Background Lip reading uses lip images for visual speech recognition.Deep-learning-based lip reading has greatly improved performance in current datasets;however,most existing research ignores the significance of shor...Background Lip reading uses lip images for visual speech recognition.Deep-learning-based lip reading has greatly improved performance in current datasets;however,most existing research ignores the significance of short-term temporal dependencies of lip-shape variations between adjacent frames,which leaves space for further improvement in feature extraction.Methods This article presents a spatiotemporal feature fusion network(STDNet)that compensates for the deficiencies of current lip-reading approaches in short-term temporal dependency modeling.Specifically,to distinguish more similar and intricate content,STDNet adds a temporal feature extraction branch based on a 3D-CNN,which enhances the learning of dynamic lip movements in adjacent frames while not affecting spatial feature extraction.In particular,we designed a local–temporal block,which aggregates interframe differences,strengthening the relationship between various local lip regions through multiscale convolution.We incorporated the squeeze-and-excitation mechanism into the Global-Temporal Block,which processes a single frame as an independent unitto learn temporal variations across the entire lip region more effectively.Furthermore,attention pooling was introduced to highlight meaningful frames containing key semantic information for the target word.Results Experimental results demonstrated STDNet's superior performance on the LRW and LRW-1000,achieving word-level recognition accuracies of 90.2% and 53.56%,respectively.Extensive ablation experiments verified the rationality and effectiveness of its modules.Conclusions The proposed model effectively addresses short-term temporal dependency limitations in lip reading,and improves the temporal robustness of the model against variable-length sequences.These advancements validate the importance of explicit short-term dynamics modeling for practical lip-reading systems.展开更多
Temporal and spatial scales play important roles in fishery ecology,and an inappropriate spatio-temporal scale may result in large errors in modeling fish distribution.The objective of this study is to evaluate the ro...Temporal and spatial scales play important roles in fishery ecology,and an inappropriate spatio-temporal scale may result in large errors in modeling fish distribution.The objective of this study is to evaluate the roles of spatio-temporal scales in habitat suitability modeling,with the western stock of winter-spring cohort of neon flying squid （Ornmastrephes bartramii） in the northwest Pacific Ocean as an example.In this study,the fishery-dependent data from the Chinese Mainland Squid Jigging Technical Group and sea surface temperature （SST） from remote sensing during August to October of 2003-2008 were used.We evaluated the differences in a habitat suitability index model resulting from aggregating data with 36 different spatial scales with a combination of three latitude scales （0.5°,1 ° and 2°）,four longitude scales （0.5°,1°,2° and 4°）,and three temporal scales （week,fortnight,and month）.The coefficients of variation （CV） of the weekly,biweekly and monthly suitability index （SI） were compared to determine which temporal and spatial scales of SI model are more precise.This study shows that the optimal temporal and spatial scales with the lowest CV are month,and 0.5° latitude and 0.5° longitude for O.bartramii in the northwest Pacific Ocean.This suitability index model developed with an optimal scale can be cost-effective in improving forecasting fishing ground and requires no excessive sampling efforts.We suggest that the uncertainty associated with spatial and temporal scales used in data aggregations needs to be considered in habitat suitability modeling.展开更多
Objectives:The prevalence of cyber-aggression is increasing worldwide,resulting in significant negative impacts on both perpetrators and victims.This study aimed to investigate the relationship between cyber-ostracism...Objectives:The prevalence of cyber-aggression is increasing worldwide,resulting in significant negative impacts on both perpetrators and victims.This study aimed to investigate the relationship between cyber-ostracism and cyber-aggression among college students,clarify the role of various types of rumination in this dynamic.Methods:A total of 1198 Chinese college students(67.4%female;mean age 20.78 years;SD=1.12)were recruited through cluster random sampling and completed the Cyber-ostracism Experience Scale(COES),Positive and Negative Rumination Scale(PANRS),and Adolescent Online Aggression Behavior Scale(AOABS).Thestructural equation model(SEM)was employed to examine the relationship between cyber-ostracism,negative rumination,and cyber-aggression,as well as the moderating effect of positive rumination.Results:The results indicate that cyber-ostracism(β=0.128,p<0.001)positively predicts cyber-aggression.Negative ruminationmediates this relationship(effect size=0.027,95%CI=[0.007,0.014]).Positive rumination moderates the direct effect of cyber-ostracism on cyber-aggression(β=0.103,p<0.001).It alsomoderates both the first half(β=0.148,p<0.001)and the second half(β=0.138,p<0.001)of themediating pathway.Conclusion:This study suggests that cyber-ostracism influences cyber-aggression through negative rumination among Chinese college students.Positive rumination moderates this effect,although its impact is relatively limited.These findings offer valuable guidance for preventing and intervening in cyber-aggression among college students.展开更多
Since April 2002,the Gravity Recovery and Climate Experiment Satellite(GRACE)has provided monthly total water storage anomalies(TWSAs)on a global scale.However,these TWSAs are discontinuous because some GRACE observat...Since April 2002,the Gravity Recovery and Climate Experiment Satellite(GRACE)has provided monthly total water storage anomalies(TWSAs)on a global scale.However,these TWSAs are discontinuous because some GRACE observation data are missing.This study presents a combined machine learning-based modeling algorithm without hydrological model data.The TWSA time-series data for 11 large regions worldwide were divided into training and test sets.Autoregressive integrated moving average(ARIMA),long short-term memory(LSTM),and an ARIMA-LSTM combined model were used.The model predictions were compared with GRACE observations,and the model accuracy was evaluated using fi ve metrics:the Nash-Sutcliff e effi ciency coeffi cient(NSE),Pearson correlation coeffi cient(CC),root mean square error(RMSE),normalized RMSE(NRMSE),and mean absolute percentage error.The results show that at the basin scale,the mean CC,NSE,and NRMSE for the ARIMA-LSTM model were 0.93,0.83,and 0.12,respectively.At the grid scale,this study compared the spatial distribution and cumulative distribution function curves of the metrics in the Amazon and Volga River basins.The ARIMA-LSTM model had mean CC and NSE values of 0.89 and 0.61 and 0.92 and 0.61 in the Amazon and Volga River basins,respectively,which are superior to those of the ARIMA model(0.86 and 0.48 and 0.88 and 0.46,respectively)and the LSTM model(0.80 and 0.41 and 0.89 and 0.31,respectively).In the ARIMA-LSTM model,the proportions of grid cells with NSE>0.50 for the two basins were 63.3%and 80.8%,while they were 54.3%and 51.3%in the ARIMA model and 53.7%and 43.2%in the LSTM model.The ARIMA-LSTM model significantly improved the NSE values of the predictions while guaranteeing high CC values in the GRACE data reconstruction at both scales,which can aid in fi lling in discontinuous data in temporal gravity fi eld models..展开更多
