IndRT-GCNets: Knowledge Reasoning with Independent Recurrent Temporal Graph Convolutional Representations

Due to the structural dependencies among concurrent events in the knowledge graph and the substantial amount of sequential correlation information carried by temporally adjacent events,we propose an Independent Recurrent Temporal Graph Convolution Networks(IndRT-GCNets)framework to efficiently and accurately capture event attribute information.The framework models the knowledge graph sequences to learn the evolutionary represen-tations of entities and relations within each period.Firstly,by utilizing the temporal graph convolution module in the evolutionary representation unit,the framework captures the structural dependency relationships within the knowledge graph in each period.Meanwhile,to achieve better event representation and establish effective correlations,an independent recurrent neural network is employed to implement auto-regressive modeling.Furthermore,static attributes of entities in the entity-relation events are constrained andmerged using a static graph constraint to obtain optimal entity representations.Finally,the evolution of entity and relation representations is utilized to predict events in the next subsequent step.On multiple real-world datasets such as Freebase13(FB13),Freebase 15k(FB15K),WordNet11(WN11),WordNet18(WN18),FB15K-237,WN18RR,YAGO3-10,and Nell-995,the results of multiple evaluation indicators show that our proposed IndRT-GCNets framework outperforms most existing models on knowledge reasoning tasks,which validates the effectiveness and robustness.

Spatial-temporal Evolution of the Urban-rural Coordination Relationship in Northeast China in 1990-2018

To comprehensively understand the law of urban-rural relationship and propose scientific measures of urban-rural coordinated development in Northeast China,this study uses the coupling coordination degree model and geographically and temporally weighted regression(GTWR)model to analyze the spatial-temporal patterns and the corresponding driving mechanisms of its urban-rural coordination since 1990.The results are as follows.First,the urban-rural coupling coordination degree in Northeast China was very low and improved slowly,but its stages of evolution is a good interpretation of the strategic arrangements of China's urbanization.Second,the urban-rural coupling coordination degree in Northeast China had spatial differences and was characterized by central polarization,converging on urban agglomeration,which was high in the south and low in the north.Moreover,the gap between the north and south weakened.Third,the spatial-temporal evolution of the urban-rural coordination relationship in Northeast China was influenced by pulling from the central cities,pushing from rural transformation,and government regulations.The influence intensity of the three mechanisms was weak,but the pulling from the central cities was stronger than that of the other two mechanisms.Furthermore,the spatial difference between the three mechanisms determines the spatial pattern and its evolution of the urban-rural coordination relationship in Northeast China.Fourth,to promote the development of urban-rural coordination in Northeast China,it is essential to advance urban-rural economic correlation,enhance the government^role in regulating and guiding,and adopt different policies for each region in Northeast China.

Spatio-temporal Pattern Characteristics of Relationship Between Urbanization and Economic Development at County Level in China

The relationship between China’s urbanization and economic development(RCUED) is an important concern nationwide. As important actors in regional strategy and policy, county-level regions have played an increasingly significant role in the development of China’s social economy. However, the existing research on the RCUED lacks the fine depiction of the county-level administrative units.Using 2000 and 2010 census data and the statistical analysis method, we uncovered the evolution characteristics of China’s urbanization and economic development and conducted a quantitative identification for the RCUED with improved methods using the quadrant map approach. In addition, we investigated the spatial correlation effect of the RCUED using the spatial autocorrelation analysis method. The results were as follows: 1) In general, a high degree of matching exists between China’s urbanization and economic development at the county level at the significance level of 0.01. The correlation coefficients between China’s urbanization and economic development in2000 and 2010 were 0.608 and 0.603, respectively. 2) A significant regional difference exists in the RCUED at the county level. Based on a comparative analysis of 2276 county units in China in the two years, we found that county units can be categorized as under-urbanized, basic coordination and over-urbanized in various areas. No situation was observed where urbanization seriously lagged behind the economic development level, so the levels of urbanization and economic development appear to be basically coordinated,and the coordination state may be gradually optimized over time. 3) Over time, the spatial dependency of the RCUED has weakened and the spatial heterogeneity has increased. Northeast China has always been an area characterized by over-urbanization. The number of county units classified as under-urbanized has begun to decline in eastern coastal urban agglomeration areas, while counties rich in resources have transformed from having point-shaped over-urbanization to plane-shaped under-urbanization along the northern border,and the number of over-urbanized county units has increased in the middle reaches of the Yangtze River. 4)’Lag-lag’ type and ’advance-advance’ type accounted for 68% of all counties in China, and these counties were shown to have obvious spatial differentiation characteristics.

