Abstract
In many contexts one encounters Hermitian operators M on a Hilbert space whose dimension is so large that it is impossible to write down all matrix entries in an orthonormal basis. How does one determine whether such M is positive semidefinite? Here we approach this problem by deriving asymptotically optimal bounds to the distance to the positive semidefinite cone in Schatten p-norm for all integer \(p\in [1,\infty )\), assuming that we know the moments \(\mathbf {tr}(M^k)\) up to a certain order \(k=1,\ldots , m\). We then provide three methods to compute these bounds and relaxations thereof: the sos polynomial method (a semidefinite program), the Handelman method (a linear program relaxation), and the Chebyshev method (a relaxation not involving any optimization). We investigate the analytical and numerical performance of these methods and present a number of example computations, partly motivated by applications to tensor networks and to the theory of free spectrahedra.
Similar content being viewed by others
Notes
If \(r=1\), then there is a simple criterion to determine whether M is psd: M is psd if and only if each \(A^{[i]}\) is either psd or negative semidefinite, and the number of negative semidefinite matrices is even. But we are not aware of any such simple criterion for \(r>1\).
This has been pointed out to us by Richard Kueng.
One way to see this is by the Portmanteau theorem: the cumulative distribution functions of \(\mu \) and \(\nu _t\) differ by at most \(t^{-1}\) at every point, and therefore we have (even uniform) convergence as \(t\rightarrow \infty \), which implies weak-\(^*\) convergence \(\nu _t \rightarrow \mu \).
Although M is only of size \(2\times 2\), one can clearly achieve the same moments on larger matrices by simply repeating the eigenvalues.
Communicated to us by Boaz Barak.
This is not to be confused with the sos polynomial method of [4], which is a semidefinite program that computes \(\min _p \Vert M - p(M) \Vert _1\), where p is a sos polynomial of given degree m. The goal of the method of [4] is to approximate M as well as possible with a sos polynomial (as this provides a purification), which is possible only if M is psd. Note moreover that \(\Vert M - p(M) \Vert _1\) cannot be computed from the moments of M.
This can also be deduced directly from Theorem 3, using \(x=x^2+x(1-x)\) and \(1-x=(1-x)^2+x(1-x)\).
References
Ben-Tal, A., Nemirovski, A.: On tractable approximations of uncertain linear matrix inequalities affected by interval uncertainty. SIAM J. Optim. 12(3), 811–833 (2002)
Bhatia, R.: Matrix Analysis. Springer, Berlin (1991)
Blekherman, G., Parrilo, P.A., Thomas, R.R. (eds.): Semidefinite optimization and convex algebraic geometry. MOS-SIAM Series on Optimization (2013)
De las Cuevas, G., Schuch, N., Perez-Garcia, D., Cirac, J .I.: Purifications of multipartite states: limitations and constructive methods. New J. Phys. 15, 123021 (2013)
De las Cuevas, G., Cubitt, T .S., Cirac, J .I., Wolf, M .M., Perez-Garcia, D.: Fundamental limitations in the purifications of tensor networks. J. Math. Phys 57, 071902 (2016)
Fannes, M., Nachtergaele, B., Werner, R.F.: Finitely correlated states on quantum spin chains. Commun. Math. Phys. 144, 443 (1992)
Fritz, T., Netzer, T., Thom, A.: Spectrahedral containment and operator systems with finite-dimensional realization. SIAM J. Appl. Algebra Geom. 1, 556 (2016)
Gil, A., Segura, J., Temme, N.M.: Numerical Methods for Special Functions. Society for Industrial and Applied Mathematics (SIAM), Philadelphia (2007)
Handelman, D.: Representing polynomials by positive linear functions on compact convex polyhedra. Pac. J. Math. 132, 35 (1988)
Helton, J.W., Klep, I., McCullough, S.: Free convex algebraic geometry. In: Semidefinite Optimization and Convex Algebraic Geometry, volume 13 of MOS-SIAM Ser. Optim., pp. 341–405. SIAM, Philadelphia, PA (2013)
Helton, J .W., Klep, I., McCullough, S.: The matricial relaxation of a linear matrix inequality. Math. Program. 138(1–2, Ser. A), 401–445 (2013)
Hiai, F., Nakamura, Y.: Majorizations for generalized S-numbers in semifinite von Neumann algebras. Math. Z. 195(1), 17–27 (1987)
Higham, N.J.: Functions of Matrices. SIAM, Philadelphia (2008)
Kliesch, M., Gross, D., Eisert, J.: Matrix product operators and states: NP-hardness and undecidability. Phys. Rev. Lett. 113, 160503 (2014)
Kreuin, M.G., Nudel’man, A.A.: The Markov Moment Problem and Extremal Problems. American Mathematical Society, Providence (1977)
