Principal Manifolds and Nonlinear Dimensionality Reduction via Tangent Space Alignment

We present a new algorithm for manifold learning and nonlinear dimensionality reduction. Based on a set of unorganized data points sampled with noise from a parameterized manifold, the local geometry of the manifold is learned by constructing an approximation for the tangent space at each point, and those tangent spaces are then aligned to give the global coordinates of the data points with respect to the underlying manifold. We also present an error analysis of our algorithm showing that reconstruction errors can be quite small in some cases. We illustrate our algorithm using curves and surfaces both in 2D/3D Euclidean spaces and higher dimensional Euclidean spaces. We also address several theoretical and algorithmic issues for further research and improvements.

Expanders, Group Extensions, Hadamard Manifolds and Certain Banach Spaces

In this note,we prove that expanders cannot be coarsely embedded into group extensions of sequences of groups which are coarsely embeddable into Hardamad manifolds and certain Banach spaces due to the similar concentration theorems.

Compact Space-like Submanifolds in Pseudo-Riemann Manifold

We first establish a integral inequality for compact maximal space-like subman-ifolds in pseudo-Riemannian manifolds N^np+p. Then, we investigate compact space-like sub-manifolds and hupersurfaces with parallel second fundamental form in N^np+p and some other N^np+pambient spaces. We obtain some distribution theorems for the square norm of the secondfundamental form.

Secret Sharing Scheme Based on the Differential Manifold

In this paper, the concepts of topological space and differential manifold are introduced, and it is proved that the surface determined by function F (x2, x2, …, xt) of class Cr in Euelidean Rt is a differential manifold. Using the intersection of the tangent plane and the hypernormal of the differential manifold to construct the shared master key of participants, an intuitive, secure and complete (t,n)-threshold secret sharing scheme is designed. The paper is proved to be safe, and the probability of successful attack of attackers is only 1/pt-1. When the prime number p is sufficiently large, the probability is almost 0. The results show that this scheme has the characteristics of single-parameter representation of the master key in the geometric method, and is more practical and easy to implement than the Blakley threshold secret sharing scheme.

一种基于流形的机械臂动作构型知识压缩表达方法

针对在机械臂分拣任务中,存在物体形状各异、大小不一、训练神经网络成本过高的问题,提出一种基于流形空间的机械臂快速分拣方法。通过自主设计的一款简易实验装置模拟代替机械臂进行实验。对高维数据进行压缩,结合三维快速凸包求解算法,对体积大小不同的同类物体的流形空间进行分割,以凸包形式将稳定性较高的点集包裹起来。实验结果表明,体积大小不同的同类物体的高稳定流形子空间是一致的。该方法可以通过对一种物体的流形子结构进行尺度放缩,得到不同大小的同类物体的高稳定分拣区域,用于生成高效、可靠的机械臂分拣任务中的6D位姿构型,以提高分拣作业的工作效率。

Ricci Curvature of Certain Submanifolds in Kenmotsu Space Forms

In this paper, we obtain some sharp inequalities between the Ricci cur- vature and the squared mean curvature for bi-slant and semi-slant submanifolds in Kenmotsu space forms. Estimates of the scalar curvature and the k-Ricci curvature, in terms of the squared mean curvature, are also proved respectively.

有亏格为1的Heegaard分解的三维流形中的环面纽结

将存在亏格为1的Heegaard分解T’_(1)∪F T’_(2)的三维流形记为M=L(p,q),其中p和q是互素整数,q/p为T’_(2)的纬线在T’_(1)上的斜率.若环面F上的简单闭曲线γ在M中非平凡,则称γ是M中的环面纽结.本文对在M中沿环面纽结作m/n-Dehn手术所得流形进行了分类,并给出了两个实心环体沿边界上平环作融合所得流形是L(p,q)中环面纽结补的特征描述.

正曲率齐性Finsler空间的分类:偶数维情形下的一种新方法(英文)

本文介绍了正曲率齐性Finsler流形的分类。在偶数维的情形下,给出了一种新方法,证明了偶数维光滑陪集空间上有正曲率齐性Finsler度量,当且仅当其上面有正曲率齐性黎曼度量。

A NEW PROOF OF GAFFNEY’S INEQUALITY FOR DIFFERENTIAL FORMS ON MANIFOLDS-WITH-BOUNDARY:THE VARIATIONAL APPROACH à LA KOZONO-YANAGISAWA

Let(M,g_(0))be a compact Riemannian manifold-with-boundary.We present a new proof of the classical Gaffney inequality for differential forms in boundary value spaces over M,via a variational approach a la Kozono-Yanagisawa[Lr-variational inequality for vector fields and the Helmholtz-Weyl decomposition in bounded domains,Indiana Univ.Math.J.58(2009),1853-1920],combined with global computations based on the Bochner technique.

