of a scanned area, such as an exposed rock surface in an underground tunnel,with mi...Terrestrial laser scanning(TLS) is a useful technology for rock mass characterization. A laser scanner produces a massive point cloud of a scanned area, such as an exposed rock surface in an underground tunnel,with millimeter precision. The density of the point cloud depends on several parameters from both the TLS operational conditions and the specifications of the project, such as the resolution and the quality of the laser scan, the section of the tunnel, the distance between scanning stations, and the purpose of the scans. One purpose of the scan can be to characterize the rock mass and statistically analyze the discontinuities that compose it for further discontinuous modeling. In these instances, additional data processing and a detailed analysis should be performed on the point cloud to extract the parameters to define a discrete fracture network(DFN) for each discontinuity set. I-site studio is a point cloud processing software that allows users to edit and process laser scans. This software contains a set of geotechnical analysis tools that assist engineers during the structural mapping process, allowing for greater and more representative data regarding the structural information of the rock mass, which may be used for generating DFNs. This paper presents the procedures used during a laser scan for characterizing discontinuities in an underground limestone mine and the results of the scan as applied to the generation of DFNs for further discontinuous modeling.展开更多
Heterogeneity is an inherent component of rock and may be present in different forms including mineralheterogeneity, geometrical heterogeneity, weak grain boundaries and micro-defects. Microcracks areusually observed ...Heterogeneity is an inherent component of rock and may be present in different forms including mineralheterogeneity, geometrical heterogeneity, weak grain boundaries and micro-defects. Microcracks areusually observed in crystalline rocks in two forms： natural and stress-induced; the amount of stressinducedmicrocracking increases with depth and in-situ stress. Laboratory results indicate that thephysical properties of rocks such as strength, deformability, P-wave velocity and permeability areinfluenced by increase in microcrack intensity. In this study, the finite-discrete element method （FDEM）is used to model microcrack heterogeneity by introducing into a model sample sets of microcracks usingthe proposed micro discrete fracture network （mDFN） approach. The characteristics of the microcracksrequired to create mDFN models are obtained through image analyses of thin sections of Lac du Bonnetgranite adopted from published literature. A suite of two-dimensional laboratory tests including uniaxial,triaxial compression and Brazilian tests is simulated and the results are compared with laboratory data.The FDEM-mDFN models indicate that micro-heterogeneity has a profound influence on both the mechanicalbehavior and resultant fracture pattern. An increase in the microcrack intensity leads to areduction in the strength of the sample and changes the character of the rock strength envelope. Spallingand axial splitting dominate the failure mode at low confinement while shear failure is the dominantfailure mode at high confinement. Numerical results from simulated compression tests show thatmicrocracking reduces the cohesive component of strength alone, and the frictional strength componentremains unaffected. Results from simulated Brazilian tests show that the tensile strength is influenced bythe presence of microcracks, with a reduction in tensile strength as microcrack intensity increases. Theimportance of microcrack heterogeneity in reproducing a bi-linear or S-shape failure envelope and itseffects on the mechanisms leading to spalling damage near an underground opening are also discussed.展开更多
Fractured reservoirs are an important target for oil and gas exploration in the Tarim Basin and the prediction of this type of reservoir is challenging.Due to the complicated fracture system in the Tarim Basin,the con...Fractured reservoirs are an important target for oil and gas exploration in the Tarim Basin and the prediction of this type of reservoir is challenging.Due to the complicated fracture system in the Tarim Basin,the conventional AVO inversion method based on HTI theory to predict fracture development will result in some errors.Thus,an integrated research concept for fractured reservoir prediction is put forward in this paper.Seismic modeling plays a bridging role in this concept,and the establishment of an anisotropic fracture model by Discrete Fracture Network （DFN） is the key part.Because the fracture system in the Tarim Basin shows complex anisotropic characteristics,it is vital to build an effective anisotropic model.Based on geological,well logging and seismic data,an effective anisotropic model of complex fracture systems can be set up with the DFN method.The effective elastic coefficients,and the input data for seismic modeling can be calculated.Then seismic modeling based on this model is performed,and the seismic response characteristics are analyzed.The modeling results can be used in the following AVO inversion for fracture detection.展开更多