Research on the Optimal Vegetation Cover for Remote Sensing Assessment of Soil Erosion Risk Using the Temporal Matching Relationship between Rainfall and Vegetation

Vegetation cover derived from remote sensing image is widely used for soil erosion risk assessment, but there is no clear guideline to select the most appropriate temporal satellite data. It is common practice that satellite data during growing season are randomly selected and used in soil erosion risk assessment. However, the effectiveness of vegetation in protecting the soil is quite different even if it is the same growing season since vegetation covers change as they grow. This article aims to provide a method of choosing optimal vegetation cover for studying soil erosion risk using remote sensing, that is, the vegetation cover in the most appropriate temporal period. Based on the temporal relationship of the two most active impact factors, rainfall and vegetation, an index of RV is developed and used to indicate the relative erosion risk during the year. The results show that annual variation of rainfall is significant, and vegetation is relatively stable, resulting in their matching relationship is different in each year. The correlation coefficient reaches 0.89 between RV and real sediment transport during the period when rainfall can cause soil erosion. In other words, RV is a good indicator of soil erosion. Therefore, there is a good correlation between RV maximum and the optimal vegetation cover, which can help facilitate erosion research in the future, showing good potential for successful application in other places.

On Exploiting Temporal, Social, and Geographical Relationships for Data Forwarding in Delay Tolerant Networks

Because of unpredictable node mobility and absence of global information in Delay Tolerant Networks (DTNs), effective data forwarding has become a significant challenge in such network. Currently, most of existing data forwarding mechanisms select nodes with high cumulative contact capability as forwarders. However, for the heterogeneity of the transient node contact patterns, these selection approaches may not be the best relay choices within a short time period. This paper proposes an appropriate data forwarding mechanism, which combines time, location, and social characteristics into one coordinate system, to improve the performance of data forwarding in DTNs. The Temporal-Social Relationship and the Temporal-Geographical Relationship reveal the implied connection information among these three factors. This mechanism is formulated and verified in the experimental studies of realistic DTN traces. The empirical results show that our proposed mechanism can achieve better performance compared to the existing schemes with similar forwarding costs (e.g. end-to-end delay and delivery success ratio).

Two new approaches for image registration based onspatial-temporal relationship

How to improve the probability of registration and precision of localization is a hard problem, which is desiderated to solve. The two basic approaches (normalized cross-correlation and phase correlation) for image registration are analysed, two improved approaches based on spatial-temporal relationship are presented. This method adds the correlation matrix according to the displacements in x- cirection and y- directions, and the registration pose is searched in the added matrix. The method overcomes the shortcoming that the probability of registration decreasing with area increasing owing to geometric distortion, improves the probability and the robustness of registration.

A Lightweight Temporal Convolutional Network for Human Motion Prediction

A lightweight multi-layer residual temporal convolutional network model(RTCN)is proposed to target the highly complex kinematics and temporal correlation of human motion.RTCN uses 1-D convolution to efficiently obtain the spatial structure information of human motion and extract the correlation in the time series of human motion.The residual structure is applied to the proposed network model to alleviate the problem of gradient disappearance in the deep network.Experiments on the Human 3.6M dataset demonstrate that the proposed method effectively reduces the errors of motion prediction compared with previous methods,especially of long-term prediction.

Action Recognition Using Multi-Scale Temporal Shift Module and Temporal Feature Difference Extraction Based on 2D CNN

