Positive definite functions of noncommuting contractions, Hua-Bellman matrices, and a new distance metric

We study positive definite functions on noncommuting strict contractions. In particular, we study functions that induce positive definite Hua-Bellman matrices (i.e., matrices of the form [det(I − Ai Aj)−α]ij where Ai and Aj are strict contractions and α ∈ C). We start by revisiting a 1959 work of Bellman (R. Bellman Representation theorems and inequalities for Hermitian matrices; Duke Mathematical J., 26(3), 1959) that studies Hua-Bellman matrices and claims a strengthening of Hua’s representation theoretic results on their positive definiteness (L.-K. Hua, Inequalities involving determinants; Acta Mathematica Sinica, 5(1955), pp. 463–470). We uncover a critical error in Bellman’s proof that has surprisingly escaped notice to date. We “fix” this error and provide conditions under which det(I − A∗B)−α is a positive definite function; our conditions correct Bellman’s claim and subsume both Bellman’s and Hua’s prior results. Subsequently, we build on our result and introduce a new hyperbolic-like geometry on noncommuting contractions, and remark on its potential applications.

[1]  Mitsuru Uchiyama,et al.  Subadditivity of eigenvalue sums , 2005 .

[2]  Elena Deza,et al.  Encyclopedia of Distances , 2014 .

[3]  T. Andô POSITIVITY OF OPERATOR-MATRICES OF HUA-TYPE , 2008 .

[4]  Harry Crane Some algebraic identities for the α-permanent , 2013 .

[5]  T. Ando Hua—Marcus inequalities , 1980 .

[6]  Marco Congedo,et al.  Riemannian Optimization and Approximate Joint Diagonalization for Blind Source Separation , 2018, IEEE Transactions on Signal Processing.

[7]  D'aniel Virosztek,et al.  The metric property of the quantum Jensen-Shannon divergence , 2019, Advances in Mathematics.

[8]  Joe Harris,et al.  Representation Theory: A First Course , 1991 .

[9]  P. Brändén Solutions to two problems on permanents , 2012 .

[10]  David Vere-Jones A Generalization of Permanents and Determinants , 1988 .

[11]  On the Variation of the Determinant of a Positive Definite Matrix 1) , 1951 .

[12]  Alan D. Sokal,et al.  Complete monotonicity for inverse powers of some combinatorially defined polynomials , 2013, 1301.2449.

[13]  F. Yger,et al.  Riemannian Approaches in Brain-Computer Interfaces: A Review , 2017, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[14]  H. Loo INEQUALITIES INVOLVING DETERMINANTS , 1955 .

[15]  Shuangzhe Liu,et al.  Hadamard, Khatri-Rao, Kronecker and Other Matrix Products , 2008 .

[16]  Romain Couillet,et al.  Random matrix improved covariance estimation for a large class of metrics , 2019, ICML.

[17]  Maher Moakher,et al.  Means of Hermitian positive-definite matrices based on the log-determinant α-divergence function , 2012 .

[18]  G. Szegö,et al.  Über gewisse Potenzreihen mit lauter positiven Koeffizienten , 1933 .

[19]  Suvrit Sra,et al.  Conic Geometric Optimization on the Manifold of Positive Definite Matrices , 2013, SIAM J. Optim..

[20]  S. Sra Positive definite matrices and the S-divergence , 2011, 1110.1773.

[21]  Fuzhen Zhang,et al.  Revisiting hua-marcus-bellman-ando inequalities on contractive matrices , 2009 .

[22]  Petter Brand'en Solutions to two problems on permanents , 2011 .

[23]  Fuzhen Zhang,et al.  Contractive matrices of Hua type , 2011 .

[24]  T. Shirai REMARKS ON THE POSITIVITY OF α-DETERMINANTS , 2007 .

[25]  S. Sra On the matrix square root via geometric optimization , 2015, 1507.08366.

[26]  Freddie Åström,et al.  Unsupervised Label Learning on Manifolds by Spatially Regularized Geometric Assignment , 2018, GCPR.

[27]  A determinantal inequality , 2002 .

[28]  Anoop Cherian,et al.  Jensen-Bregman LogDet Divergence with Application to Efficient Similarity Search for Covariance Matrices , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.