This paper concernes analysis for the global exponential stability of a class of recurrent neural networks with mixed discrete and distributed delays. It first proves the existence and uniqueness of the balance point,...This paper concernes analysis for the global exponential stability of a class of recurrent neural networks with mixed discrete and distributed delays. It first proves the existence and uniqueness of the balance point, then by employing the Lyapunov-Krasovskii functional and Young inequality, it gives the sufficient condition of global exponential stability of cellular neural network with mixed discrete and distributed delays, in addition, the example is provided to illustrate the applicability of the result.展开更多
Mapping crop distribution with remote sensing data is of great importance for agricultural production, food security and agricultural sustainability. Winter rape is an important oil crop, which plays an important role...Mapping crop distribution with remote sensing data is of great importance for agricultural production, food security and agricultural sustainability. Winter rape is an important oil crop, which plays an important role in the cooking oil market of China. The Jianghan Plain and Dongting Lake Plain (JPDLP) are major agricultural production areas in China. Essential changes in winter rape distribution have taken place in this area during the 21st century. However, the pattern of these changes remains unknown. In this study, the spatial and temporal dynamics of winter rape from 2000 to 2017 on the JPDLP were analyzed. An artificial neural network (ANN)-based classification method was proposed to map fractional winter rape distribution by fusing moderate resolution imaging spectrometer (MODIS) data and high-resolution imagery. The results are as follows:(1) The total winter rape acreages on the JPDLP dropped significantly, especially on the Jianghan Plain with a decline of about 45% during 2000 and 2017.(2) The winter rape abundance keeps changing with about 20–30% croplands changing their abundance drastically in every two consecutive observation years.(3) The winter rape has obvious regional differentiation for the trend of its change at the county level, and the decreasing trend was observed more strongly in the traditionally dominant agricultural counties.展开更多
Structure plane is one of the important factors affecting the stability and failure mode of rock mass engineering.Rock mass structure characterization is the basic work of rock mechanics research and the important con...Structure plane is one of the important factors affecting the stability and failure mode of rock mass engineering.Rock mass structure characterization is the basic work of rock mechanics research and the important content of numerical simulation.A new 3-dimensional rough discrete fracture network(RDFN3D)model and its modeling method based on the Weierstrass-Mandelbrot(W-M)function were presented in this paper.The RDFN3D model,which improves and unifies the modelling methods for the complex structural planes,has been realized.The influence of fractal dimension,amplitude,and surface precision on the modeling parameters of RDFN3D was discussed.The reasonable W-M parameters suitable for the roughness coefficient of JRC were proposed,and the relationship between the mathematical model and the joint characterization was established.The RDFN3D together with the smooth 3-dimensional discrete fracture network(DFN3D)models were successfully exported to the drawing exchange format,which will provide a wide application in numerous numerical simulation codes including both the continuous and discontinuous methods.The numerical models were discussed using the COMSOL Multiphysics code and the 3-dimensional particle flow code,respectively.The reliability of the RDFN3D model was preliminarily discussed and analyzed.The roughness and spatial connectivity of the fracture networks have a dominant effect on the fluid flow patterns.The research results can provide a new geological model and analysis model for numerical simulation and engineering analysis of jointed rock mass.展开更多
The accuracy of target threat estimation has a great impact on command decision-making.The Bayesian network,as an effective way to deal with the problem of uncertainty,can be used to track the change of the target thr...The accuracy of target threat estimation has a great impact on command decision-making.The Bayesian network,as an effective way to deal with the problem of uncertainty,can be used to track the change of the target threat level.Unfortunately,the traditional discrete dynamic Bayesian network(DDBN)has the problems of poor parameter learning and poor reasoning accuracy in a small sample environment with partial prior information missing.Considering the finiteness and discreteness of DDBN parameters,a fuzzy k-nearest neighbor(KNN)algorithm based on correlation of feature quantities(CF-FKNN)is proposed for DDBN parameter learning.Firstly,the correlation between feature quantities is calculated,and then the KNN algorithm with fuzzy weight is introduced to fill the missing data.On this basis,a reasonable DDBN structure is constructed by using expert experience to complete DDBN parameter learning and reasoning.Simulation results show that the CF-FKNN algorithm can accurately fill in the data when the samples are seriously missing,and improve the effect of DDBN parameter learning in the case of serious sample missing.With the proposed method,the final target threat assessment results are reasonable,which meets the needs of engineering applications.展开更多