Convolutional neural networks, which have achieved outstanding performance in image recognition, have been extensively applied to action recognition. The mainstream approaches to video understanding can be categorized into two-dimensional and three-dimensional convolutional neural networks. Although three-dimensional convolutional filters can learn the temporal correlation between different frames by extracting the features of multiple frames simultaneously, it results in an explosive number of parameters and calculation cost. Methods based on two-dimensional convolutional neural networks use fewer parameters;they often incorporate optical flow to compensate for their inability to learn temporal relationships. However, calculating the corresponding optical flow results in additional calculation cost;further, it necessitates the use of another model to learn the features of optical flow. We proposed an action recognition framework based on the two-dimensional convolutional neural network;therefore, it was necessary to resolve the lack of temporal relationships. To expand the temporal receptive field, we proposed a multi-scale temporal shift module, which was then combined with a temporal feature difference extraction module to extract the difference between the features of different frames. Finally, the model was compressed to make it more compact. We evaluated our method on two major action recognition benchmarks: the HMDB51 and UCF-101 datasets. Before compression, the proposed method achieved an accuracy of 72.83% on the HMDB51 dataset and 96.25% on the UCF-101 dataset. Following compression, the accuracy was still impressive, at 95.57% and 72.19% on each dataset. The final model was more compact than most related works.

SPATIO-TEMPORAL VARIATION OF ACTUAL EVAPOTRANSPIRATION AND ITS RELATION WITH CLIMATE PARAMETERS IN THE PEARL RIVER BASIN,CHINA

Spatio-temporal variation of actual evapotranspiration(ETa) in the Pearl River basin from 1961 to 2010 are analyzed based on daily data from 60 national observed stations. ETa is calculated by the Advection-Aridity model(AA model) in the current study, and Mann-Kendall test(MK) and Inverse Distance Weighted interpolation method(IDW)were applied to detect the trends and spatial variation pattern. The relations of ETa with climate parameters and radiation/dynamic terms are analyzed by Person correlation method. Our findings are shown as follows: 1) Mean annual ETa in the Pearl River basin is about 665.6 mm/a. It has significantly decreased in 1961-2010 at a rate of-24.3mm/10 a. Seasonally, negative trends of summer and autumn ETa are higher than that of spring and winter. 2) The value of ETa is higher in the southeast coastal area than in the northwest region of the Pearl River basin, while the latter has shown the strongest negative trend. 3) Negative trends of ETa in the Pearl River basin are most probably due to decreasing radiation term and increasing dynamic term. The decrease of the radiation term is related with declining diurnal temperature range and sunshine duration, and rising atmospheric pressure as well. The contribution of dynamic term comes from increasing average temperature, maximum and minimum temperatures in the basin. Meanwhile, the decreasing average wind speed weakens dynamic term and finally, to a certain extent, it slows down the negative trend of the ETa.

Surface Humidity Changes in Different Temporal Scales

As the key driven factor of hydrological cycles and global energy transfer processes, water vapour in the atmosphere is important for observing and understanding climatic system changes. In this study, we utilized the multi-dimensional Kolmogorov-Zurbenko filter (KZ filter) to assimilate a near-global high-resolution (monthly 1°?× 1°?grid) humidity climate observation database that provided consistent humidity estimates from 1973 onwards;then we examined the global humidity movements based on different temporal scales that separated out according to the average spectral features of specific humidity data. Humidity climate components were restored with KZ filters to represent the long-term trends and El Nino-like interannual movements. Movies of thermal maps based on these two climate components were used to visualize the water vapour fluctuation patterns over the Earth. Current results suggest that increases in water vapour are found over a large part of the oceans and the land of Eurasia, and the most confirmed increasing pattern is over the south part of North Atlantic and around the India subcontinent;meanwhile, the surface moisture levels over lands of south hemisphere are becoming less.

东北地区城镇化与资源环境承载力耦合关系的时空变化

理清城镇化与资源环境承载力耦合关系可为城市可持续发展转型提供决策参考。在分析2003—2019年东北地区城镇化与资源环境承载力时空格局的基础上,采用耦合协调度模型定量剖析了二者耦合关系的时空变化特征。结果显示:2003—2019年,东北地区综合城镇化水平提升了114.29%,总体呈现“东高西低”、辽中南城市群和哈长城市群核心城市引领全域发展的局面;资源环境承载力呈增加趋势,增幅36.09%,空间上聚集于哈大线周边区域;资源环境承载力水平与综合城镇化水平的耦合协调度总体上呈上升趋势,近年增幅有所下降,且辽西北、吉林南部和黑龙江西北部耦合协调度相对较低,核心城市和重要城市耦合协调度普遍较高。亟须提升核心城市对中小城镇的辐射引领能力,并加强生态退化地区的环境保护和高质量发展。

