Early screening of diabetes retinopathy(DR)plays an important role in preventing irreversible blindness.Existing research has failed to fully explore effective DR lesion information in fundus maps.Besides,traditional ...Early screening of diabetes retinopathy(DR)plays an important role in preventing irreversible blindness.Existing research has failed to fully explore effective DR lesion information in fundus maps.Besides,traditional attention schemes have not considered the impact of lesion type differences on grading,resulting in unreasonable extraction of important lesion features.Therefore,this paper proposes a DR diagnosis scheme that integrates a multi-level patch attention generator(MPAG)and a lesion localization module(LLM).Firstly,MPAGis used to predict patches of different sizes and generate a weighted attention map based on the prediction score and the types of lesions contained in the patches,fully considering the impact of lesion type differences on grading,solving the problem that the attention maps of lesions cannot be further refined and then adapted to the final DR diagnosis task.Secondly,the LLM generates a global attention map based on localization.Finally,the weighted attention map and global attention map are weighted with the fundus map to fully explore effective DR lesion information and increase the attention of the classification network to lesion details.This paper demonstrates the effectiveness of the proposed method through extensive experiments on the public DDR dataset,obtaining an accuracy of 0.8064.展开更多
在环境复杂的柔性制造系统中进行合理地AMR(autonomous mobile robot)调度具有重要意义。针对AMR调度中的路径规划与任务分配,以最小化AMR完成任务时间为目标函数建立数学模型,并通过拓扑法进行地图建模;采用贪婪算法对间隔时间服从定...在环境复杂的柔性制造系统中进行合理地AMR(autonomous mobile robot)调度具有重要意义。针对AMR调度中的路径规划与任务分配,以最小化AMR完成任务时间为目标函数建立数学模型,并通过拓扑法进行地图建模;采用贪婪算法对间隔时间服从定长分布的订单进行任务分配,通过对AMR工作状态分类以减少算法运算量;基于Dijkstra算法进行全局路径规划,搜索AMR的全局最短路径,并通过AMR的激光雷达进行局部避障路径规划;最后,通过在openTCS平台进行调度仿真实验验证其有效性。展开更多
We report a design and implementation of a field-programmable-gate-arrays(FPGA)based hardware platform,which is used to realize control and signal readout of trapped-ion-based multi-level quantum systems.This platform...We report a design and implementation of a field-programmable-gate-arrays(FPGA)based hardware platform,which is used to realize control and signal readout of trapped-ion-based multi-level quantum systems.This platform integrates a four-channel 2.8 Gsps@14 bits arbitrary waveform generator,a 16-channel 1 Gsps@14 bits direct-digital-synthesisbased radio-frequency generator,a 16-channel 8 ns resolution pulse generator,a 10-channel 16 bits digital-to-analogconverter module,and a 2-channel proportion integration differentiation controller.The hardware platform can be applied in the trapped-ion-based multi-level quantum systems,enabling quantum control of multi-level quantum system and highdimensional quantum simulation.The platform is scalable and more channels for control and signal readout can be implemented by utilizing more parallel duplications of the hardware.The hardware platform also has a bright future to be applied in scaled trapped-ion-based quantum systems.展开更多
With the construction of the power Internet of Things(IoT),communication between smart devices in urban distribution networks has been gradually moving towards high speed,high compatibility,and low latency,which provi...With the construction of the power Internet of Things(IoT),communication between smart devices in urban distribution networks has been gradually moving towards high speed,high compatibility,and low latency,which provides reliable support for reconfiguration optimization in urban distribution networks.Thus,this study proposed a deep reinforcement learning based multi-level dynamic reconfiguration method for urban distribution networks in a cloud-edge collaboration architecture to obtain a real-time optimal multi-level dynamic reconfiguration solution.First,the multi-level dynamic reconfiguration method was discussed,which included feeder-,transformer-,and substation-levels.Subsequently,the multi-agent system was combined with the cloud-edge collaboration architecture to build a deep reinforcement learning model for multi-level dynamic reconfiguration in an urban distribution network.The cloud-edge collaboration architecture can