Human Activity Recognition(HAR)is an active research area due to its applications in pervasive computing,human-computer interaction,artificial intelligence,health care,and social sciences.Moreover,dynamic environments...Human Activity Recognition(HAR)is an active research area due to its applications in pervasive computing,human-computer interaction,artificial intelligence,health care,and social sciences.Moreover,dynamic environments and anthropometric differences between individuals make it harder to recognize actions.This study focused on human activity in video sequences acquired with an RGB camera because of its vast range of real-world applications.It uses two-stream ConvNet to extract spatial and temporal information and proposes a fine-tuned deep neural network.Moreover,the transfer learning paradigm is adopted to extract varied and fixed frames while reusing object identification information.Six state-of-the-art pre-trained models are exploited to find the best model for spatial feature extraction.For temporal sequence,this study uses dense optical flow following the two-stream ConvNet and Bidirectional Long Short TermMemory(BiLSTM)to capture longtermdependencies.Two state-of-the-art datasets,UCF101 and HMDB51,are used for evaluation purposes.In addition,seven state-of-the-art optimizers are used to fine-tune the proposed network parameters.Furthermore,this study utilizes an ensemble mechanism to aggregate spatial-temporal features using a four-stream Convolutional Neural Network(CNN),where two streams use RGB data.In contrast,the other uses optical flow images.Finally,the proposed ensemble approach using max hard voting outperforms state-ofthe-art methods with 96.30%and 90.07%accuracies on the UCF101 and HMDB51 datasets.展开更多
文摘The ability to accurately predict urban traffic flows is crucial for optimising city operations.Consequently,various methods for forecasting urban traffic have been developed,focusing on analysing historical data to understand complex mobility patterns.Deep learning techniques,such as graph neural networks(GNNs),are popular for their ability to capture spatio-temporal dependencies.However,these models often become overly complex due to the large number of hyper-parameters involved.In this study,we introduce Dynamic Multi-Graph Spatial-Temporal Graph Neural Ordinary Differential Equation Networks(DMST-GNODE),a framework based on ordinary differential equations(ODEs)that autonomously discovers effective spatial-temporal graph neural network(STGNN)architectures for traffic prediction tasks.The comparative analysis of DMST-GNODE and baseline models indicates that DMST-GNODE model demonstrates superior performance across multiple datasets,consistently achieving the lowest Root Mean Square Error(RMSE)and Mean Absolute Error(MAE)values,alongside the highest accuracy.On the BKK(Bangkok)dataset,it outperformed other models with an RMSE of 3.3165 and an accuracy of 0.9367 for a 20-min interval,maintaining this trend across 40 and 60 min.Similarly,on the PeMS08 dataset,DMST-GNODE achieved the best performance with an RMSE of 19.4863 and an accuracy of 0.9377 at 20 min,demonstrating its effectiveness over longer periods.The Los_Loop dataset results further emphasise this model’s advantage,with an RMSE of 3.3422 and an accuracy of 0.7643 at 20 min,consistently maintaining superiority across all time intervals.These numerical highlights indicate that DMST-GNODE not only outperforms baseline models but also achieves higher accuracy and lower errors across different time intervals and datasets.
基金supported by The Henan Province Science and Technology Research Project(242102211046)the Key Scientific Research Project of Higher Education Institutions in Henan Province(25A520039)+1 种基金theNatural Science Foundation project of Zhongyuan Institute of Technology(K2025YB011)the Zhongyuan University of Technology Graduate Education and Teaching Reform Research Project(JG202424).
文摘Electrocardiogram (ECG) analysis is critical for detecting arrhythmias, but traditional methods struggle with large-scale Electrocardiogram data and rare arrhythmia events in imbalanced datasets. These methods fail to perform multi-perspective learning of temporal signals and Electrocardiogram images, nor can they fully extract the latent information within the data, falling short of the accuracy required by clinicians. Therefore, this paper proposes an innovative hybrid multimodal spatiotemporal neural network to address these challenges. The model employs a multimodal data augmentation framework integrating visual and signal-based features to enhance the classification performance of rare arrhythmias in imbalanced datasets. Additionally, the spatiotemporal fusion module incorporates a spatiotemporal graph convolutional network to jointly model temporal and spatial features, uncovering complex dependencies within the Electrocardiogram data and improving the model’s ability to represent complex patterns. In experiments conducted on the MIT-BIH arrhythmia dataset, the model achieved 99.95% accuracy, 99.80% recall, and a 99.78% F1 score. The model was further validated for generalization using the clinical INCART arrhythmia dataset, and the results demonstrated its effectiveness in terms of both generalization and robustness.
基金Australian Research Council Linkage Program(LP200301404)for sponsoring this researchthe financial support provided by the State Key Laboratory of Geohazard Prevention and Geoenvironment Protection(Chengdu University of Technology,SKLGP2021K002)National Natural Science Foundation of China(52374101,32111530138).