浅析汉语和韩国语表“瞬时”关系时间连接标记的句法位置和句法限制

“瞬时”关系指两个事件之间的时间间隔非常短,即后一事件紧接着前一事件进行、发生或完成。汉语和韩国语表示两个事件之间的“瞬时”关系时,汉语主要依靠词汇手段,而韩国语主要依靠语法手段,而韩国语的语法标记是强制性手段。旨在为汉韩、韩汉双语教学及翻译实践提供事实依据和参考,该研究考察汉语和韩国语表示“瞬时”关系时间连接标记的句法位置和句法限制,即探讨主语的位置及主语一致性问题。

多维城市化与乡村地域功能演化的时空关联研究:以江汉平原为例

探索城市化与乡村地域功能演化的复杂性关联规律既是新时期城乡融合与乡村振兴的重要命题,也是乡村地理学研究的时代前沿。以江汉平原为研究案例,以2000—2019年为研究时段,在建立城市化指标体系和乡村地域功能评价模型的基础上,运用系统动态面板模型、时空地理加权回归模型等方法,深入探究多维城市化与乡村地域功能之间的时序关联规律及空间关联特征。结果表明:(1)江汉平原乡村地域功能演化具有明显结构分异性,其中,乡村地域综合功能在时序上呈现阶段性变化,空间上呈现沿焦柳-长荆发展轴的倒“U”形分布格局向汉宜线-东部平原带的“T”形分布格局演化。(2)江汉平原多维城市化与乡村地域功能在时序关联上呈现复杂性特征,其中人口、经济城市化与乡村地域综合功能呈现“N”形曲线规律,土地城市化与乡村地域综合功能呈现单调递增态势,但弹性系数值低于后期人口和经济城市化正效应值。(3)江汉平原多维城市化与乡村地域功能在空间关联上呈现明显规律性和异质性特征。研究结果有利于丰富乡村地理学理论内涵,为案例地区的城乡融合与乡村振兴实践提供支持。

基于互补相关理论的青藏高原蒸散发时空变化及其影响因素

蒸散发在陆地水热循环过程中起着重要的纽带作用,青藏高原(QTP)因独特的地理环境,其蒸散发研究受到极大的关注。基于0.05°分辨率的气象格网数据集和MODIS遥感数据,利用互补相关理论估算了2002—2019年青藏高原逐日蒸散发,探讨了蒸散发的时空变化特征,及其与冻土类型、气象要素的关系。结果表明:(1)互补相关理论在青藏高原地区估算的蒸散发结果是可靠的。(2)青藏高原2002—2019年平均蒸散发量约371.54 mm/a,具有明显的季节性差异,主要集中在夏季;蒸散发在青藏高原呈现出东南向西北递减的空间分布特征,趋势变化上在东部呈增加趋势,而西部呈减少趋势;蒸散发稳定性的空间差异显著,波动强度表现出从东南向西北递减的趋势。(3)蒸散发量与冻土类型空间分布密切相关,冻土水热状态越稳定,蒸散发量越小。(4)气象要素对蒸散发时空变化的影响具有明显的差异性,蒸散发对净辐射敏感性最高,风速的相对变化对蒸散发变化的贡献最大。本文结果加强了气候变化背景下青藏高原水循环的响应机制的研究,为区域生态环境保护措施的制定提供了依据。

城市路内违章停车时空特征及影响因素作用模型

为揭示城市路内违章停车的时空模式及影响机理,结合持续两个月共10396条电子警察违停抓拍数据,从土地利用、停车供给与道路设计这3个维度选取16个影响因素进行分析。利用标准椭圆方法、层次聚类法及皮尔逊相关性检验等方法,分析路内违停时空特征并进行影响因素筛选,进一步根据因变量特征构建贝叶斯分位数回归模型,解析早晚高峰违停数量与影响因素的关系。研究结果表明:从违停数量的时间特征来看,工作日违停数量明显高于周末违停数量,且主要集中在周中,工作日违停数量按小时聚类结果显示,违停数量分为3个时间模式,分别为早晚高峰时间段(8:00-10:00和17:00-19:00)、早晚高峰汇聚时间段(10:00-11:00和15:00-17:00)与其他时间段;从违停数量的空间特征来看,工作日违停数量的整体空间分布主要集中在商业和居住密度较高的区域,其在一天内的空间变化趋势呈现“扩撒-稳定-收缩”的态势;分位数回归结果显示,土地使用、停车供给及道路设计这3个维度因素对路内违停存在非线性影响与阈值效应,且早晚高峰违停产生机制具有差异,路内违停治理需要考虑时空特征进行差异化治理。

