• Corpus ID: 44093237

A faster hafnian formula for complex matrices and its benchmarking on the Titan supercomputer

  title={A faster hafnian formula for complex matrices and its benchmarking on the Titan supercomputer},
  author={Andreas Bj{\"o}rklund and Brajesh Gupt and Nicol{\'a}s Quesada},
We introduce new and simple algorithms for the calculation of the number of perfect matchings of complex weighted, undirected graphs with and without loops. Our compact formulas for the hafnian and loop hafnian of $n \times n $ complex matrices run in $O(n^3 2^{n/2})$ time, are embarrassingly parallelizable and, to the best of our knowledge, are the fastest exact algorithms to compute these quantities. Despite our highly optimized algorithm, numerical benchmarks on the Titan supercomputer with… 

Figures from this paper

Hafnian of two-parameter matrices
The concept of the hafnian first appeared in the works on quantum field theory by E. R. Caianiello. However, it also has an important combinatorial property: the hafnian of the adjacency matrix of an
Measuring the similarity of graphs with a Gaussian boson sampler
This work shows that Gaussian boson samplers have a potential practical application: samples from these devices can be used to construct a feature vector that embeds a graph in Euclidean space, where similarity measures between graphs---so-called graph kernels---can be naturally defined.
Graph isomorphism and Gaussian boson sampling
Abstract We introduce a connection between a near-term quantum computing device, specifically a Gaussian boson sampler, and the graph isomorphism problem. We propose a scheme where graphs are encoded
A quantum hardware-induced graph kernel based on Gaussian Boson Sampling
A feature map and graph similarity measure or ‘graph kernel’ is constructed using samples from the Gaussian Boson Sampler, and it is shown that the kernel performs well compared to standard graph kernels on typical benchmark datasets.
The hafnian of Toeplitz matrices of a special type, perfect matchings and Bessel polynomials
We present a simple and convenient analytical formula for efficient exact computation of the hafnian of Toeplitz matrices of a special type. An interpretation of the obtained results is given in the
Classical benchmarking of Gaussian Boson Sampling on the Titan supercomputer
This work employs a recent exact sampling algorithm for GBS with threshold detectors to perform classical simulations on the Titan supercomputer and indicates that large losses can be tolerated and that the use of threshold detectors is preferable over using photon-number-resolving detectors postselected on collision-free outputs.
Quantum experiments and graphs II: Quantum interference, computation, and state generation
It is shown that one can interpret quantum experiments composed of linear optics and probabilistic sources with graph theory, and an approach to describe state-of-the-art photonic quantum experiments using graph theory is presented, where the quantum states are given by the coherent superpositions of perfect matchings.
Strawberry Fields: A Software Platform for Photonic Quantum Computing
Strawberry Fields, an open-source quantum programming architecture for light-based quantum computers, is introduced, and its key features are detailed.
Linear multiport photonic interferometers: loss analysis of temporally-encoded architectures
It is concluded that implementing a chain of reconfigurable beamsplitters and delay loops on an integrated lithum niobate platform could outperform comparable architectures in the near future.


Counting perfect matchings as fast as Ryser
This work shows that there is a polynomial space algorithm that counts the number of perfect matchings in an n-vertex graph in O*(2n/2) ⊂ O(1.415n) time, and presents it in the more general setting of computing the hafnian over an arbitrary ring analogously to Ryser's algorithm for permanent computation.
Polynomial Time Algorithms to Approximate Permanents and Mixed Discriminants Within a Simply Exponential Factor
  • A. Barvinok
  • Mathematics, Computer Science
    Random Struct. Algorithms
  • 1999
We present real, complex, and quaternionic versions of a simple ran-domized polynomial time algorithm to approximate the permanent of a non-negative matrix and, more generally, the mixed discriminant
Fast Polynomial-Space Algorithms Using Möbius Inversion: Improving on Steiner Tree and Related Problems
The concept of branching walks is introduced and the Inclusion-Exclusion algorithm of Karp is extended for counting Hamiltonian paths and polynomial-space $\mathcal{O}^*(2^n)$ algorithms for several spanning tree and partition problems are obtained.
Hafnians, perfect matchings and Gaussian matrices
It is concluded that Barvinok's estimator gives a polynomial-time algorithm for the approximate (up to subexponential errors) evaluation of the number of perfect matchings.
Faster exponential-time algorithms in graphs of bounded average degree
Exact Algorithms for Exact Satisfiability and Number of Perfect Matchings
This work shows that the Exact Satisfiability problem of size l with m clauses can be solved in time 2mlO(1) and polynomial space, and shows how to count the number of perfect matchings in time O(1.732n) and exponential space.
Approximating permanents and hafnians
We prove that the logarithm of the permanent of an nxn real matrix A and the logarithm of the hafnian of a 2nx2n real symmetric matrix A can be approximated within an additive error 1 > epsilon > 0
Computing Permanents for Boson Sampling on Tianhe-2 Supercomputer
The permanent of the largest matrix is computed using up to 312,000 CPU cores of Tianhe-2, and it is inferred from the current most efficient permanent-computing algorithms that an upper bound on the performance of Tiane-2 is one 50-photon sample per ~100 min.
Gaussian boson sampling for perfect matchings of arbitrary graphs
With the proposed method, a Gaussian Boson Sampling device can be used to estimate the number of perfect matchings significantly faster and with lower energy consumption compared to a classical computer.
Alternative Algorithms for Counting All Matchings in Graphs
Two new methods for counting all matchings in a graph are presented, one of which is a generalization of a Godman-Godsil estimator and the other uses importance sampling.