The global clustering of inventive talent shapes innovation capacity and drives economic growth.For China,this process is especially crucial in sustaining its development momentum.This paper draws on data from the EPO...The global clustering of inventive talent shapes innovation capacity and drives economic growth.For China,this process is especially crucial in sustaining its development momentum.This paper draws on data from the EPO Worldwide Patent Statistical Database(PATSTAT)to extract global inventive talent mobility information and analyzes the spatial structural evolution of the global inventive talent flow network.The study finds that this network is undergoing a multi-polar transformation,characterized by the rising importance of a few central countries-such as the United States,Germany,and China-and the increasing marginalization of many peripheral countries.In response to this typical phenomenon,the paper constructs an endogenous migration model and conducts empirical testing using the Temporal Exponential Random Graph Model(TERGM).The results reveal several endogenous mechanisms driving global inventive talent flows,including reciprocity,path dependence,convergence effects,transitivity,and cyclic structures,all of which contribute to the network’s multi-polar trend.In addition,differences in regional industrial structures significantly influence talent mobility choices and are a decisive factor in the formation of poles within the multi-polar landscape.Based on these findings,it is suggested that efforts be made to foster two-way channels for talent exchange between China and other global innovation hubs,in order to enhance international collaboration and knowledge flow.We should aim to reduce the migration costs and institutional barriers faced by R&D personnel,thereby encouraging greater mobility of high-skilled talent.Furthermore,the government is advised to strategically leverage regional strengths in high-tech industries as a lever to capture competitive advantages in emerging technologies and products,ultimately strengthening the country’s position in the global innovation landscape.展开更多
As one of the main characteristics of atmospheric pollutants,PM_(2.5) severely affects human health and has received widespread attention in recent years.How to predict the variations of PM_(2.5) concentrations with h...As one of the main characteristics of atmospheric pollutants,PM_(2.5) severely affects human health and has received widespread attention in recent years.How to predict the variations of PM_(2.5) concentrations with high accuracy is an important topic.The PM_(2.5) monitoring stations in Xinjiang Uygur Autonomous Region,China,are unevenly distributed,which makes it challenging to conduct comprehensive analyses and predictions.Therefore,this study primarily addresses the limitations mentioned above and the poor generalization ability of PM_(2.5) concentration prediction models across different monitoring stations.We chose the northern slope of the Tianshan Mountains as the study area and took the January−December in 2019 as the research period.On the basis of data from 21 PM_(2.5) monitoring stations as well as meteorological data(temperature,instantaneous wind speed,and pressure),we developed an improved model,namely GCN−TCN−AR(where GCN is the graph convolution network,TCN is the temporal convolutional network,and AR is the autoregression),for predicting PM_(2.5) concentrations on the northern slope of the Tianshan Mountains.The GCN−TCN−AR model is composed of an improved GCN model,a TCN model,and an AR model.The results revealed that the R2 values predicted by the GCN−TCN−AR model at the four monitoring stations(Urumqi,Wujiaqu,Shihezi,and Changji)were 0.93,0.91,0.93,and 0.92,respectively,and the RMSE(root mean square error)values were 6.85,7.52,7.01,and 7.28μg/m^(3),respectively.The performance of the GCN−TCN−AR model was also compared with the currently neural network models,including the GCN−TCN,GCN,TCN,Support Vector Regression(SVR),and AR.The GCN−TCN−AR outperformed the other current neural network models,with high prediction accuracy and good stability,making it especially suitable for the predictions of PM_(2.5)concentrations.This study revealed the significant spatiotemporal variations of PM_(2.5)concentrations.First,the PM_(2.5) concentrations exhibited clear seasonal fluctuations,with higher levels typically observed in winter and differences presented between months.Second,the spatial distribution analysis revealed that cities such as Urumqi and Wujiaqu have high PM_(2.5) concentrations,with a noticeable geographical clustering of pollutions.Understanding the variations in PM_(2.5) concentrations is highly important for the sustainable development of ecological environment in arid areas.展开更多
This paper presents a conceptual data model, the STA-model, for handling spatial, temporal and attribute aspects of objects in GIS. The model is developed on the basis of object-oriented modeling approach. This model ...This paper presents a conceptual data model, the STA-model, for handling spatial, temporal and attribute aspects of objects in GIS. The model is developed on the basis of object-oriented modeling approach. This model includes two major parts: (a) modeling the signal objects by STA-object elements, and (b) modeling relationships between STA-objects. As an example, the STA-model is applied for modeling land cover change data with spatial, temporal and attribute components.展开更多
ISDTM,based on an augmented Allen's interval temporal logic(ITL)and first-order predicate calculus,is a formal temporal model for representing intrusion signatures.It is augmented with some real time extensions wh...ISDTM,based on an augmented Allen's interval temporal logic(ITL)and first-order predicate calculus,is a formal temporal model for representing intrusion signatures.It is augmented with some real time extensions which enhance the expressivity.Intrusion scenarios usually are the set of events and system states,where-the temporal sequence is their basic relation.Intrusion signatures description,therefore,is to represent such temporal relations in a sense.While representing these signatures,ISDTM decomposes the intrusion process into the sequence of events according to their relevant intervals,and then specifies network states in these Intervals.The uncertain intrusion signatures as well as basic temporal modes of events,which consist of the parallel mode,the sequential mode and the hybrid mode,can be succinctly and naturally represented in ISDTM.Mode chart is the visualization of intrusion signatures in ISDTM,which makes the formulas more readable.The intrusion signatures descriptions in ISDTM have advantages of compact construct,concise syntax,scalability and easy implementation.展开更多
Root-cause identification plays a vital role in business decision making by providing effective future directions for the organizations.Aspect extraction and sentiment extraction plays a vital role in identifying the ...Root-cause identification plays a vital role in business decision making by providing effective future directions for the organizations.Aspect extraction and sentiment extraction plays a vital role in identifying the rootcauses.This paper proposes the Ensemble based temporal weighting and pareto ranking(ETP)model for Root-cause identification.Aspect extraction is performed based on rules and is followed by opinion identification using the proposed boosted ensemble model.The obtained aspects are validated and ranked using the proposed aspect weighing scheme.Pareto-rule based aspect selection is performed as the final selection mechanism and the results are presented for business decision making.Experiments were performed with the standard five product benchmark dataset.Performances on all five product reviews indicate the effective performance of the proposed model.Comparisons are performed using three standard state-of-the-art models and effectiveness is measured in terms of F-Measure and Detection rates.The results indicate improved performances exhibited by the proposed model with an increase in F-Measure levels at 1%–15%and detection rates at 4%–24%compared to the state-of-the-art models.展开更多