Multi-Scale Object Perception with Embedding Textural Space

This paper mainly focuses on the issues about generic multi-scale object perception for detection or recognition. A novel computational model in visually-feature space is presented for scene & object representation to purse the underlying textural manifold statistically in nonparametric manner. The associative method approximately makes perceptual hierarchy in human-vision biologically coherency in specific quad-tree-pyramid structure, and the appropriate scale-value of different objects can automatically be selected by evaluating from well-defined scale function without any priori knowledge. The sufficient experiments truly demonstrate the effectiveness of scale determination in textural manifold with object localization rapidly.

Completing Einstein’s Spacetime

The four-dimensional character of Einstein’s spacetime is generally accepted in mainstream physics as beyond reasonable doubt correct. However the real problem is when we require scale invariance and that this spacetime be four-dimensional on all scales. It is true that on our classical scale, the 4D decouples into 3D plus one time dimension and that on very large scale only the curvature of spacetime becomes noticeable. However the critical problem is that such spacetime must remain 4D no matter how small the scale we are probing is. This is something of crucial importance for quantum physics. The present work addresses this basic, natural and logical requirement and shows how many contradictory results and shortcomings of relativity and quantum gravity could be eliminated when we “complete” Einstein’s spacetime in such a geometrical gauge invariant way. Concurrently the work serves also as a review of the vast Literature on E-Infinity theory used here.

Including Space-Time in the Extended Group Cl3* of Relativistic Form-Invariance

The inclusion of space-time in the extended group of relativistic form-invariance, Cl3*, is specified as the inclusion of the whole space-time manifold in this multiplicative Lie group. First physical results presented here are: the geometric origin of the time arrow, a better understanding of the non-simultaneity in optics and a mainly geometric origin for the universe expansion, and its recent acceleration.

A Quantum Representation of the Homogeneous 5D Manifold and the Perelman Mappings of 5D onto Non-Homogeneous Lorentz 4D Manifolds

The expression of the Maxwell magnetic monopole was employed to correlate the space to space projection that gives rise to the Gell-Mann standard model, and space to time projection which gives the leptons;and how does it correlate to the Perelman mappings from the homogeneous 5D manifold to the Lorentz 4D manifold, together with correlating the physical consequences caused by the breaking of the Diagonal Long Range Order [DLRO] of the monopoles quantum states affected by the motion of massive particles in the Lorentz 4D boundary of the 5D manifold, which leads to gravitons and the gravity field via the General Relativity covariant Riemannian 4D curvatures metric equation.

数字孪生赋能下的互动生成式规划与治理

21世纪以来,数字孪生进入城市规划与治理领域,加快了城市的感知和问题反馈的过程,形成了时间、空间和人之间新的互动机制。数字孪生城市的挑战来自城市的复杂性,提出以人为主的沉浸参与式角度来重新定义空间,强调空间的多维度延伸,以空间流形的模型为基础,建立起数字孪生城市的空间表达方式及其升维与降维的计算逻辑。基于此,探讨数字孪生赋能下的“互动生成式”规划模式:通过数字孪生的参数化设计,以人为主体,不断调整和实时反馈城市的各项规划参数,构建不断演进的数字孪生的规划治理系统及政策。

