In this paper, a new approach is presented for finite-time control problems for linear systems subject to time-varying parametric uncertainties and exogenous disturbance. The disturbance is assumed to be time varying ...In this paper, a new approach is presented for finite-time control problems for linear systems subject to time-varying parametric uncertainties and exogenous disturbance. The disturbance is assumed to be time varying and bounded. Sufficient conditions are obtained for the existence of a linear parameter-dependent state feedback gain, which can ensure that the closed-loop system is finite-time bounded （FTB）. The conditions can be reduced to feasibility problems involving LMIs. Numerical examples show the validity of the proposed methodology.展开更多
We propose two more general methods to construct nullnorms on bounded lattices. By some illustrative examples, we demonstrate that the new method differ from the existing approaches.
In this paper, a simplified definition of boundedness of the sets in probabilistic normed linear space was introduced. By means of the probabilistic norm of linear operators, the linear operator theory on probabilisti...In this paper, a simplified definition of boundedness of the sets in probabilistic normed linear space was introduced. By means of the probabilistic norm of linear operators, the linear operator theory on probabilistic normed linear space was further studied. On probabilistic normed linear operator space, some resonance theorems dealing with probabilistic bounded sets, probabilistic semi_bounded sets, and probabilistic non_unbounded sets are obtained.展开更多
It is proved that there is only one L^P-matricially normed space of dimension 1 and that quotient spaces of L^P-matricially normed spaces are also L^P-matricially normed spaces. Some properties of L^P-matricially norm...It is proved that there is only one L^P-matricially normed space of dimension 1 and that quotient spaces of L^P-matricially normed spaces are also L^P-matricially normed spaces. Some properties of L^P-matricially normed spaces are given.展开更多
We prove a new version of the Holevo bound employing the Hilbert-Schmidt norm instead of the Kullback-Leibler divergence. Suppose Alice is sending classical information to Bob by using a quantum channel while Bob is p...We prove a new version of the Holevo bound employing the Hilbert-Schmidt norm instead of the Kullback-Leibler divergence. Suppose Alice is sending classical information to Bob by using a quantum channel while Bob is performing some projective measurements. We bound the classical mutual information in terms of the Hilbert-Schmidt norm by its quantum Hilbert-Schmidt counterpart. This constitutes a Holevo-type upper bound on the classical information transmission rate via a quantum channel. The resulting inequality is rather natural and intuitive relating classical and quantum expressions using the same measure.展开更多
边界框回归分支是深度目标跟踪器的关键模块,其性能直接影响跟踪器的精度.评价精度的指标之一是交并比(Intersection over union,IoU).基于IoU的损失函数取代了l_(n)-norm损失成为目前主流的边界框回归损失函数,然而IoU损失函数存在2个...边界框回归分支是深度目标跟踪器的关键模块,其性能直接影响跟踪器的精度.评价精度的指标之一是交并比(Intersection over union,IoU).基于IoU的损失函数取代了l_(n)-norm损失成为目前主流的边界框回归损失函数,然而IoU损失函数存在2个固有缺陷:1)当预测框与真值框不相交时IoU为常量0,无法梯度下降更新边界框的参数;2)在IoU取得最优值时其梯度不存在,边界框很难收敛到IoU最优处.揭示了在回归过程中IoU最优的边界框各参数之间蕴含的定量关系,指出在边界框中心处于特定位置时存在多种尺寸不同的边界框使IoU损失最优的情况,这增加了边界框尺寸回归的不确定性.从优化两个统计分布之间散度的视角看待边界框回归问题,提出了光滑IoU(Smooth-IoU,SIoU)损失,即构造了在全局上光滑(即连续可微)且极值唯一的损失函数,该损失函数自然蕴含边界框各参数之间特定的最优关系,其唯一取极值的边界框可使IoU达到最优.光滑性确保了在全局上梯度存在使得边界框更容易回归到极值处,而极值唯一确保了在全局上可梯度下降更新参数,从而避开了IoU损失的固有缺陷.提出的光滑损失可以很容易取代IoU损失集成到现有的深度目标跟踪器上训练边界框回归,在LaSOT、GOT-10k、TrackingNet、OTB2015和VOT2018测试基准上所取得的结果,验证了光滑IoU损失的易用性和有效性.展开更多
基金the Scientific Innovation Team Project of Hubei Provincial Department of Education (T200809)the Science Foundationof Education Commission of Hubei Province (No. D20081306)the Doctoral Pre-research Foundation of Three Gorges University
文摘In this paper, a new approach is presented for finite-time control problems for linear systems subject to time-varying parametric uncertainties and exogenous disturbance. The disturbance is assumed to be time varying and bounded. Sufficient conditions are obtained for the existence of a linear parameter-dependent state feedback gain, which can ensure that the closed-loop system is finite-time bounded （FTB）. The conditions can be reduced to feasibility problems involving LMIs. Numerical examples show the validity of the proposed methodology.
文摘We propose two more general methods to construct nullnorms on bounded lattices. By some illustrative examples, we demonstrate that the new method differ from the existing approaches.
文摘In this paper, a simplified definition of boundedness of the sets in probabilistic normed linear space was introduced. By means of the probabilistic norm of linear operators, the linear operator theory on probabilistic normed linear space was further studied. On probabilistic normed linear operator space, some resonance theorems dealing with probabilistic bounded sets, probabilistic semi_bounded sets, and probabilistic non_unbounded sets are obtained.
文摘It is proved that there is only one L^P-matricially normed space of dimension 1 and that quotient spaces of L^P-matricially normed spaces are also L^P-matricially normed spaces. Some properties of L^P-matricially normed spaces are given.
文摘We prove a new version of the Holevo bound employing the Hilbert-Schmidt norm instead of the Kullback-Leibler divergence. Suppose Alice is sending classical information to Bob by using a quantum channel while Bob is performing some projective measurements. We bound the classical mutual information in terms of the Hilbert-Schmidt norm by its quantum Hilbert-Schmidt counterpart. This constitutes a Holevo-type upper bound on the classical information transmission rate via a quantum channel. The resulting inequality is rather natural and intuitive relating classical and quantum expressions using the same measure.
文摘边界框回归分支是深度目标跟踪器的关键模块,其性能直接影响跟踪器的精度.评价精度的指标之一是交并比(Intersection over union,IoU).基于IoU的损失函数取代了l_(n)-norm损失成为目前主流的边界框回归损失函数,然而IoU损失函数存在2个固有缺陷:1)当预测框与真值框不相交时IoU为常量0,无法梯度下降更新边界框的参数;2)在IoU取得最优值时其梯度不存在,边界框很难收敛到IoU最优处.揭示了在回归过程中IoU最优的边界框各参数之间蕴含的定量关系,指出在边界框中心处于特定位置时存在多种尺寸不同的边界框使IoU损失最优的情况,这增加了边界框尺寸回归的不确定性.从优化两个统计分布之间散度的视角看待边界框回归问题,提出了光滑IoU(Smooth-IoU,SIoU)损失,即构造了在全局上光滑(即连续可微)且极值唯一的损失函数,该损失函数自然蕴含边界框各参数之间特定的最优关系,其唯一取极值的边界框可使IoU达到最优.光滑性确保了在全局上梯度存在使得边界框更容易回归到极值处,而极值唯一确保了在全局上可梯度下降更新参数,从而避开了IoU损失的固有缺陷.提出的光滑损失可以很容易取代IoU损失集成到现有的深度目标跟踪器上训练边界框回归,在LaSOT、GOT-10k、TrackingNet、OTB2015和VOT2018测试基准上所取得的结果,验证了光滑IoU损失的易用性和有效性.