A new temporal gravity field model called WHU-Grace01s solely recovered from Gravity Recovery and Climate Experiment （GRACE） K-Band Range Rate （KBRR） data based on dynamic integral approach is presented in this pa...A new temporal gravity field model called WHU-Grace01s solely recovered from Gravity Recovery and Climate Experiment （GRACE） K-Band Range Rate （KBRR） data based on dynamic integral approach is presented in this paper. After meticulously preprocessing of the GRACE KBRR data, the root mean square of its post residuals is about 0.2 micrometers per second, and seventy-two monthly temporal solutions truncated to degree and order 60 are computed for the period from January 2003 to December 2008. After applying the combi- nation filter in WHU-Grace01s, the global temporal signals show obvious periodical change rules in the large-scale fiver basins. In terms of the degree variance, our solution is smaller at high degrees, and shows a good consistency at the rest of degrees with the Release 05 models from Center for Space Research （CSR）, GeoForschungsZentrum Potsdam （GFZ） and Jet Pro- pulsion Laboratory 0PL）. Compared with other published models in terms of equivalent water height distribution, our solution is consistent with those published by CSR, GFZ, JPL, Delft institute of Earth Observation and Space system （DEOS）, Tongji University （Tongji）, Institute of Theoretical Geodesy （ITG）, Astronomical Institute in University of Bern （AIUB） and Groupe de Recherche de Geodesie Spatiale （GRGS}, which indicates that the accuracy of WHU-Grace01s has a good consistency with the previously published GRACE solutions.展开更多
Time is an important dimension for information in the geographical information system. Data, such as the historical state of target property space and related events causing the state to be changed, should be saved as...Time is an important dimension for information in the geographical information system. Data, such as the historical state of target property space and related events causing the state to be changed, should be saved as important files. This should be applied to property management. This paper designs and constructs a spatial temporal model, which is suitable to the property data changing management and spatial temporal query by analyzing the basic types and characteristics of property management spatial changing time and date. This model uses current and historical situational layers to organize and set up the relationship between current situation data and historical dates according to spatial temporal topological relations in property entities. By using Map Basic, housing property management and spatial query is realized.展开更多
A high altitude platform station （HAPS） based communications scenario for providing Intemet access and broadband multimedia services to the passengers on board of a high speed train （traveling up to 300km/h） is pr...A high altitude platform station （HAPS） based communications scenario for providing Intemet access and broadband multimedia services to the passengers on board of a high speed train （traveling up to 300km/h） is proposed. Regarding the addressed scenario, when the propagation link between HAPS and train is blocked by obstacles, a three-dimensional （3-D） geometrical single cylinder spatial-temporal channel model is presented, in which closed form, mathematically tractable space-time correlation functions are obtained. It shows that the correlation functions determined by the 3-D model are of significant difference with those of the conventional 2-D model. Based on the analysis model, the paper derives a realized simulation model using sum-of-sinusoids approach, and applies method of equal areas （MEA） and modified method of equal areas （MMEA） to determine the model parameters. The fitting performance of the simulation model with the analysis one is evaluated by two means-square error （MSE） performance criteria. Finally, numerical simulation results verify the mathematical analysis conclusion, when N ≥21, simulation model has an excellent fitness with the analysis one.展开更多
As there is datum redundancy in tradition database and temporal database in existence and the quantities of temporal database are increasing fleetly. We put forward compress storage tactics for temporal datum which co...As there is datum redundancy in tradition database and temporal database in existence and the quantities of temporal database are increasing fleetly. We put forward compress storage tactics for temporal datum which combine compress technology in existence in order to settle datum redundancy in the course of temporal datum storage and temporal datum of slow acting domain and momentary acting domain are accessed by using each from independence clock method and mutual clock method .We also bring forward strategy of gridding storage to resolve the problems of temporal datum rising rapidly.展开更多
The hands and face are the most important parts for expressing sign language morphemes in sign language videos.However,we find that existing Continuous Sign Language Recognition(CSLR)methods lack the mining of hand an...The hands and face are the most important parts for expressing sign language morphemes in sign language videos.However,we find that existing Continuous Sign Language Recognition(CSLR)methods lack the mining of hand and face information in visual backbones or use expensive and time-consuming external extractors to explore this information.In addition,the signs have different lengths,whereas previous CSLR methods typically use a fixed-length window to segment the video to capture sequential features and then perform global temporal modeling,which disturbs the perception of complete signs.In this study,we propose a Multi-Scale Context-Aware network(MSCA-Net)to solve the aforementioned problems.Our MSCA-Net contains two main modules:(1)Multi-Scale Motion Attention(MSMA),which uses the differences among frames to perceive information of the hands and face in multiple spatial scales,replacing the heavy feature extractors;and(2)Multi-Scale Temporal Modeling(MSTM),which explores crucial temporal information in the sign language video from different temporal scales.We conduct extensive experiments using three widely used sign language datasets,i.e.,RWTH-PHOENIX-Weather-2014,RWTH-PHOENIX-Weather-2014T,and CSL-Daily.The proposed MSCA-Net achieve state-of-the-art performance,demonstrating the effectiveness of our approach.展开更多
Temporal attention mechanisms are essential for video action recognition,enabling models to focus on semantically informative moments.However,these models frequently exhibit temporal infidelitymisaligned attention wei...Temporal attention mechanisms are essential for video action recognition,enabling models to focus on semantically informative moments.However,these models frequently exhibit temporal infidelitymisaligned attention weights caused by limited training diversity and the absence of fine-grained temporal supervision.While video-level labels provide coarse-grained action guidance,the lack of detailed constraints allows attention noise to persist,especially in complex scenarios with distracting spatial elements.To address this issue,we propose temporal fidelity enhancement(TFE),a competitive learning paradigm based on the disentangled information bottleneck(DisenIB)theory.TFE mitigates temporal infidelity by decoupling action-relevant semantics from spurious correlations through adversarial feature disentanglement.Using pre-trained representations for initialization,TFE establishes an adversarial process in which segments with elevated temporal attention compete against contexts with diminished action relevance.This mechanism ensures temporal consistency and enhances the fidelity of attention patterns without requiring explicit fine-grained supervision.Extensive studies on UCF101,HMDB-51,and Charades benchmarks validate the effectiveness of our method,with significant improvements in action recognition accuracy.展开更多