文摘Discrete fracture network(DFN)commonly existing in natural rock masses plays an important role in geological complexity which can influence rock fracturing behaviour during fluid injection.This paper simulated the hydraulic fracturing process in lab-scale coal samples with DFNs and the induced seismic activities by the discrete element method(DEM).The effects of DFNs on hydraulic fracturing,induced seismicity and elastic property changes have been concluded.Denser DFNs can comprehensively decrease the peak injection pressure and injection duration.The proportion of strong seismic events increases first and then decreases with increasing DFN density.In addition,the relative modulus of the rock mass is derived innovatively from breakdown pressure,breakdown fracture length and the related initiation time.Increasing DFN densities among large(35–60 degrees)and small(0–30 degrees)fracture dip angles show opposite evolution trends in relative modulus.The transitional point(dip angle)for the opposite trends is also proportionally affected by the friction angle of the rock mass.The modelling results have much practical meaning to infer the density and geometry of pre-existing fractures and the elastic property of rock mass in the field,simply based on the hydraulic fracturing and induced seismicity monitoring data.
基金supported by National Natural Science Foundation of China(Grant Nos.52074312 and 52211530097)CNPC Science and Technology Innovation Foundation(Grant No.2021DQ02-0505).
文摘Wellbore breakout is one of the critical issues in drilling due to the fact that the related problems result in additional costs and impact the drilling scheme severely.However,the majority of such wellbore breakout analyses were based on continuum mechanics.In addition to failure in intact rocks,wellbore breakouts can also be initiated along natural discontinuities,e.g.weak planes and fractures.Furthermore,the conventional models in wellbore breakouts with uniform distribution fractures could not reflect the real drilling situation.This paper presents a fully coupled hydro-mechanical model of the SB-X well in the Tarim Basin,China for evaluating wellbore breakouts in heavily fractured rocks under anisotropic stress states using the distinct element method(DEM)and the discrete fracture network(DFN).The developed model was validated against caliper log measurement,and its stability study was carried out by stress and displacement analyses.A parametric study was performed to investigate the effects of the characteristics of fracture distribution(orientation and length)on borehole stability by sensitivity studies.Simulation results demonstrate that the increase of the standard deviation of orientation when the fracture direction aligns parallel or perpendicular to the principal stress direction aggravates borehole instability.Moreover,an elevation in the average fracture length causes the borehole failure to change from the direction of the minimum in-situ horizontal principal stress(i.e.the direction of wellbore breakouts)towards alternative directions,ultimately leading to the whole wellbore failure.These findings provide theoretical insights for predicting wellbore breakouts in heavily fractured rocks.
基金Under the auspices of National High Technology Research and Development Program of China (No.2007AA12Z242)
文摘The technique of incremental updating,which can better guarantee the real-time situation of navigational map,is the developing orientation of navigational road network updating.The data center of vehicle navigation system is in charge of storing incremental data,and the spatio-temporal data model for storing incremental data does affect the efficiency of the response of the data center to the requirements of incremental data from the vehicle terminal.According to the analysis on the shortcomings of several typical spatio-temporal data models used in the data center and based on the base map with overlay model,the reverse map with overlay model (RMOM) was put forward for the data center to make rapid response to incremental data request.RMOM supports the data center to store not only the current complete road network data,but also the overlays of incremental data from the time when each road network changed to the current moment.Moreover,the storage mechanism and index structure of the incremental data were designed,and the implementation algorithm of RMOM was developed.Taking navigational road network in Guangzhou City as an example,the simulation test was conducted to validate the efficiency of RMOM.Results show that the navigation database in the data center can response to the requirements of incremental data by only one query with RMOM,and costs less time.Compared with the base map with overlay model,the data center does not need to temporarily overlay incremental data with RMOM,so time-consuming of response is significantly reduced.RMOM greatly improves the efficiency of response and provides strong support for the real-time situation of navigational road network.
基金supported in part by the Science and Technology Project of Hebei Education Department(No.ZD2021088)in part by the S&T Major Project of the Science and Technology Ministry of China(No.2017YFE0135700)。
文摘Spatio-temporal cellular network traffic prediction at wide-area level plays an important role in resource reconfiguration,traffic scheduling and intrusion detection,thus potentially supporting connected intelligence of the sixth generation of mobile communications technology(6G).However,the existing studies just focus on the spatio-temporal modeling of traffic data of single network service,such as short message,call,or Internet.It is not conducive to accurate prediction of traffic data,characterised by diverse network service,spatio-temporality and supersize volume.To address this issue,a novel multi-task deep learning framework is developed for citywide cellular network traffic prediction.Functionally,this framework mainly consists of a dual modular feature sharing layer and a multi-task learning layer(DMFS-MT).The former aims at mining long-term spatio-temporal dependencies and local spatio-temporal fluctuation trends in data,respectively,via a new combination of convolutional gated recurrent unit(ConvGRU)and 3-dimensional convolutional neural network(3D-CNN).For the latter,each task is performed for predicting service-specific traffic data based on a fully connected network.On the real-world Telecom Italia dataset,simulation results demonstrate the effectiveness of our proposal through prediction performance measure,spatial pattern comparison and statistical distribution verification.
