Sampling Matrices From Harish-Chandra-Itzykson-Zuber Densities
Colin McSwiggen (Courant Institute, New York University)
4021 - 4032 of 19283 Results
Format results
Friendship in the Axiverse
David Cyncynates University of Washington
Continuous Maximum Entropy Distributions
Jonathan Leake (Weierstrass Institute Berlin)Discriminating between theories of the very early universe
Jerome Quintin University of Waterloo
Constrained Optimization On Riemannian Manifolds
Melanie Weber (Oxford, Mathematical Institute)Near Optimal Sample Complexity For Matrix And Tensor Normal Models Via Geodesic Convexity
Akshay Ramachandran (University of Amsterdam)Pivot Hamiltonians: a tale of symmetry, entanglement, and quantum criticality
Nathanan Tantivasadakarn California Institute of Technology (Caltech)
Computing The Nc-Rank Via Discrete Convex Optimization On Cat(0) Spaces
Hiroshi Hirai (University of Tokyo)Norm Minimization, Invariant Theory, And Quantum Tomography
Cole Franks (MIT)Regulating Loops in dS
Akhil Premkumar University of California, San Diego
Spectral analysis of product formulas for quantum simulation
Changhao Yi University of New Mexico
Twisted bilayers: magic continuum, noncollinear magnetism and more
Zhu-Xi Luo University of California, Santa Barbara
Sampling Matrices From Harish-Chandra-Itzykson-Zuber Densities
Colin McSwiggen (Courant Institute, New York University)Exponential densities on unitary conjugation orbits of Hermitian matrices, known as Harish-Chandra-Itzykson-Zuber (HCIZ) densities, are important in various settings in physics and random matrix theory. However, the basic question of efficient sampling from these densities has remained open. We present two efficient algorithms to sample matrices from distributions that are close to the HCIZ distribution. Both algorithms exploit a natural self-reducible structure that arises from continuous symmetries of the underlying unitary orbit. We will also discuss applications to quantum inference and differentially private rank-k approximation.Friendship in the Axiverse
David Cyncynates University of Washington
A generic low-energy prediction of string theory is the existence of a large collection of axions, commonly known as a string axiverse. String axions can be distributed over many orders of magnitude in mass, and are expected to interact with one another through their joint potential. In this talk, I will show how non-linearities in this potential lead to a new type of resonant energy transfer between axions with nearby masses. This resonance generically transfers energy from axions with larger decay constants to those with smaller decay constants, leading to a multitude of signatures. These include enhanced direct detection prospects for a resonant pair comprising even a small subcomponent of dark matter, and boosted small-scale structure if the pair is the majority of DM. Near-future iterations of experiments such as ADMX and DM Radio will be sensitive to this scenario, as will astrophysical probes of DM substructure.
Continuous Maximum Entropy Distributions
Jonathan Leake (Weierstrass Institute Berlin)Maximum entropy distributions on discrete domains have been studied in a number of contexts over the past few decades. In particular, they have been used in deterministic algorithms for the approximation of the permanent, scaling problems for torus actions, applications of spanning tree sampling problems, and beyond. These applications are intimately linked to distributions on some finite support set, with a given expectation, and which maximize entropy. In this talk, we will discuss the generalization of maximum entropy distributions from finite discrete support sets to continuous manifolds. Such distributions arise in connection to a number of topics; for example, quantum entropy, interior point methods, matrix distributions studied in statistics, the isotropic constant, and low-rank matrix approximation. We will discuss a few of these applications, along with some of the computability issues which arise when moving from discrete support to continuous. We will also briefly discuss the connection between this generalization and the norm minimization techniques used for general scaling problems. This is joint work with Nisheeth Vishnoi.Discriminating between theories of the very early universe
Jerome Quintin University of Waterloo
There exist various scenarios for the very early universe that could potentially be the explanation for the observed properties of the cosmic microwave background. The current paradigm -- inflationary cosmology -- has rightfully received much attention, but it is not the only theoretically viable explanation. Indeed, several alternative scenarios exist, for example a contracting universe prior to a bounce or a slowly expanding emerging universe. It thus bares the question: how can we discriminate between the various theories, both from a theoretical and an observational point of view? A few pathways to answering this question are discussed in this talk.