Deep learning has found widespread application across diverse domains owing to its exceptional performance.Nevertheless,the lack of transparency in deep learning models’decision-making processes undermines their usab...Deep learning has found widespread application across diverse domains owing to its exceptional performance.Nevertheless,the lack of transparency in deep learning models’decision-making processes undermines their usability,especially in critical contexts.While researchers have made noteworthy advancements in explaining these models,they have frequently overlooked the differences between static and temporal models during explanation generation.In temporal models,features change over time,posing new challenges in the generation of explanations.Though extensive research has been dedicated to surmounting these hurdles,a survey summarizing these contributions is currently absent.To bridge this gap,this paper endeavors to summarize existing methods and their contributions in terms of both static and temporal models,highlighting their disparities.Additionally,we propose an innovative classification approach based on the comprehensibility of explanations,demonstrating that different explanation methods vary in their understandability for users.Finally,to assess the limitations of the explanation capabilities of existing methods,we specifically choose knowledge tracing to analyze the evolution of explanation methods in this context of temporal modeling and interpretations.展开更多
Intelligent spatial-temporal data analysis,leveraging data such as multivariate time series and geographic information,provides researchers with powerful tools to uncover multiscale patterns and enhance decision-makin...Intelligent spatial-temporal data analysis,leveraging data such as multivariate time series and geographic information,provides researchers with powerful tools to uncover multiscale patterns and enhance decision-making processes.As artificial intelligence advances,intelligent spatial-temporal algorithms have found extensive applications across various disciplines,such as geosciences,biology,and public health.1 Compared to traditional methods,these algorithms are data driven,making them well suited for addressing the complexities of modeling real-world systems.However,their reliance on substantial domain-specific expertise limits their broader applicability.Recently,significant advancements have been made in spatial-temporal large models.Trained on large-scale data,these models exhibit a vast parameter scale,superior generalization capabilities,and multitasking advantages over previous methods.Their high versatility and scalability position them as promising super hubs for multidisciplinary research,integrating knowledge,intelligent algorithms,and research communities from different fields.Nevertheless,achieving this vision will require overcoming numerous critical challenges,offering an expansive and profound space for future exploration.展开更多
Large-eddy simulations （LES） based on the temporal approximate deconvolution model were performed for a forced homogeneous isotropic turbulence （FHIT） with polymer additives at moderate Taylor Reynolds number. Fin...Large-eddy simulations （LES） based on the temporal approximate deconvolution model were performed for a forced homogeneous isotropic turbulence （FHIT） with polymer additives at moderate Taylor Reynolds number. Finitely extensible nonlinear elastic in the Peterlin approximation model was adopted as the constitutive equation for the filtered conformation tensor of the polymer molecules. The LES results were verified through comparisons with the direct numerical simulation results. Using the LES database of the FHIT in the Newtonian fluid and the polymer solution flows, the polymer effects on some important parameters such as strain, vorticity, drag reduction, and so forth were studied. By extracting the vortex structures and exploring the flatness factor through a high-order correlation function of velocity derivative and wavelet analysis, it can be found that the small-scale vortex structures and small-scale intermittency in the FHIT are all inhibited due to the existence of the polymers. The extended self-similarity scaling law in the polymer solution flow shows no apparent difference from that in the Newtonian fluid flow at the currently simulated ranges of Reynolds and Weissenberg numbers.展开更多
The Gravity Recovery and Climate Experiment（GRACE） mission can significantly improve our knowledge of the temporal variability of the Earth＇s gravity field.We obtained monthly gravity field solutions based on varia...The Gravity Recovery and Climate Experiment（GRACE） mission can significantly improve our knowledge of the temporal variability of the Earth＇s gravity field.We obtained monthly gravity field solutions based on variational equations approach from GPS-derived positions of GRACE satellites and K-band range-rate measurements.The impact of different fixed data weighting ratios in temporal gravity field recovery while combining the two types of data was investigated for the purpose of deriving the best combined solution.The monthly gravity field solution obtained through above procedures was named as the Institute of Geodesy and Geophysics（IGG） temporal gravity field models.IGG temporal gravity field models were compared with GRACE Release05（RL05） products in following aspects：（i） the trend of the mass anomaly in China and its nearby regions within 2005-2010; （ii） the root mean squares of the global mass anomaly during 2005-2010; （iii） time-series changes in the mean water storage in the region of the Amazon Basin and the Sahara Desert between 2005 and 2010.The results showed that IGG solutions were almost consistent with GRACE RL05 products in above aspects（i）-（iii）.Changes in the annual amplitude of mean water storage in the Amazon Basin were 14.7 ± 1.2 cm for IGG,17.1 ± 1.3 cm for the Centre for Space Research（CSR）,16.4 ± 0.9 cm for the GeoForschungsZentrum（GFZ） and 16.9 ± 1.2 cm for the Jet Propulsion Laboratory（JPL） in terms of equivalent water height（EWH）,respectively.The root mean squares of the mean mass anomaly in Sahara were 1.2 cm,0.9 cm,0.9 cm and 1.2 cm for temporal gravity field models of IGG,CSR,GFZ and JPL,respectively.Comparison suggested that IGG temporal gravity field solutions were at the same accuracy level with the latest temporal gravity field solutions published by CSR,GFZ and JPL.展开更多
This paper proposes a mechanism theory on regional development by using a modified Logistic model. It reveals regional evolution is an integration of fluctuation in temporal dimension and disparity in spatial dimensio...This paper proposes a mechanism theory on regional development by using a modified Logistic model. It reveals regional evolution is an integration of fluctuation in temporal dimension and disparity in spatial dimension. T = S model is established by using Logistic model to simulate the growth of per capita GDP in China from 1990 to 1999. The result shows that T=S model accurately simulates the tracks of economic growth.展开更多
基金Supported by the National Key Research and Development Program of China(2023YFC3306201)the National Natural Science Foundation of China(61772125)the Fundamental Research Funds for the Central Universities(N2317004).