基金support from the National Natural Science Foundation of China(Grant Nos.42320104003,42177175,and 42077247)the Fundamental Research Funds for the Central Universities.
文摘Accurate prediction of hydraulic fracture propagation is vital for Enhanced Geothermal System(EGS)design.We study the first hydraulic fracturing job at the GR1 well in the Gonghe Basin using field data,where the overall direction of hydraulic fractures does not show a delineated shape parallel to the maximum principal stress orientation.A field-scale numerical model based on the distinct element method is set up to carry out a fully coupled hydromechanical simulation,with the explicit representation of natural fractures via the discrete fracture network(DFN)approach.The effects of injection parameters and in situ stress on hydraulic fracture patterns are then quantitatively assessed.The study reveals that shear-induced deformation primarily governs the fracturing morphology in the GR1 well,driven by smaller injection rates and viscosities that promote massive activation of natural fractures,ultimately dominating the direction of hydraulic fracturing.Furthermore,the increase of in situ differential stress may promote shear damage of natural fracture surfaces,with the exact influence pattern depending on the combination of specific discontinuity properties and in situ stress state.Finally,we provide recommendations for EGS fracturing based on the influence characteristics of multiple parameters.This study can serve as an effective basis and reference for the design and optimization of EGS in the Gonghe basin and other sites.
基金supported by the National Natural Science Foundation of China(Grant No.12071042)the Beijing Natural Science Foundation(Grant No.1202004)Promoting the Development of University Classification-Student Innovation and Entrepreneurship Training Programme(Grant No.5112410857)。
文摘In this paper,we propose a symmetric difference data enhancement physics-informed neural network(SDE-PINN)to study soliton solutions for discrete nonlinear lattice equations(NLEs).By considering known and unknown symmetric points,numerical simulations are conducted to one-soliton and two-soliton solutions of a discrete KdV equation,as well as a one-soliton solution of a discrete Toda lattice equation.Compared with the existing discrete deep learning approach,the numerical results reveal that within the specified spatiotemporal domain,the prediction accuracy by SDE-PINN is excellent regardless of the interior or extrapolation prediction,with a significant reduction in training time.The proposed data enhancement technique and symmetric structure development provides a new perspective for the deep learning approach to solve discrete NLEs.The newly proposed SDE-PINN can also be applied to solve continuous nonlinear equations and other discrete NLEs numerically.
基金Kinross Gold and MITACS for their financial support(Grant No.FR42880).
文摘The goal of this research is to develop mine-scale discrete fracture network(DFN)models in which the influence of the spatial heterogeneity of fracture distributions may be investigated on the rock wedge stability of an open pit slope.For this purpose,spatially conditioned DFN models were developed for the pit walls at Tasiast mine using comprehensive structural data from the mine.Using Sequential Gaussian Simulation(SGS),volumetric fracture intensities(P32)were modeled across the entire mine site in the form of 3D block models.The simulated P32 block models were used as the input constraints for conditional DFN fracture generation,where the DFN grid dimension is the same as the SGS 3D blocks.The spatially constrained DFN models were further calibrated using aerial fracture intensities(P21)data from the pit walls,obtained by a survey of the pit walls using an unmanned aerial vehicle(UAV)and measured traces of joints from 3D point cloud data.The final DFN model is expected to honor the fracture intensities gathered through different means with optimal model accuracy.Finally,bench-scale and interramp scale rock wedge slope stability analyses were conducted using the calibrated conditional DFN models.This work proves the significance of conditioned DFN models in rock wedge stability analysis.Such models provide detailed information regarding rock wedge stability so that site monitoring and prevention plans can be conducted with higher efficiency.