effectively support the multi-agent system to conduct“centralized training and decentralized execution”operation modes and improve the learning efficiency of the model.Thereafter,for a multi-agent system,this study adopted a combination of offline and online learning to endow the model with the ability to realize automatic optimization and updation of the strategy.In the offline learning phase,a Q-learning-based multi-agent conservative Q-learning(MACQL)algorithm was proposed to stabilize the learning results and reduce the risk of the next online learning phase.In the online learning phase,a multi-agent deep deterministic policy gradient(MADDPG)algorithm based on policy gradients was proposed to explore the action space and update the experience pool.Finally,the effectiveness of the proposed method was verified through a simulation analysis of a real-world 445-node system.展开更多
Massive rural-to-urban migration in China is consequential for political trust: rural-to-urban migrants have been found to hold lower levels of trust in local government than their rural peers who choose to stay in th...Massive rural-to-urban migration in China is consequential for political trust: rural-to-urban migrants have been found to hold lower levels of trust in local government than their rural peers who choose to stay in the countryside (mean 4.92 and 6.34 out of 10, respectively, p < 0.001). This article explores why migrants have a certain level of political trust in their county-level government. Using data of rural-to-urban migrants from the China Family Panel Survey, this study performs a hierarchical linear modeling (HLM) to unpack the multi-level explanatory factors of rural-to-urban migrants’ political trust. Findings show that the individual-level socio-economic characteristics and perceptions of government performance (Level-1), the neighborhood-level characteristics-the physical and social status and environment of neighborhoods (Level-2), and the objective macroeconomic performance of county-level government (Level-3), work together to explain migrants’ trust levels. These results suggest that considering the effects of neighborhood-level factors on rural-to-urban migrants’ political trust merits policy and public management attention in rapidly urbanizing countries.展开更多
自由导航移动平台(自主移动机器人(Autonomous Mobile Robots,简称"AMRs"))可以在不停机的情况下避开障碍物或人员。因此,ISO3691-4所要求的安全功能非常复杂。特别是在转弯时,必须能够在多个保护区域之间切换,以确保防止人...自由导航移动平台(自主移动机器人(Autonomous Mobile Robots,简称"AMRs"))可以在不停机的情况下避开障碍物或人员。因此,ISO3691-4所要求的安全功能非常复杂。特别是在转弯时,必须能够在多个保护区域之间切换,以确保防止人与机器之间的碰撞,且避免不必要的停机。展开更多
近年来,国际上对于强震前的加速矩释放(AMR)现象是否可作为一种可靠的、带有普遍性的地震前兆现象争议较大.本文以2008年3月21日新疆于田 M_s 7.3地震为例,试图从前兆存在的客观性和与地震发生的物理相关性两方面考察本次地震前的 AMR ...近年来,国际上对于强震前的加速矩释放(AMR)现象是否可作为一种可靠的、带有普遍性的地震前兆现象争议较大.本文以2008年3月21日新疆于田 M_s 7.3地震为例,试图从前兆存在的客观性和与地震发生的物理相关性两方面考察本次地震前的 AMR 现象.用"破裂时间分析"方程中的幂指数 m 作为描述震前加速矩释放"程度"的参量,在时间-空间-地震序列截止震级组成的三维参数空间(T,R,M_c)内考察 AMR 现象存存的客观性.考虑了多种因素对 m(T,R,M_c)分布图像可能的影响,其中,余震是否删除和 M_c 对计算影响不大,但 M_L 6.0以上"干扰"事件的影响则较大.结果表明,于田地震前的确存在 AMR 现象,但得到的 m(T,R,M_c)分布图像较为复杂,可观测到两个明显的 AMR 集中分布区.此外,在以实际震中为圆心的多个圆形区域内,使用固定时间窗向实际发震时刻滑动逼近,可观测到 m 值逐渐减小,即加速特征逐渐明显的过程.对震前矩释放程度 m 值的时-空扫描结果显示,出现 AMR 现象的空间区域与震中位置似有较好的对应,但其时-空演化图像与滑动时-空窗的选取有关.这表明,本次 M_s 7.3地震前的确存在 AMR 现象,并与其孕震过程在物理上相关.但本文仅是一个震例的研究,无法给出具有统计显著性的结论。此外,用 AMR 来约束地震发生的时间看来是困难的.展开更多
Based on the stochastic AMR model, this paper constructs man-made earthquake catalogues to investigate the property of parameter estimation of the model. Then the stochastic AMR model is applied to the study of severa...Based on the stochastic AMR model, this paper constructs man-made earthquake catalogues to investigate the property of parameter estimation of the model. Then the stochastic AMR model is applied to the study of several strong earthquakes in China and New Zealand. Akaikes AIC criterion is used to discriminate whether an accelerating mode of earthquake activity precedes those events or not. Finally, regional accelerating seismic activity and possible prediction approach for future strong earthquakes are discussed.展开更多
基金supported in part by the Research on the Application of Multimodal Artificial Intelligence in Diagnosis and Treatment of Type 2 Diabetes under Grant No.2020SK50910in part by the Hunan Provincial Natural Science Foundation of China under Grant 2023JJ60020.