Constrained Optimization On Riemannian Manifolds
Melanie Weber (Oxford, Mathematical Institute)Many applications involve non-Euclidean data, where exploiting Riemannian geometry can deliver algorithms that are computationally superior to standard nonlinear programming approaches. This observation has resulted in an increasing interest in Riemannian methods in the optimization and machine learning community. In this talk, we consider the problem of optimizing a function on a Riemannian manifold subject to convex constraints. We introduce Riemannian Frank-Wolfe (RFW) methods, a class of projection-free algorithms for constrained geodesically convex optimization. To understand the algorithm’s efficiency, we discuss (1) its iteration complexity, (2) the complexity of computing the Riemannian gradient, and (3) the complexity of the Riemannian “linear” oracle (RLO), a crucial subroutine at the heart of the algorithm. We complement our theoretical results with an empirical comparison of RFW against state-of-the-art Riemannian optimization methods. Joint work with Suvrit Sra (MIT).Near Optimal Sample Complexity For Matrix And Tensor Normal Models Via Geodesic Convexity
Akshay Ramachandran (University of Amsterdam)The matrix normal model, the family of Gaussian matrix-variate distributions whose covariance matrix is the Kronecker product of two lower dimensional factors, is frequently used to model matrix-variate data. The tensor normal model generalizes this family to Kronecker products of three or more factors. We study the estimation of the Kronecker factors of the covariance matrix in the matrix and tensor models. We show nonasymptotic bounds for the error achieved by the maximum likelihood estimator (MLE) in several natural metrics. In contrast to existing bounds, our results do not rely on the factors being well-conditioned or sparse. For the matrix normal model, all our bounds are minimax optimal up to logarithmic factors, and for the tensor normal model our bound for the largest factor and overall covariance matrix are minimax optimal up to constant factors provided there are enough samples for any estimator to obtain constant Frobenius error. In the same regimes as our sample complexity bounds, we show that an iterative procedure to compute the MLE known as the flip-flop algorithm converges linearly with high probability. Our main tool is geodesic strong convexity in the geometry on positive-definite matrices induced by the Fisher information metric. This strong convexity is determined by the expansion of certain random quantum channels. We also provide numerical evidence that combining the flip-flop algorithm with a simple shrinkage estimator can improve performance in the undersampled regime.Pivot Hamiltonians: a tale of symmetry, entanglement, and quantum criticality
Nathanan Tantivasadakarn California Institute of Technology (Caltech)
I will introduce the notion of Pivot Hamiltonians, a special class of Hamiltonians that can be used to "generate" both entanglement and symmetry. On the entanglement side, pivot Hamiltonians can be used to generate unitary operators that prepare symmetry-protected topological (SPT) phases by "rotating" the trivial phase into the SPT phase. This process can be iterated: the SPT can itself be used as a pivot to generate more SPTs, giving a rich web of dualities. Furthermore, a full rotation can have a trivial action in the bulk, but pump lower dimensional SPTs to the boundary, allowing the practical application of scalably preparing cluster states as SPT phases for measurement-based quantum computation. On the symmetry side, pivot Hamiltonians can naturally generate U(1) symmetries at the transition between the aforementioned trivial and SPT phases. The sign-problem free nature of the construction gives a systematic approach to realize quantum critical points between SPT phases in higher dimensions that can be numerically studied. As an example, I will discuss a quantum Monte Carlo study of a 2D lattice model where we find evidence of a direct transition consistent with a deconfined quantum critical point with emergent SO(5) symmetry.