基金Youth Innovation Promotion Association CAS,Grant/Award Number:2021103Strategic Priority Research Program of Chinese Academy of Sciences,Grant/Award Number:XDC02060500。
文摘Appropriately characterising the mixed space-time relations of the contagion process caused by hybrid space and time factors remains the primary challenge in COVID-19 forecasting.However,in previous deep learning models for epidemic forecasting,spatial and temporal variations are captured separately.A unified model is developed to cover all spatio-temporal relations.However,this measure is insufficient for modelling the complex spatio-temporal relations of infectious disease transmission.A dynamic adaptive spatio-temporal graph network(DASTGN)is proposed based on attention mechanisms to improve prediction accuracy.In DASTGN,complex spatio-temporal relations are depicted by adaptively fusing the mixed space-time effects and dynamic space-time dependency structure.This dual-scale model considers the time-specific,space-specific,and direct effects of the propagation process at the fine-grained level.Furthermore,the model characterises impacts from various space-time neighbour blocks under time-varying interventions at the coarse-grained level.The performance comparisons on the three COVID-19 datasets reveal that DASTGN achieves state-of-the-art results with a maximum improvement of 17.092%in the root mean-square error and 11.563%in the mean absolute error.Experimental results indicate that the mechanisms of designing DASTGN can effectively detect some spreading characteristics of COVID-19.The spatio-temporal weight matrices learned in each proposed module reveal diffusion patterns in various scenarios.In conclusion,DASTGN has successfully captured the dynamic spatio-temporal variations of COVID-19,and considering multiple dynamic space-time relationships is essential in epidemic forecasting.
基金the National Natural Science Foundation of China(61573285).
文摘Owing to the wide range of applications in various fields,generative models have become increasingly popular.However,they do not handle spatio-temporal features well.Inspired by the recent advances in these models,this paper designs a distributed spatio-temporal generative adversarial network(STGAN-D)that,given some initial data and random noise,generates a consecutive sequence of spatio-temporal samples which have a logical relationship.This paper builds a spatio-temporal discriminator to distinguish whether the samples generated by the generator meet the requirements for time and space coherence,and builds a controller for distributed training of the network gradient updated to separate the model training and parameter updating,to improve the network training rate.The model is trained on the skeletal dataset and the traffic dataset.In contrast to traditional generative adversarial networks(GANs),the proposed STGAN-D can generate logically coherent samples with the corresponding spatial and temporal features while avoiding mode collapse.In addition,this paper shows that the proposed model can generate different styles of spatio-temporal samples given different random noise inputs,and the controller can improve the network training rate.This model will extend the potential range of applications of GANs to areas such as traffic information simulation and multiagent adversarial simulation.
基金the National Natural Science Fundation of China (10377014).
文摘The manner and conditions of running the decision-making system with self-defense electronic jamming are given. After proposing the scenario of applying discrete dynamic Bayesian network to the decision making with self-defense electronic jamming, a decision-making model with self-defense electronic jamming based on the discrete dynamic Bayesian network is established. Then jamming decision inferences by the aid of the algorithm of discrete dynamic Bayesian network are carried on. The simulating result shows that this method is able to synthesize different targets which are not predominant. In this way, various features at the same time, as well as the same feature appearing at different time complement mutually; in addition, the accuracy and reliability of electronic jamming decision making are enhanced significantly.
基金funded by the NIOSH Mining Program under Contract No. 200-2016-91300
文摘Terrestrial laser scanning(TLS) is a useful technology for rock mass characterization. A laser scanner produces a massive point cloud of a scanned area, such as an exposed rock surface in an underground tunnel,with millimeter precision. The density of the point cloud depends on several parameters from both the TLS operational conditions and the specifications of the project, such as the resolution and the quality of the laser scan, the section of the tunnel, the distance between scanning stations, and the purpose of the scans. One purpose of the scan can be to characterize the rock mass and statistically analyze the discontinuities that compose it for further discontinuous modeling. In these instances, additional data processing and a detailed analysis should be performed on the point cloud to extract the parameters to define a discrete fracture network(DFN) for each discontinuity set. I-site studio is a point cloud processing software that allows users to edit and process laser scans. This software contains a set of geotechnical analysis tools that assist engineers during the structural mapping process, allowing for greater and more representative data regarding the structural information of the rock mass, which may be used for generating DFNs. This paper presents the procedures used during a laser scan for characterizing discontinuities in an underground limestone mine and the results of the scan as applied to the generation of DFNs for further discontinuous modeling.