Treadmill exercise improves hippocampal neural plasticity and relieves cognitive deficits in a mouse model of epilepsy

Epilepsy frequently leads to cognitive dysfunction and approaches to treatment remain limited.Although regular exercise effectively improves learning and memory functions across multiple neurological diseases,its application in patients with epilepsy remains controversial.Here,we adopted a 14-day treadmill-exercise paradigm in a pilocarpine injection-induced mouse model of epilepsy.Cognitive assays confirmed the improvement of object and spatial memory after endurance training,and electrophysiological studies revealed the maintenance of hippocampal plasticity as a result of physical exercise.Investigations of the mechanisms underlying this effect revealed that exercise protected parvalbumin interneurons,probably via the suppression of neuroinflammation and improved integrity of blood-brain barrier.In summary,this work identified a previously unknown mechanism through which exercise improves cognitive rehabilitation in epilepsy.

基于长短期时间关系网络的视频行人重识别

行人重识别是计算机视觉领域中的一个重要研究方向,其目的是在不同的监控摄像头中识别并跟踪同一行人.由于视频帧间存在多种时间关系,从这些关系中可以获取到对象的运动模式以及细粒度特征,因此视频重识别相比图像重识别拥有更丰富的时空线索,也更接近实际应用.问题的关键是如何挖掘这些时空线索作为视频重识别的特征.本文针对视频行人重识别问题,提出了一种基于Transformer的长短期时间关系网络(Long and Short Time Transformer,LSTT).该网络包含长短期时间关系模块,提取重要时序信息并强化特征表示.长期时间关系模块利用记忆线索存储每帧信息,并在每一帧建立全局联系;短期时间关系模块则考虑相邻帧之间交互,学习细粒度目标信息,提高特征表示能力.此外,为了提高模型对不同目标特征的适配性,本文还设计了一个包含不同规格卷积核的多尺度模块.该模块具有多种卷积感受野,能够更全面覆盖目标区域,从而进一步提高模型的泛化性能.在MARS、MARS_DL和iLIDS-VID 3个数据集上的实验结果表明,LSTT模型性能最优.

长江上游凯江流域径流侵蚀功率时空分布与输沙关系

水土流失是长江上游面临的首要生态环境问题,准确把握区域水土流失的规律及其发展趋势是关键。基于SWAT模型和径流侵蚀功率理论,分析了该区域典型流域凯江的径流侵蚀功率时空分布特征及其与输沙的关系。结果表明:径流模拟结果在率定期和验证期基本满足R^(2)>0.85,NSE>0.75,PBIAS<±15%,评价指标均达可信以上程度。2009—2018年径流侵蚀功率年际变化大,在多年尺度上干支流表现出“上游大、下游小”的空间分布特征。另外,多种水沙关系对比分析表明,输沙模数与径流侵蚀功率主要呈幂函数关系。研究成果可为科学治理水土流失和准确预测输沙提供理论依据。

基于时空注意力的空间关联三维形貌重建

聚焦形貌恢复通过对场景深度和散焦模糊之间的潜在关系进行建模实现三维形貌重建。但现有的三维形貌重建网络无法有效利用图像序列的时序关联进行表征学习,因此,提出一种基于多景深图像序列空间关联特征的深度网络框架——三维空间相关水平分析模型(3D SCHAM)进行三维形貌重建。该模型不仅可以精确捕获单帧图像中聚焦区域到离焦区域的边缘特征,而且可有效利用不同图像帧之间的空间依赖性特征。首先,通过构建深度、宽度和感受野复合扩展的网络构造三维形貌重建的时域连续模型,进而确定单点深度结果;其次,引入基于空间关联的注意力模块,充分学习帧与帧间的“邻接性”与“距离性”空间依赖关系;另外,利用残差反转瓶颈进行重采样,以保持跨尺度的语义丰富性。在DDFF 12-Scene真实场景数据集上的实验结果显示,相较于DfFintheWild模型,3D SCHAM在深度值准确度度量的3个阈值1.25,1.25^(2),1.25^(3)上的精确度分别提升了15.34%、3.62%、0.86%,验证了该模型在真实场景的鲁棒性。