文摘Background Lip reading uses lip images for visual speech recognition.Deep-learning-based lip reading has greatly improved performance in current datasets;however,most existing research ignores the significance of short-term temporal dependencies of lip-shape variations between adjacent frames,which leaves space for further improvement in feature extraction.Methods This article presents a spatiotemporal feature fusion network(STDNet)that compensates for the deficiencies of current lip-reading approaches in short-term temporal dependency modeling.Specifically,to distinguish more similar and intricate content,STDNet adds a temporal feature extraction branch based on a 3D-CNN,which enhances the learning of dynamic lip movements in adjacent frames while not affecting spatial feature extraction.In particular,we designed a local–temporal block,which aggregates interframe differences,strengthening the relationship between various local lip regions through multiscale convolution.We incorporated the squeeze-and-excitation mechanism into the Global-Temporal Block,which processes a single frame as an independent unitto learn temporal variations across the entire lip region more effectively.Furthermore,attention pooling was introduced to highlight meaningful frames containing key semantic information for the target word.Results Experimental results demonstrated STDNet's superior performance on the LRW and LRW-1000,achieving word-level recognition accuracies of 90.2% and 53.56%,respectively.Extensive ablation experiments verified the rationality and effectiveness of its modules.Conclusions The proposed model effectively addresses short-term temporal dependency limitations in lip reading,and improves the temporal robustness of the model against variable-length sequences.These advancements validate the importance of explicit short-term dynamics modeling for practical lip-reading systems.
基金funded by National High Technology Research and Development Program of China (863 Program,2012AA092303)Project of Shanghai Science and Technology Innovation (12231203900)+2 种基金Industrialization Program of National Development and Reform Commission (2159999)National Science and Technology Support Program (2013BAD13B01)Shanghai Leading Academic Discipline Project
文摘Temporal and spatial scales play important roles in fishery ecology,and an inappropriate spatio-temporal scale may result in large errors in modeling fish distribution.The objective of this study is to evaluate the roles of spatio-temporal scales in habitat suitability modeling,with the western stock of winter-spring cohort of neon flying squid （Ornmastrephes bartramii） in the northwest Pacific Ocean as an example.In this study,the fishery-dependent data from the Chinese Mainland Squid Jigging Technical Group and sea surface temperature （SST） from remote sensing during August to October of 2003-2008 were used.We evaluated the differences in a habitat suitability index model resulting from aggregating data with 36 different spatial scales with a combination of three latitude scales （0.5°,1 ° and 2°）,four longitude scales （0.5°,1°,2° and 4°）,and three temporal scales （week,fortnight,and month）.The coefficients of variation （CV） of the weekly,biweekly and monthly suitability index （SI） were compared to determine which temporal and spatial scales of SI model are more precise.This study shows that the optimal temporal and spatial scales with the lowest CV are month,and 0.5° latitude and 0.5° longitude for O.bartramii in the northwest Pacific Ocean.This suitability index model developed with an optimal scale can be cost-effective in improving forecasting fishing ground and requires no excessive sampling efforts.We suggest that the uncertainty associated with spatial and temporal scales used in data aggregations needs to be considered in habitat suitability modeling.
基金supported by the Key Research Base Project of Humanities and Social Sciences in Jiangxi Colleges and Universities(JD23056)the Characteristic Programmers of Mental Health Education for Students,University of Tibetan Medicine(2024XLJKZD01).
文摘Objectives:The prevalence of cyber-aggression is increasing worldwide,resulting in significant negative impacts on both perpetrators and victims.This study aimed to investigate the relationship between cyber-ostracism and cyber-aggression among college students,clarify the role of various types of rumination in this dynamic.Methods:A total of 1198 Chinese college students(67.4%female;mean age 20.78 years;SD=1.12)were recruited through cluster random sampling and completed the Cyber-ostracism Experience Scale(COES),Positive and Negative Rumination Scale(PANRS),and Adolescent Online Aggression Behavior Scale(AOABS).Thestructural equation model(SEM)was employed to examine the relationship between cyber-ostracism,negative rumination,and cyber-aggression,as well as the moderating effect of positive rumination.Results:The results indicate that cyber-ostracism(β=0.128,p<0.001)positively predicts cyber-aggression.Negative ruminationmediates this relationship(effect size=0.027,95%CI=[0.007,0.014]).Positive rumination moderates the direct effect of cyber-ostracism on cyber-aggression(β=0.103,p<0.001).It alsomoderates both the first half(β=0.148,p<0.001)and the second half(β=0.138,p<0.001)of themediating pathway.Conclusion:This study suggests that cyber-ostracism influences cyber-aggression through negative rumination among Chinese college students.Positive rumination moderates this effect,although its impact is relatively limited.These findings offer valuable guidance for preventing and intervening in cyber-aggression among college students.
基金financially supported by The National Natural Science Foundation of China (42374004)the Open Fund of Hubei Luojia Laboratory (220100045)the Natural Science Foundation of Sichuan Province (2022NSFSC1047)。
文摘Since April 2002,the Gravity Recovery and Climate Experiment Satellite(GRACE)has provided monthly total water storage anomalies(TWSAs)on a global scale.However,these TWSAs are discontinuous because some GRACE observation data are missing.This study presents a combined machine learning-based modeling algorithm without hydrological model data.The TWSA time-series data for 11 large regions worldwide were divided into training and test sets.Autoregressive integrated moving average(ARIMA),long short-term memory(LSTM),and an ARIMA-LSTM combined model were used.The model predictions were compared with GRACE observations,and the model accuracy was evaluated using fi ve metrics:the Nash-Sutcliff e effi ciency coeffi cient(NSE),Pearson correlation coeffi cient(CC),root mean square error(RMSE),normalized RMSE(NRMSE),and mean absolute percentage error.The results show that at the basin scale,the mean CC,NSE,and NRMSE for the ARIMA-LSTM model were 0.93,0.83,and 0.12,respectively.At the grid scale,this study compared the spatial distribution and cumulative distribution function curves of the metrics in the Amazon and Volga River basins.The ARIMA-LSTM model had mean CC and NSE values of 0.89 and 0.61 and 0.92 and 0.61 in the Amazon and Volga River basins,respectively,which are superior to those of the ARIMA model(0.86 and 0.48 and 0.88 and 0.46,respectively)and the LSTM model(0.80 and 0.41 and 0.89 and 0.31,respectively).In the ARIMA-LSTM model,the proportions of grid cells with NSE>0.50 for the two basins were 63.3%and 80.8%,while they were 54.3%and 51.3%in the ARIMA model and 53.7%and 43.2%in the LSTM model.The ARIMA-LSTM model significantly improved the NSE values of the predictions while guaranteeing high CC values in the GRACE data reconstruction at both scales,which can aid in fi lling in discontinuous data in temporal gravity fi eld models..