文摘Heterogeneity is an inherent component of rock and may be present in different forms including mineralheterogeneity, geometrical heterogeneity, weak grain boundaries and micro-defects. Microcracks areusually observed in crystalline rocks in two forms： natural and stress-induced; the amount of stressinducedmicrocracking increases with depth and in-situ stress. Laboratory results indicate that thephysical properties of rocks such as strength, deformability, P-wave velocity and permeability areinfluenced by increase in microcrack intensity. In this study, the finite-discrete element method （FDEM）is used to model microcrack heterogeneity by introducing into a model sample sets of microcracks usingthe proposed micro discrete fracture network （mDFN） approach. The characteristics of the microcracksrequired to create mDFN models are obtained through image analyses of thin sections of Lac du Bonnetgranite adopted from published literature. A suite of two-dimensional laboratory tests including uniaxial,triaxial compression and Brazilian tests is simulated and the results are compared with laboratory data.The FDEM-mDFN models indicate that micro-heterogeneity has a profound influence on both the mechanicalbehavior and resultant fracture pattern. An increase in the microcrack intensity leads to areduction in the strength of the sample and changes the character of the rock strength envelope. Spallingand axial splitting dominate the failure mode at low confinement while shear failure is the dominantfailure mode at high confinement. Numerical results from simulated compression tests show thatmicrocracking reduces the cohesive component of strength alone, and the frictional strength componentremains unaffected. Results from simulated Brazilian tests show that the tensile strength is influenced bythe presence of microcracks, with a reduction in tensile strength as microcrack intensity increases. Theimportance of microcrack heterogeneity in reproducing a bi-linear or S-shape failure envelope and itseffects on the mechanisms leading to spalling damage near an underground opening are also discussed.
基金co-supported by the National Basic Research Program of China(Grant No.2011CB201103)the National Science and Technology Major Project(GrantNo.2011ZX05004003)
文摘Fractured reservoirs are an important target for oil and gas exploration in the Tarim Basin and the prediction of this type of reservoir is challenging.Due to the complicated fracture system in the Tarim Basin,the conventional AVO inversion method based on HTI theory to predict fracture development will result in some errors.Thus,an integrated research concept for fractured reservoir prediction is put forward in this paper.Seismic modeling plays a bridging role in this concept,and the establishment of an anisotropic fracture model by Discrete Fracture Network （DFN） is the key part.Because the fracture system in the Tarim Basin shows complex anisotropic characteristics,it is vital to build an effective anisotropic model.Based on geological,well logging and seismic data,an effective anisotropic model of complex fracture systems can be set up with the DFN method.The effective elastic coefficients,and the input data for seismic modeling can be calculated.Then seismic modeling based on this model is performed,and the seismic response characteristics are analyzed.The modeling results can be used in the following AVO inversion for fracture detection.
基金Project supported by the National Natural Science Foundations of China(Grant No.70871056)the Society Science Foundation from Ministry of Education of China(Grant No.08JA790057)the Advanced Talents'Foundation and Student's Foundation of Jiangsu University,China(Grant Nos.07JDG054 and 07A075)
文摘This paper concernes analysis for the global exponential stability of a class of recurrent neural networks with mixed discrete and distributed delays. It first proves the existence and uniqueness of the balance point, then by employing the Lyapunov-Krasovskii functional and Young inequality, it gives the sufficient condition of global exponential stability of cellular neural network with mixed discrete and distributed delays, in addition, the example is provided to illustrate the applicability of the result.
基金supported by the Natural Science Foundation of Hubei Province, China (2017CFB434)the National Natural Science Foundation of China (41506208 and 61501200)the Basic Research Funds for Yellow River Institute of Hydraulic Research, China (HKYJBYW-2016-06)
文摘Mapping crop distribution with remote sensing data is of great importance for agricultural production, food security and agricultural sustainability. Winter rape is an important oil crop, which plays an important role in the cooking oil market of China. The Jianghan Plain and Dongting Lake Plain (JPDLP) are major agricultural production areas in China. Essential changes in winter rape distribution have taken place in this area during the 21st century. However, the pattern of these changes remains unknown. In this study, the spatial and temporal dynamics of winter rape from 2000 to 2017 on the JPDLP were analyzed. An artificial neural network (ANN)-based classification method was proposed to map fractional winter rape distribution by fusing moderate resolution imaging spectrometer (MODIS) data and high-resolution imagery. The results are as follows:(1) The total winter rape acreages on the JPDLP dropped significantly, especially on the Jianghan Plain with a decline of about 45% during 2000 and 2017.(2) The winter rape abundance keeps changing with about 20–30% croplands changing their abundance drastically in every two consecutive observation years.(3) The winter rape has obvious regional differentiation for the trend of its change at the county level, and the decreasing trend was observed more strongly in the traditionally dominant agricultural counties.