Kuhlmann, S., Marshall, M., Schwartz, N.: Positivity, sums of squares and the multi-dimensional moment problem. Adv. Geom. 5(4), 583–606 (2005)
Lasserre, J.B.: Bounding the support of a measure from its marginal moments. Proc. Am. Math. Soc. 139(9), 3375–3382 (2011)
Marshall, M., Kuhlmann, S.: Positivity, sums of squares and the multi-dimensional moment problem. Trans. Am. Math. Soc. 354(11), 4285–4301 (2002)
Orús, R.: A practical introduction to tensor networks: matrix product states and projected entangled pair states. Ann. Phys. 349, 117 (2014)
Perez-Garcia, D., Verstraete, F., Wolf, M.M., Cirac, J.I.: Matrix product state representations. Quantum Inf. Comput. 7, 401 (2007)
Riesz, M.: Sur le problème des moments. Troisième note, Arkiv för Matematik, Astronomi och Fysik 17(16) (1923)
Rivlin, T.J.: An Introduction to the Approximation of Functions. Dover, New York (1981)
Rudin, W.: Functional Analysis. International Series in Pure and Applied Mathematics, 2nd edn. McGraw-Hill Inc, New York (1991)
Shohat, J.A., Tamarkin, J.D.: The Problem of Moments, 4th edn. American Mathematical Society, New York (1970)
Verstraete, F., Garcia-Ripoll, J.J., Cirac, J.I.: Matrix product density operators: simulation of finite-temperature and dissipative systems. Phys. Rev. Lett. 93, 207204 (2004)
Vidal, G.: Efficient classical simulation of slightly entangled quantum computation. Phys. Rev. Lett. 91, 147902 (2003)
Villani, C.: Optimal transport: old and new. In: Grundlehren der mathematischen Wissenschaften, chapter 338. Springer (2009)
Werner, A.H., Jaschke, D., Silvi, P., Calarco, T., Eisert, J., Montangero, S.: A positive tensor network approach for simulating open quantum many-body systems. Phys. Rev. Lett. 116, 23720 (2016)
Acknowledgements
GDLC acknowledges funding of the Elise Richter Program of the FWF. TN acknowledges funding through the FWF Project P 29496-N35 (free semialgebraic geometry and convexity). Most of this work was conducted while TF was at the Max Planck Institute for Mathematics in the Sciences. We thank Hilary Carteret and Andreas Thom for helpful comments on the topic.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by M. M. Wolf
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extension to Von Neumann Algebras
Extension to Von Neumann Algebras
Let \({\mathcal {N}}\) be a von Neumann algebra equipped with a faithful normal trace \(\mathbf {tr}\) satisfying \(\mathbf {tr}(1) = 1\). For example, \({\mathcal {N}}\) may be the group von Neumann algebra of a discrete group \(\varGamma \), defined as the weak operator closure of the group algebra \({\mathbb {C}}[\varGamma ]\) as acting on \(\ell ^2(\varGamma )\); the trace is given by \(\mathbf {tr}(x) := \langle e,xe\rangle \) with \(e\in \varGamma \) being the unit. In the case where \({\mathcal {N}}= M_s({\mathbb {C}})\) is just the matrix algebra, the discussion presented here specializes to that of the main text.
For a Hermitian element \(M\in {\mathcal {N}}\), the Schatten p-norm is again defined as
where the absolute value and power functions are defined in terms of functional calculus. We can now ask the same question as in the main text: suppose we have M of which we only know the values of the first m moments
Then what can we say about the p-distance from M to the cone of positive elements?
It is straightforward to see that essentially all of the methods of the main text still apply without any change. The only differences are the following:
Remark 1 no longer applies, since the estimate that we used there involves the matrix size explicitly. Hence there is no bound on \(\Vert M\Vert _\infty \) that could be computed from the moments. So in order to scale M such that \(\Vert M\Vert _\infty \le 1\) is guaranteed, an a priori bound on \(\Vert M\Vert _\infty \) needs to be known in addition to the moments. Otherwise our methods will not apply in their current form.
The proof of Proposition 1 is still essentially the same, but \(\sigma \) needs to be generalized to the spectral scale, and the corresponding inequality is [12, Corollary 3.3(1)].
Everything else is completely unchanged, including the fact that we are secretly addressing a version of the Hausdorff moment problem.
Rights and permissions
About this article
Cite this article
De las Cuevas, G., Fritz, T. & Netzer, T. Optimal Bounds on the Positivity of a Matrix from a Few Moments. Commun. Math. Phys. 375, 105–126 (2020). https://doi.org/10.1007/s00220-020-03720-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00220-020-03720-5