基金supported by the Major Project of the National Social Science Fund of China,titled“Design Path Selection for the Mechanism of New and Old Growth Driver Conversion”(Grant No.18ZDA077)by the Joint Special Major Research Project of the Yangtze River Delta Economics and Social Development Research Center at Nanjing University and the Collaborative Innovation Center for China Economy(CICCE),titled“Practicing Innovation in China’s Development Economics for the Yangtze River Delta:From Industrial Clusters to Technological Clusters”(Grant No.CYD2022006).
文摘The global clustering of inventive talent shapes innovation capacity and drives economic growth.For China,this process is especially crucial in sustaining its development momentum.This paper draws on data from the EPO Worldwide Patent Statistical Database(PATSTAT)to extract global inventive talent mobility information and analyzes the spatial structural evolution of the global inventive talent flow network.The study finds that this network is undergoing a multi-polar transformation,characterized by the rising importance of a few central countries-such as the United States,Germany,and China-and the increasing marginalization of many peripheral countries.In response to this typical phenomenon,the paper constructs an endogenous migration model and conducts empirical testing using the Temporal Exponential Random Graph Model(TERGM).The results reveal several endogenous mechanisms driving global inventive talent flows,including reciprocity,path dependence,convergence effects,transitivity,and cyclic structures,all of which contribute to the network’s multi-polar trend.In addition,differences in regional industrial structures significantly influence talent mobility choices and are a decisive factor in the formation of poles within the multi-polar landscape.Based on these findings,it is suggested that efforts be made to foster two-way channels for talent exchange between China and other global innovation hubs,in order to enhance international collaboration and knowledge flow.We should aim to reduce the migration costs and institutional barriers faced by R&D personnel,thereby encouraging greater mobility of high-skilled talent.Furthermore,the government is advised to strategically leverage regional strengths in high-tech industries as a lever to capture competitive advantages in emerging technologies and products,ultimately strengthening the country’s position in the global innovation landscape.
基金supported by the Program of Support Xinjiang by Technology(2024E02028,B2-2024-0359)Xinjiang Tianchi Talent Program of 2024,the Foundation of Chinese Academy of Sciences(B2-2023-0239)the Youth Foundation of Shandong Natural Science(ZR2023QD070).
文摘As one of the main characteristics of atmospheric pollutants,PM_(2.5) severely affects human health and has received widespread attention in recent years.How to predict the variations of PM_(2.5) concentrations with high accuracy is an important topic.The PM_(2.5) monitoring stations in Xinjiang Uygur Autonomous Region,China,are unevenly distributed,which makes it challenging to conduct comprehensive analyses and predictions.Therefore,this study primarily addresses the limitations mentioned above and the poor generalization ability of PM_(2.5) concentration prediction models across different monitoring stations.We chose the northern slope of the Tianshan Mountains as the study area and took the January−December in 2019 as the research period.On the basis of data from 21 PM_(2.5) monitoring stations as well as meteorological data(temperature,instantaneous wind speed,and pressure),we developed an improved model,namely GCN−TCN−AR(where GCN is the graph convolution network,TCN is the temporal convolutional network,and AR is the autoregression),for predicting PM_(2.5) concentrations on the northern slope of the Tianshan Mountains.The GCN−TCN−AR model is composed of an improved GCN model,a TCN model,and an AR model.The results revealed that the R2 values predicted by the GCN−TCN−AR model at the four monitoring stations(Urumqi,Wujiaqu,Shihezi,and Changji)were 0.93,0.91,0.93,and 0.92,respectively,and the RMSE(root mean square error)values were 6.85,7.52,7.01,and 7.28μg/m^(3),respectively.The performance of the GCN−TCN−AR model was also compared with the currently neural network models,including the GCN−TCN,GCN,TCN,Support Vector Regression(SVR),and AR.The GCN−TCN−AR outperformed the other current neural network models,with high prediction accuracy and good stability,making it especially suitable for the predictions of PM_(2.5)concentrations.This study revealed the significant spatiotemporal variations of PM_(2.5)concentrations.First,the PM_(2.5) concentrations exhibited clear seasonal fluctuations,with higher levels typically observed in winter and differences presented between months.Second,the spatial distribution analysis revealed that cities such as Urumqi and Wujiaqu have high PM_(2.5) concentrations,with a noticeable geographical clustering of pollutions.Understanding the variations in PM_(2.5) concentrations is highly important for the sustainable development of ecological environment in arid areas.
文摘This paper presents a conceptual data model, the STA-model, for handling spatial, temporal and attribute aspects of objects in GIS. The model is developed on the basis of object-oriented modeling approach. This model includes two major parts: (a) modeling the signal objects by STA-object elements, and (b) modeling relationships between STA-objects. As an example, the STA-model is applied for modeling land cover change data with spatial, temporal and attribute components.
基金Supported by the National Natural Science Foundation of China(60073074)
文摘ISDTM,based on an augmented Allen's interval temporal logic(ITL)and first-order predicate calculus,is a formal temporal model for representing intrusion signatures.It is augmented with some real time extensions which enhance the expressivity.Intrusion scenarios usually are the set of events and system states,where-the temporal sequence is their basic relation.Intrusion signatures description,therefore,is to represent such temporal relations in a sense.While representing these signatures,ISDTM decomposes the intrusion process into the sequence of events according to their relevant intervals,and then specifies network states in these Intervals.The uncertain intrusion signatures as well as basic temporal modes of events,which consist of the parallel mode,the sequential mode and the hybrid mode,can be succinctly and naturally represented in ISDTM.Mode chart is the visualization of intrusion signatures in ISDTM,which makes the formulas more readable.The intrusion signatures descriptions in ISDTM have advantages of compact construct,concise syntax,scalability and easy implementation.