Gravitational Space-Time Curve Generation via Accelerated Charged Particles

A force with an acceleration that is equal to multiples greater than the speed of light per unit time is exerted on a cloud of charged particles. The particles are resultantly accelerated to within an infinitesimal fraction of the speed of light. As the force or acceleration increases, the particles’ velocity asymptotically approaches but never achieves the speed of light obeying relativity. The asymptotic increase in the particles’ velocity toward the speed of light as acceleration increasingly surpasses the speed of light per unit time does not compensate for the momentum value produced on the particles at sub-light velocities. Hence, the particles’ inertial mass value must increase as acceleration increases. This increase in the particles’ inertial mass as the particles are accelerated produce a gravitational field which is believed to occur in the oscillation of quarks achieving velocities close to the speed of light. The increased inertial mass of the density of accelerated charged particles becomes the source mass (or Big “M”) in Newton’s equation for gravitational force. This implies that a space-time curve is generated by the accelerated particles. Thus, it is shown that the acceleration number (or multiple of the speed of light greater than 1 per unit of time) and the number of charged particles in the cloud density are surjectively mapped to points on a differential manifold or space-time curved surface. Two aspects of Einstein’s field equations are used to describe the correspondence between the gravitational field produced by the accelerated particles and the resultant space-time curve. The two aspects are the Schwarzchild metric and the stress energy tensor. Lastly, the possibility of producing a sufficient acceleration or electromagnetic force on the charged particles to produce a gravitational field is shown through the Lorentz force equation. Moreover, it is shown that a sufficient voltage can be generated to produce an acceleration/force on the particles that is multiples greater than the speed of light per unit time thereby generating gravity.

Feature Extraction of Kernel Regress Reconstruction for Fault Diagnosis Based on Self-organizing Manifold Learning

The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.

面向相交多流形聚类的标签传播算法

经典的流形学习算法假设样本数据位于高维单流形上,但在现实生活中的真实数据通常位于高维多流形上,且这些数据往往相互交叠,导致流形学习算法效果不佳。传统的标签传播算法通过相似性矩阵构建连接矩阵,实现良好分离数据的聚类,但不能有效聚类相互交叠的多流形数据。针对该问题,提出一种面向相交多流形的标签传播算法LPAMMC。采用局部主成分分析算法确定相交多流形数据的相交区域,并基于混合概率主成分分析(MPPCA)模型和多流形的拓扑结构划分相互交叠的子流形,构建“must-link”和“cannot-link”聚类约束,通过约束构建适合相交多流形数据的传播矩阵,实现标签传播算法。LPAMMC算法通过MPPCA模型和多流形拓扑结构划分出子流形,提高相交多流形数据的聚类精度,且MPPCA模型仅用于多流形数据的相交区域,降低了计算复杂度。实验结果表明,LPAMMC算法不仅具有标签传播算法速度快的特点,且能有效聚类相交多流形数据。在Two spirals数据集上的聚类精度、标准互信息和调整兰德系数取得了与SMMC算法相同的性能,运行时间缩短86.7个百分点。

ON THE INVARIANT SUBMANIFOLDS OF RIEMANNIAN PRODUCT MANIFOLD

In this paper, the vertical and horizontal distributions of an invariant sub-manifold of a Riemannian product manifold are discussed. An invariant real space form in a Riemannian product manifold is researched. Finally, necessary and sufficient conditions are given on an invariant submanifold of a Riemannian product manifold to be a locally symmetric and real space form.

A Design Method of Noncoherent Unitary Space-Time Codes

We generalized an constructing method of noncoherent unitary space time codes (N-USTC) over Rayleigh flat fading channels. A family of N-USTCs with T symbol peroids, M transmit and N receive antennas was constructed by the exponential mapping method based on the tangent subspace of the Grassmann manifold. This exponential mapping method can transform the coherent space time codes (C-STC) into the N-USTC on the Grassmann manifold. We infered an universal framework of constructing a C-STC that is designed by using the algebraic number theory and has full rate and full diversity (FRFD) for t symbol periods and same antennas, where M, N, T, t are general positive integer. We discussed the constraint condition that the exponential mapping has only one solution, from which we presented a approach of searching the optimum adjustive factor αopt that can generate an optimum noncoherent codeword. For different code parameters M, N, T, t and the optimum adjustive factor αopt, we gave the simulation results of the several N-USTCs.

内外特征一致性对齐的非等距三维模型对应关系计算

针对非等距三维模型对应关系计算准确率低且难以自动化的问题,提出一种基于局部流形调和基与内外积空间对齐的三维模型对应关系计算算法.首先利用局部流形调和基作为模型本征信息,并结合笛卡儿坐标等外部信息嵌入内外积空间,实现本征信息和外部信息对齐;其次将计算对齐的目标函数与一致性点偏移算法融合,提高结果的稳定性和准确性;最后利用基于交替方向乘子法的交替优化算法来求解目标函数,构建最终的对应关系结果.与已有算法的对比实验结果表明,所提算法在SMAL,SHERC’19,TOSCA和SHERC’16 Topology数据集上构建的对应关系测地误差最小,全局对应关系准确率最高,同时能够处理拓扑噪声和模型自身对称性影响对应关系计算的问题.