|
Nonnegative tensor factorizations using an alternating direction method
Xingju CAI, Yannan CHEN, Deren HAN
Front Math Chin. 2013, 8 (1): 3-18.
https://doi.org/10.1007/s11464-012-0264-8
The nonnegative tensor (matrix) factorization finds more and more applications in various disciplines including machine learning, data mining, and blind source separation, etc. In computation, the optimization problem involved is solved by alternatively minimizing one factor while the others are fixed. To solve the subproblem efficiently, we first exploit a variable regularization term which makes the subproblem far from ill-condition. Second, an augmented Lagrangian alternating direction method is employed to solve this convex and well-conditioned regularized subproblem, and two accelerating skills are also implemented. Some preliminary numerical experiments are performed to show the improvements of the new method.
References |
Related Articles |
Metrics
|
|
lk,s-Singular values and spectral radius of rectangular tensors
Chen LING, Liqun QI
Front Math Chin. 2013, 8 (1): 63-83.
https://doi.org/10.1007/s11464-012-0265-7
The real rectangular tensors arise from the strong ellipticity condition problem in solid mechanics and the entanglement problem in quantum physics. In this paper, we study the singular values/vectors problem of real nonnegative partially symmetric rectangular tensors. We first introduce the concepts of lk,s-singular values/vectors of real partially symmetric rectangular tensors. Then, based upon the presented properties of lk,s-singular values /vectors, some properties of the related lk,s-spectral radius are discussed. Furthermore, we prove two analogs of Perron-Frobenius theorem and weak Perron-Frobenius theorem for real nonnegative partially symmetric rectangular tensors.
References |
Related Articles |
Metrics
|
|
Maximal number of distinct H-eigenpairs for a two-dimensional real tensor
Kelly J. PEARSON, Tan ZHANG
Front Math Chin. 2013, 8 (1): 85-105.
https://doi.org/10.1007/s11464-012-0263-9
Based on the generalized characteristic polynomial introduced by J. Canny in Generalized characteristic polynomials [J. Symbolic Comput., 1990, 9(3): 241–250], it is immediate that for any m-order n-dimensional real tensor, the number of distinct H-eigenvalues is less than or equal to n(m-1)n-1. However, there is no known bounds on the maximal number of distinct Heigenvectors in general. We prove that for any m≥2, an m-order 2-dimensional tensor A exists such that A has 2(m - 1) distinct H-eigenpairs. We give examples of 4-order 2-dimensional tensors with six distinct H-eigenvalues as well as six distinct H-eigenvectors. We demonstrate the structure of eigenpairs for a higher order tensor is far more complicated than that of a matrix. Furthermore, we introduce a new class of weakly symmetric tensors, called p-symmetric tensors, and show under certain conditions, p-symmetry will effectively reduce the maximal number of distinct H-eigenvectors for a given two-dimensional tensor. Lastly, we provide a complete classification of the H-eigenvectors of a given 4-order 2-dimensional nonnegative p-symmetric tensor. Additionally, we give sufficient conditions which prevent a given 4-order 2-dimensional nonnegative irreducible weakly symmetric tensor from possessing six pairwise distinct H-eigenvectors.
References |
Related Articles |
Metrics
|
|
Factorization of simple modules for certain restricted two-parameter quantum groups
Min LI, Xiuling WANG
Front Math Chin. 2013, 8 (1): 169-190.
https://doi.org/10.1007/s11464-012-0236-z
We study the representations of the restricted two-parameter quantum groups of types B and G. For these restricted two-parameter quantum groups, we give some explicit conditions which guarantee that a simple module can be factored as the tensor product of a one-dimensional module with a module that is naturally a module for the quotient by central group-like elements. That is, given θ a primitive lth root of unity, the factorization of simple ?θy,θz,( )-modules is possible, if and only if (2(y - z), l) = 1 for =??2n+1; (3(y - z), l) = 1 for g= G2.
References |
Related Articles |
Metrics
|
|
Nowhere-zero 3-flows in matroid base graph
Yinghao ZHANG, Guizhen LIU
Front Math Chin. 2013, 8 (1): 217-227.
https://doi.org/10.1007/s11464-012-0246-x
The base graph of a simple matroid M=(E,?) is the graph G such that V(G)=? and E(G)={BB′:B,B′∈?,|B\B′|=1}, where the same notation is used for the vertices of G and the bases of M. It is proved that the base graph G of connectedsimple matroid M is Z3-connected if |V (G)|≥5. We also proved that if M is not a connected simple matroid, then the base graph G of M does not admit a nowhere-zero 3-flow if and only if |V (G)| = 4. Furthermore, if for every connected component Ei (i≥2) of M, the matroid ase graph Gi of Mi = M|Ei has |V (Gi)|≥5, then G is Z3-connected which also implies that G admits nowhere-zero 3-flow immediately.
References |
Related Articles |
Metrics
|
15 articles
|