基金This work was financially supported by the National Key R&D Program of China(No.2021YFC2900500)the National Natural Science Foundation of China(Nos.52074020 and 42202306)+2 种基金the Open Fund of State Key Laboratory of Water Resource Protection and Utilization in Coal Mining(No.WPUKFJJ2019-06)the Interdisciplinary Research Project for Young Teachers of USTB(Fundamental Research Funds for the Central Universities)(No.FRF-IDRY-21001)the Natural Science Foundation of Jiangsu Province,China(No.BK20200993).
文摘Structure plane is one of the important factors affecting the stability and failure mode of rock mass engineering.Rock mass structure characterization is the basic work of rock mechanics research and the important content of numerical simulation.A new 3-dimensional rough discrete fracture network(RDFN3D)model and its modeling method based on the Weierstrass-Mandelbrot(W-M)function were presented in this paper.The RDFN3D model,which improves and unifies the modelling methods for the complex structural planes,has been realized.The influence of fractal dimension,amplitude,and surface precision on the modeling parameters of RDFN3D was discussed.The reasonable W-M parameters suitable for the roughness coefficient of JRC were proposed,and the relationship between the mathematical model and the joint characterization was established.The RDFN3D together with the smooth 3-dimensional discrete fracture network(DFN3D)models were successfully exported to the drawing exchange format,which will provide a wide application in numerous numerical simulation codes including both the continuous and discontinuous methods.The numerical models were discussed using the COMSOL Multiphysics code and the 3-dimensional particle flow code,respectively.The reliability of the RDFN3D model was preliminarily discussed and analyzed.The roughness and spatial connectivity of the fracture networks have a dominant effect on the fluid flow patterns.The research results can provide a new geological model and analysis model for numerical simulation and engineering analysis of jointed rock mass.
基金supported by the Fundamental Scientific Research Business Expenses for Central Universities(3072021CFJ0803)the Advanced Marine Communication and Information Technology Ministry of Industry and Information Technology Key Laboratory Project(AMCIT21V3).
文摘The accuracy of target threat estimation has a great impact on command decision-making.The Bayesian network,as an effective way to deal with the problem of uncertainty,can be used to track the change of the target threat level.Unfortunately,the traditional discrete dynamic Bayesian network(DDBN)has the problems of poor parameter learning and poor reasoning accuracy in a small sample environment with partial prior information missing.Considering the finiteness and discreteness of DDBN parameters,a fuzzy k-nearest neighbor(KNN)algorithm based on correlation of feature quantities(CF-FKNN)is proposed for DDBN parameter learning.Firstly,the correlation between feature quantities is calculated,and then the KNN algorithm with fuzzy weight is introduced to fill the missing data.On this basis,a reasonable DDBN structure is constructed by using expert experience to complete DDBN parameter learning and reasoning.Simulation results show that the CF-FKNN algorithm can accurately fill in the data when the samples are seriously missing,and improve the effect of DDBN parameter learning in the case of serious sample missing.With the proposed method,the final target threat assessment results are reasonable,which meets the needs of engineering applications.
基金This work was supported by financial support from Universiti Sains Malaysia(USM)under FRGS grant number FRGS/1/2020/TK03/USM/02/1the School of Computer Sciences USM for their support.
文摘Human Activity Recognition(HAR)is an active research area due to its applications in pervasive computing,human-computer interaction,artificial intelligence,health care,and social sciences.Moreover,dynamic environments and anthropometric differences between individuals make it harder to recognize actions.This study focused on human activity in video sequences acquired with an RGB camera because of its vast range of real-world applications.It uses two-stream ConvNet to extract spatial and temporal information and proposes a fine-tuned deep neural network.Moreover,the transfer learning paradigm is adopted to extract varied and fixed frames while reusing object identification information.Six state-of-the-art pre-trained models are exploited to find the best model for spatial feature extraction.For temporal sequence,this study uses dense optical flow following the two-stream ConvNet and Bidirectional Long Short TermMemory(BiLSTM)to capture longtermdependencies.Two state-of-the-art datasets,UCF101 and HMDB51,are used for evaluation purposes.In addition,seven state-of-the-art optimizers are used to fine-tune the proposed network parameters.Furthermore,this study utilizes an ensemble mechanism to aggregate spatial-temporal features using a four-stream Convolutional Neural Network(CNN),where two streams use RGB data.In contrast,the other uses optical flow images.Finally,the proposed ensemble approach using max hard voting outperforms state-ofthe-art methods with 96.30%and 90.07%accuracies on the UCF101 and HMDB51 datasets.