This talk is based on arXiv:2107.04019, 2110.07599, 2110.09512Computing The Nc-Rank Via Discrete Convex Optimization On Cat(0) Spaces
Hiroshi Hirai (University of Tokyo)In this paper, we address the noncommutative rank (nc-rank) computation of a linear symbolic matrix A=A1 x1+A2 x2 +⋯+ Am xm, where each Ai is an n x n matrix over a field K, and xi (i=1,2,…,m) are noncommutative variables. For this problem, polynomial time algorithms were given by Garg, Gurvits, Oliveira, and Wigderson for K=Q, and by Ivanyos, Qiao, and Subrahmanyam for an arbitrary field K. We present a significantly different polynomial time algorithm that works on an arbitrary field K. Our algorithm is based on a combination of submodular optimization on modular lattices and convex optimization on CAT(0) spaces.Norm Minimization, Invariant Theory, And Quantum Tomography
Cole Franks (MIT)Given many identical copies of a quantum particle, what's the chance measuring their overall angular momentum yields zero? We show how to frame this problem as an optimization problem over a matrix group. We'll then discuss the relationship between this optimization problem and invariant theory, quantum tomography, and the Jacobian conjecture. This talk is based on the joint work https://arxiv.org/abs/2004.14872 with Michael Walter.Regulating Loops in dS
Akhil Premkumar University of California, San Diego
Perturbative QFT calculations in de Sitter are plagued by a variety of divergences. One particular kind, the secular growth terms, cause the naive perturbation expansion to break down at late times. Such contributions often arise from loop integrals, which are notoriously hard to compute in dS. We discuss an approach to evaluate such loop integrals, for a scalar field theory in a fixed de Sitter background. Our method is based on the Mellin-Barnes representation of correlation functions, which enables us to regulate divergences for scalars of any mass while preserving the symmetries of dS. The resulting expressions have a similar structure as a standard dimreg answer in flat space QFT. These features of the regulator are illustrated with two examples. Along the way, we illuminate the physical origin of these divergences and their interpretation within the framework of the dynamical renormalization group. Our calculations naturally reveal additional infrared divergences for massless scalar fields in de Sitter, that are not present in the massive case. Such loop corrections can be incorporated as systematic improvements to the Stochastic Inflation framework, allowing for a more precise description of the IR dynamics of massless fields in de Sitter.
Spectral analysis of product formulas for quantum simulation
Changhao Yi University of New Mexico
Trotter-Suzuki formula is a practical and efficient algorithm for Hamiltonian simulation. It has been widely used in quantum chemistry, quantum field theory and condensed matter physics. Usually, its error is quantified by the operator norm distance between the ideal evolution operator and the digital evolution operator. However, recently more and more papers discovered that, even in large Trotter step region, the quantity of interest can still be accurately simulated. These robustness phenomena imply a different approach of analyzing Trotter-Suzuki formulas. In our previous paper, by analyzing the spectral analysis of the effective Hamiltonian, we successfully established refined estimations of digital errors, and thus improved the circuit complexity of quantum phase estimation and digital adiabatic simulation.
Twisted bilayers: magic continuum, noncollinear magnetism and more
Zhu-Xi Luo University of California, Santa Barbara
Van der Waals heterostructures provide a rich venue for exotic moir\’e phenomena. In this talk, I will present a couple of unconventional examples beyond the celebrated twisted bilayer graphene. I will start by twisted bilayer of square lattice with staggered flux, which exhibits a continuum range of magic twisting angles where an exponential reduction of Dirac velocity and bandwidths occurs. Then I will discuss moir\’e magnetism arising from twisted bilayers of antiferromagnets and also ferromagnets. Despite the fact that the parent materials all exhibit collinear orderings, the bilayer system shows controllable emergent noncollinear spin textures. Time permitting, I will also discuss a theory for the potentially continuous metal-insulator transition with fractionalized electric charges in transition metal dichalcogenide moir\’e heterostructures.
Zoom Link: https://pitp.zoom.us/j/99322296758?pwd=WUNGcE1JS3FpZ1VxbklsSCtYTEJVdz09