文摘Early screening of diabetes retinopathy(DR)plays an important role in preventing irreversible blindness.Existing research has failed to fully explore effective DR lesion information in fundus maps.Besides,traditional attention schemes have not considered the impact of lesion type differences on grading,resulting in unreasonable extraction of important lesion features.Therefore,this paper proposes a DR diagnosis scheme that integrates a multi-level patch attention generator(MPAG)and a lesion localization module(LLM).Firstly,MPAGis used to predict patches of different sizes and generate a weighted attention map based on the prediction score and the types of lesions contained in the patches,fully considering the impact of lesion type differences on grading,solving the problem that the attention maps of lesions cannot be further refined and then adapted to the final DR diagnosis task.Secondly,the LLM generates a global attention map based on localization.Finally,the weighted attention map and global attention map are weighted with the fundus map to fully explore effective DR lesion information and increase the attention of the classification network to lesion details.This paper demonstrates the effectiveness of the proposed method through extensive experiments on the public DDR dataset,obtaining an accuracy of 0.8064.
文摘在环境复杂的柔性制造系统中进行合理地AMR(autonomous mobile robot)调度具有重要意义。针对AMR调度中的路径规划与任务分配,以最小化AMR完成任务时间为目标函数建立数学模型,并通过拓扑法进行地图建模;采用贪婪算法对间隔时间服从定长分布的订单进行任务分配,通过对AMR工作状态分类以减少算法运算量;基于Dijkstra算法进行全局路径规划,搜索AMR的全局最短路径,并通过AMR的激光雷达进行局部避障路径规划;最后,通过在openTCS平台进行调度仿真实验验证其有效性。
基金the Strategic Priority Research Program of CAS(Grant No.XDC07020200)the National Key R&D Program of China(Grants No.2018YFA0306600)+5 种基金the National Natural Science Foundation of China(Grant Nos.11974330 and 92165206)the Chinese Academy of Sciences(Grant No.QYZDY-SSW-SLH004)the Innovation Program for Quantum Science and Technology(Grant Nos.2021ZD0302200 and 2021ZD0301603)the Anhui Initiative in Quantum Information Technologies(Grant No.AHY050000)the Hefei Comprehensive National Science Centerthe Fundamental Research Funds for the Central Universities。
文摘We report a design and implementation of a field-programmable-gate-arrays(FPGA)based hardware platform,which is used to realize control and signal readout of trapped-ion-based multi-level quantum systems.This platform integrates a four-channel 2.8 Gsps@14 bits arbitrary waveform generator,a 16-channel 1 Gsps@14 bits direct-digital-synthesisbased radio-frequency generator,a 16-channel 8 ns resolution pulse generator,a 10-channel 16 bits digital-to-analogconverter module,and a 2-channel proportion integration differentiation controller.The hardware platform can be applied in the trapped-ion-based multi-level quantum systems,enabling quantum control of multi-level quantum system and highdimensional quantum simulation.The platform is scalable and more channels for control and signal readout can be implemented by utilizing more parallel duplications of the hardware.The hardware platform also has a bright future to be applied in scaled trapped-ion-based quantum systems.
基金supported by the National Natural Science Foundation of China under Grant 52077146.
