The singularly perturbed Robin boundary value problem for the higher order elliptic equation is considered.Under suitable conditions,the existence and asymptotic behavior of solution to the boundary value problems are...The singularly perturbed Robin boundary value problem for the higher order elliptic equation is considered.Under suitable conditions,the existence and asymptotic behavior of solution to the boundary value problems are studied.The uniform validity of its asymptotic expansion is proved by using the fixed point theorem.展开更多
液体状态机(Liquid State Machine,LSM)具有实时计算和仿生的特点,在处理时间序列数据上具有巨大潜力。为了研究如何提高神经网络模型训练性能,降低计算复杂度,文章首先梳理和回顾了近几年相关研究文献,其次提出硬件实现和软件模型两个...液体状态机(Liquid State Machine,LSM)具有实时计算和仿生的特点,在处理时间序列数据上具有巨大潜力。为了研究如何提高神经网络模型训练性能,降低计算复杂度,文章首先梳理和回顾了近几年相关研究文献,其次提出硬件实现和软件模型两个优化思路,并总结了不同优化方法的优势与不足,硬件和软件上的优化可以提高神经网络模型学习性能和训练速度,但依然存在可控性差、算法最优解未知等问题,最后针对以上问题对未来的研究方向进行了展望,可为时间序列数据处理和模式识别领域提供优化思路。展开更多
现有计算机体系架构下的神经网络难以对多任务复杂数据进行高效处理,成为制约人工智能技术发展的瓶颈之一,而人脑的并行运算方式具有高效率、低功耗和存算一体的特点,被视为打破传统冯·诺依曼计算体系最具潜力的运算体系.突触仿生...现有计算机体系架构下的神经网络难以对多任务复杂数据进行高效处理,成为制约人工智能技术发展的瓶颈之一,而人脑的并行运算方式具有高效率、低功耗和存算一体的特点,被视为打破传统冯·诺依曼计算体系最具潜力的运算体系.突触仿生器件是指从硬件层面上实现人脑神经拟态的器件,它可以模拟脑神经对信息的处理方式,即“记忆”和“信息处理”过程在同一硬件上实现,这对于构建新的运算体系具有重要的意义.近年,制备仿生突触器件的忆阻材料已获得进展,但多聚焦于神经突触功能的模拟,对于时空信息感知和传递的关键研究较为缺乏.本文通过制备一种双层结构忆阻器,实现了突触仿生器件的基本功能包括双脉冲易化和抑制、脉冲时间依赖突触可塑性(spiking time dependent plasticity,STDP)和经验式学习等,还对器件的信息感知、传递特性和稳定性进行了研究,发现该器件脉冲测试结果满足神经网络处理时空信息的基本要求,这一结果可以为忆阻器在类脑芯片中的应用提供参考.展开更多
Structure of nonnegative nontrivial and positive solutions was precisely studied for some singularly perturbed p-Laplace equations. By virtue of sub- and supersolution method, it is shown that there are many nonnegati...Structure of nonnegative nontrivial and positive solutions was precisely studied for some singularly perturbed p-Laplace equations. By virtue of sub- and supersolution method, it is shown that there are many nonnegative nontrivial spike-layer solutions and positive intermediate spike-layer solutions. Moreover, the upper and lower bound on the measure of each spike-layer were estimated when the parameter is sufficiently small.展开更多
基金Supported by the National Natural Science Foundation of China(11271247)the Excellent Youth Talented Project of the Colleges and Universities in Anhui Province(gxyqZD2016520)+1 种基金the Key Project for Teaching Research in Anhui Province(2017jyxm0591,2018jyxm0594)the Key Project for Natural Science Research in Anhui Province(KJ2015A347,KJ2019A1300).
文摘The singularly perturbed Robin boundary value problem for the higher order elliptic equation is considered.Under suitable conditions,the existence and asymptotic behavior of solution to the boundary value problems are studied.The uniform validity of its asymptotic expansion is proved by using the fixed point theorem.
文摘液体状态机(Liquid State Machine,LSM)具有实时计算和仿生的特点,在处理时间序列数据上具有巨大潜力。为了研究如何提高神经网络模型训练性能,降低计算复杂度,文章首先梳理和回顾了近几年相关研究文献,其次提出硬件实现和软件模型两个优化思路,并总结了不同优化方法的优势与不足,硬件和软件上的优化可以提高神经网络模型学习性能和训练速度,但依然存在可控性差、算法最优解未知等问题,最后针对以上问题对未来的研究方向进行了展望,可为时间序列数据处理和模式识别领域提供优化思路。
基金Supported by the National Natural Science Foundation of China(11202106)the Natural Science Foundation from the Education Bureau of Anhui Province(KJ2011A135)
文摘现有计算机体系架构下的神经网络难以对多任务复杂数据进行高效处理,成为制约人工智能技术发展的瓶颈之一,而人脑的并行运算方式具有高效率、低功耗和存算一体的特点,被视为打破传统冯·诺依曼计算体系最具潜力的运算体系.突触仿生器件是指从硬件层面上实现人脑神经拟态的器件,它可以模拟脑神经对信息的处理方式,即“记忆”和“信息处理”过程在同一硬件上实现,这对于构建新的运算体系具有重要的意义.近年,制备仿生突触器件的忆阻材料已获得进展,但多聚焦于神经突触功能的模拟,对于时空信息感知和传递的关键研究较为缺乏.本文通过制备一种双层结构忆阻器,实现了突触仿生器件的基本功能包括双脉冲易化和抑制、脉冲时间依赖突触可塑性(spiking time dependent plasticity,STDP)和经验式学习等,还对器件的信息感知、传递特性和稳定性进行了研究,发现该器件脉冲测试结果满足神经网络处理时空信息的基本要求,这一结果可以为忆阻器在类脑芯片中的应用提供参考.
文摘Structure of nonnegative nontrivial and positive solutions was precisely studied for some singularly perturbed p-Laplace equations. By virtue of sub- and supersolution method, it is shown that there are many nonnegative nontrivial spike-layer solutions and positive intermediate spike-layer solutions. Moreover, the upper and lower bound on the measure of each spike-layer were estimated when the parameter is sufficiently small.