文摘Root-cause identification plays a vital role in business decision making by providing effective future directions for the organizations.Aspect extraction and sentiment extraction plays a vital role in identifying the rootcauses.This paper proposes the Ensemble based temporal weighting and pareto ranking(ETP)model for Root-cause identification.Aspect extraction is performed based on rules and is followed by opinion identification using the proposed boosted ensemble model.The obtained aspects are validated and ranked using the proposed aspect weighing scheme.Pareto-rule based aspect selection is performed as the final selection mechanism and the results are presented for business decision making.Experiments were performed with the standard five product benchmark dataset.Performances on all five product reviews indicate the effective performance of the proposed model.Comparisons are performed using three standard state-of-the-art models and effectiveness is measured in terms of F-Measure and Detection rates.The results indicate improved performances exhibited by the proposed model with an increase in F-Measure levels at 1%–15%and detection rates at 4%–24%compared to the state-of-the-art models.
基金supported by the National 973Program of China(2013CB733302)the National Natural Science Foundation of China(41131067,41174020,41374023,41474019)+2 种基金the Open Research Fund Program of the State Key Laboratory of Geodesy and Earth's Dynamics(SKLGED2015-1-3-E)the open fund of State Key Laboratory of Geographic Information Engineering(SKLGIE2013-M-1-3)the open fund of Key Laboratory of Geospace Environment and Geodesy,Ministry of Education(13-02-05)
文摘A new temporal gravity field model called WHU-Grace01s solely recovered from Gravity Recovery and Climate Experiment （GRACE） K-Band Range Rate （KBRR） data based on dynamic integral approach is presented in this paper. After meticulously preprocessing of the GRACE KBRR data, the root mean square of its post residuals is about 0.2 micrometers per second, and seventy-two monthly temporal solutions truncated to degree and order 60 are computed for the period from January 2003 to December 2008. After applying the combi- nation filter in WHU-Grace01s, the global temporal signals show obvious periodical change rules in the large-scale fiver basins. In terms of the degree variance, our solution is smaller at high degrees, and shows a good consistency at the rest of degrees with the Release 05 models from Center for Space Research （CSR）, GeoForschungsZentrum Potsdam （GFZ） and Jet Pro- pulsion Laboratory 0PL）. Compared with other published models in terms of equivalent water height distribution, our solution is consistent with those published by CSR, GFZ, JPL, Delft institute of Earth Observation and Space system （DEOS）, Tongji University （Tongji）, Institute of Theoretical Geodesy （ITG）, Astronomical Institute in University of Bern （AIUB） and Groupe de Recherche de Geodesie Spatiale （GRGS}, which indicates that the accuracy of WHU-Grace01s has a good consistency with the previously published GRACE solutions.
文摘Time is an important dimension for information in the geographical information system. Data, such as the historical state of target property space and related events causing the state to be changed, should be saved as important files. This should be applied to property management. This paper designs and constructs a spatial temporal model, which is suitable to the property data changing management and spatial temporal query by analyzing the basic types and characteristics of property management spatial changing time and date. This model uses current and historical situational layers to organize and set up the relationship between current situation data and historical dates according to spatial temporal topological relations in property entities. By using Map Basic, housing property management and spatial query is realized.
基金Supported by the National Natural Science Foundation of China (No. 60532030).
文摘A high altitude platform station （HAPS） based communications scenario for providing Intemet access and broadband multimedia services to the passengers on board of a high speed train （traveling up to 300km/h） is proposed. Regarding the addressed scenario, when the propagation link between HAPS and train is blocked by obstacles, a three-dimensional （3-D） geometrical single cylinder spatial-temporal channel model is presented, in which closed form, mathematically tractable space-time correlation functions are obtained. It shows that the correlation functions determined by the 3-D model are of significant difference with those of the conventional 2-D model. Based on the analysis model, the paper derives a realized simulation model using sum-of-sinusoids approach, and applies method of equal areas （MEA） and modified method of equal areas （MMEA） to determine the model parameters. The fitting performance of the simulation model with the analysis one is evaluated by two means-square error （MSE） performance criteria. Finally, numerical simulation results verify the mathematical analysis conclusion, when N ≥21, simulation model has an excellent fitness with the analysis one.
文摘As there is datum redundancy in tradition database and temporal database in existence and the quantities of temporal database are increasing fleetly. We put forward compress storage tactics for temporal datum which combine compress technology in existence in order to settle datum redundancy in the course of temporal datum storage and temporal datum of slow acting domain and momentary acting domain are accessed by using each from independence clock method and mutual clock method .We also bring forward strategy of gridding storage to resolve the problems of temporal datum rising rapidly.
基金Supported by the National Natural Science Foundation of China(62072334).
文摘The hands and face are the most important parts for expressing sign language morphemes in sign language videos.However,we find that existing Continuous Sign Language Recognition(CSLR)methods lack the mining of hand and face information in visual backbones or use expensive and time-consuming external extractors to explore this information.In addition,the signs have different lengths,whereas previous CSLR methods typically use a fixed-length window to segment the video to capture sequential features and then perform global temporal modeling,which disturbs the perception of complete signs.In this study,we propose a Multi-Scale Context-Aware network(MSCA-Net)to solve the aforementioned problems.Our MSCA-Net contains two main modules:(1)Multi-Scale Motion Attention(MSMA),which uses the differences among frames to perceive information of the hands and face in multiple spatial scales,replacing the heavy feature extractors;and(2)Multi-Scale Temporal Modeling(MSTM),which explores crucial temporal information in the sign language video from different temporal scales.We conduct extensive experiments using three widely used sign language datasets,i.e.,RWTH-PHOENIX-Weather-2014,RWTH-PHOENIX-Weather-2014T,and CSL-Daily.The proposed MSCA-Net achieve state-of-the-art performance,demonstrating the effectiveness of our approach.
基金Project supported by the National Natural Science Foundation of China(Nos.62476015,62171298,and 62476181。
文摘Temporal attention mechanisms are essential for video action recognition,enabling models to focus on semantically informative moments.However,these models frequently exhibit temporal infidelitymisaligned attention weights caused by limited training diversity and the absence of fine-grained temporal supervision.While video-level labels provide coarse-grained action guidance,the lack of detailed constraints allows attention noise to persist,especially in complex scenarios with distracting spatial elements.To address this issue,we propose temporal fidelity enhancement(TFE),a competitive learning paradigm based on the disentangled information bottleneck(DisenIB)theory.TFE mitigates temporal infidelity by decoupling action-relevant semantics from spurious correlations through adversarial feature disentanglement.Using pre-trained representations for initialization,TFE establishes an adversarial process in which segments with elevated temporal attention compete against contexts with diminished action relevance.This mechanism ensures temporal consistency and enhances the fidelity of attention patterns without requiring explicit fine-grained supervision.Extensive studies on UCF101,HMDB-51,and Charades benchmarks validate the effectiveness of our method,with significant improvements in action recognition accuracy.