文摘With the construction of the power Internet of Things(IoT),communication between smart devices in urban distribution networks has been gradually moving towards high speed,high compatibility,and low latency,which provides reliable support for reconfiguration optimization in urban distribution networks.Thus,this study proposed a deep reinforcement learning based multi-level dynamic reconfiguration method for urban distribution networks in a cloud-edge collaboration architecture to obtain a real-time optimal multi-level dynamic reconfiguration solution.First,the multi-level dynamic reconfiguration method was discussed,which included feeder-,transformer-,and substation-levels.Subsequently,the multi-agent system was combined with the cloud-edge collaboration architecture to build a deep reinforcement learning model for multi-level dynamic reconfiguration in an urban distribution network.The cloud-edge collaboration architecture can effectively support the multi-agent system to conduct“centralized training and decentralized execution”operation modes and improve the learning efficiency of the model.Thereafter,for a multi-agent system,this study adopted a combination of offline and online learning to endow the model with the ability to realize automatic optimization and updation of the strategy.In the offline learning phase,a Q-learning-based multi-agent conservative Q-learning(MACQL)algorithm was proposed to stabilize the learning results and reduce the risk of the next online learning phase.In the online learning phase,a multi-agent deep deterministic policy gradient(MADDPG)algorithm based on policy gradients was proposed to explore the action space and update the experience pool.Finally,the effectiveness of the proposed method was verified through a simulation analysis of a real-world 445-node system.
文摘Massive rural-to-urban migration in China is consequential for political trust: rural-to-urban migrants have been found to hold lower levels of trust in local government than their rural peers who choose to stay in the countryside (mean 4.92 and 6.34 out of 10, respectively, p < 0.001). This article explores why migrants have a certain level of political trust in their county-level government. Using data of rural-to-urban migrants from the China Family Panel Survey, this study performs a hierarchical linear modeling (HLM) to unpack the multi-level explanatory factors of rural-to-urban migrants’ political trust. Findings show that the individual-level socio-economic characteristics and perceptions of government performance (Level-1), the neighborhood-level characteristics-the physical and social status and environment of neighborhoods (Level-2), and the objective macroeconomic performance of county-level government (Level-3), work together to explain migrants’ trust levels. These results suggest that considering the effects of neighborhood-level factors on rural-to-urban migrants’ political trust merits policy and public management attention in rapidly urbanizing countries.
文摘近年来,国际上对于强震前的加速矩释放(AMR)现象是否可作为一种可靠的、带有普遍性的地震前兆现象争议较大.本文以2008年3月21日新疆于田 M_s 7.3地震为例,试图从前兆存在的客观性和与地震发生的物理相关性两方面考察本次地震前的 AMR 现象.用"破裂时间分析"方程中的幂指数 m 作为描述震前加速矩释放"程度"的参量,在时间-空间-地震序列截止震级组成的三维参数空间(T,R,M_c)内考察 AMR 现象存存的客观性.考虑了多种因素对 m(T,R,M_c)分布图像可能的影响,其中,余震是否删除和 M_c 对计算影响不大,但 M_L 6.0以上"干扰"事件的影响则较大.结果表明,于田地震前的确存在 AMR 现象,但得到的 m(T,R,M_c)分布图像较为复杂,可观测到两个明显的 AMR 集中分布区.此外,在以实际震中为圆心的多个圆形区域内,使用固定时间窗向实际发震时刻滑动逼近,可观测到 m 值逐渐减小,即加速特征逐渐明显的过程.对震前矩释放程度 m 值的时-空扫描结果显示,出现 AMR 现象的空间区域与震中位置似有较好的对应,但其时-空演化图像与滑动时-空窗的选取有关.这表明,本次 M_s 7.3地震前的确存在 AMR 现象,并与其孕震过程在物理上相关.但本文仅是一个震例的研究,无法给出具有统计显著性的结论。此外,用 AMR 来约束地震发生的时间看来是困难的.
基金National Natural Science Foundation of China (4007401340134010)Chinese Joint Seismological Science Foundation (042002) and the project during the Tenth Five-year Plan.
文摘Based on the stochastic AMR model, this paper constructs man-made earthquake catalogues to investigate the property of parameter estimation of the model. Then the stochastic AMR model is applied to the study of several strong earthquakes in China and New Zealand. Akaikes AIC criterion is used to discriminate whether an accelerating mode of earthquake activity precedes those events or not. Finally, regional accelerating seismic activity and possible prediction approach for future strong earthquakes are discussed.