基金supported by the National Natural Science Foundation of China under Grant Nos.62272093,62137001,and 62372097.
文摘Deep learning has found widespread application across diverse domains owing to its exceptional performance.Nevertheless,the lack of transparency in deep learning models’decision-making processes undermines their usability,especially in critical contexts.While researchers have made noteworthy advancements in explaining these models,they have frequently overlooked the differences between static and temporal models during explanation generation.In temporal models,features change over time,posing new challenges in the generation of explanations.Though extensive research has been dedicated to surmounting these hurdles,a survey summarizing these contributions is currently absent.To bridge this gap,this paper endeavors to summarize existing methods and their contributions in terms of both static and temporal models,highlighting their disparities.Additionally,we propose an innovative classification approach based on the comprehensibility of explanations,demonstrating that different explanation methods vary in their understandability for users.Finally,to assess the limitations of the explanation capabilities of existing methods,we specifically choose knowledge tracing to analyze the evolution of explanation methods in this context of temporal modeling and interpretations.
基金supported by NSFC No.62372430the Youth Innovation Promotion As-sociation CAS No.2023112.
文摘Intelligent spatial-temporal data analysis,leveraging data such as multivariate time series and geographic information,provides researchers with powerful tools to uncover multiscale patterns and enhance decision-making processes.As artificial intelligence advances,intelligent spatial-temporal algorithms have found extensive applications across various disciplines,such as geosciences,biology,and public health.1 Compared to traditional methods,these algorithms are data driven,making them well suited for addressing the complexities of modeling real-world systems.However,their reliance on substantial domain-specific expertise limits their broader applicability.Recently,significant advancements have been made in spatial-temporal large models.Trained on large-scale data,these models exhibit a vast parameter scale,superior generalization capabilities,and multitasking advantages over previous methods.Their high versatility and scalability position them as promising super hubs for multidisciplinary research,integrating knowledge,intelligent algorithms,and research communities from different fields.Nevertheless,achieving this vision will require overcoming numerous critical challenges,offering an expansive and profound space for future exploration.
基金Project supported by the National Natural Science Foundation of China(Grant Nos.51206033 and 51276046)the Specialized Research Fund for the Doctoral Program of Higher Education of China(Grant No.20112302110020)+2 种基金the China Postdoctoral Science Foundation(Grant No.2011M500652)the Heilongjiang Postdoctoral Science Foundation,China(Grant No.2011LBH-Z11139)the Natural Scientific Research Innovation Foundation in Harbin Institute of Technology,China(Grant No.HIT.NSRIF.2012070)
文摘Large-eddy simulations （LES） based on the temporal approximate deconvolution model were performed for a forced homogeneous isotropic turbulence （FHIT） with polymer additives at moderate Taylor Reynolds number. Finitely extensible nonlinear elastic in the Peterlin approximation model was adopted as the constitutive equation for the filtered conformation tensor of the polymer molecules. The LES results were verified through comparisons with the direct numerical simulation results. Using the LES database of the FHIT in the Newtonian fluid and the polymer solution flows, the polymer effects on some important parameters such as strain, vorticity, drag reduction, and so forth were studied. By extracting the vortex structures and exploring the flatness factor through a high-order correlation function of velocity derivative and wavelet analysis, it can be found that the small-scale vortex structures and small-scale intermittency in the FHIT are all inhibited due to the existence of the polymers. The extended self-similarity scaling law in the polymer solution flow shows no apparent difference from that in the Newtonian fluid flow at the currently simulated ranges of Reynolds and Weissenberg numbers.
基金funded by the Major National Scientific Research Plan(2013CB733305,2012CB957703)the National Natural Science Foundation of China(41174066,41131067,41374087,41431070)
文摘The Gravity Recovery and Climate Experiment（GRACE） mission can significantly improve our knowledge of the temporal variability of the Earth＇s gravity field.We obtained monthly gravity field solutions based on variational equations approach from GPS-derived positions of GRACE satellites and K-band range-rate measurements.The impact of different fixed data weighting ratios in temporal gravity field recovery while combining the two types of data was investigated for the purpose of deriving the best combined solution.The monthly gravity field solution obtained through above procedures was named as the Institute of Geodesy and Geophysics（IGG） temporal gravity field models.IGG temporal gravity field models were compared with GRACE Release05（RL05） products in following aspects：（i） the trend of the mass anomaly in China and its nearby regions within 2005-2010; （ii） the root mean squares of the global mass anomaly during 2005-2010; （iii） time-series changes in the mean water storage in the region of the Amazon Basin and the Sahara Desert between 2005 and 2010.The results showed that IGG solutions were almost consistent with GRACE RL05 products in above aspects（i）-（iii）.Changes in the annual amplitude of mean water storage in the Amazon Basin were 14.7 ± 1.2 cm for IGG,17.1 ± 1.3 cm for the Centre for Space Research（CSR）,16.4 ± 0.9 cm for the GeoForschungsZentrum（GFZ） and 16.9 ± 1.2 cm for the Jet Propulsion Laboratory（JPL） in terms of equivalent water height（EWH）,respectively.The root mean squares of the mean mass anomaly in Sahara were 1.2 cm,0.9 cm,0.9 cm and 1.2 cm for temporal gravity field models of IGG,CSR,GFZ and JPL,respectively.Comparison suggested that IGG temporal gravity field solutions were at the same accuracy level with the latest temporal gravity field solutions published by CSR,GFZ and JPL.
基金Supported by the Knowledge Innovation Program of Chinese Academy of Sciences and National Key Technologies R &D Program in the 10th Five-Ycar Plan of china(2001BA901A40)
文摘This paper proposes a mechanism theory on regional development by using a modified Logistic model. It reveals regional evolution is an integration of fluctuation in temporal dimension and disparity in spatial dimension. T = S model is established by using Logistic model to simulate the growth of per capita GDP in China from 1990 to 1999. The result shows that T=S model accurately simulates the tracks of economic growth.
