arXiv subjects

Simon Apers

Publications and source records attributed to Simon Apers.

36 records · Page 2Linked to original sources

arXiv2023

Quantum Speedup for Graph Sparsification, Cut Approximation and Laplacian Solving

Graph sparsification underlies a large number of algorithms, ranging from approximation algorithms for cut problems to solvers for linear systems in the graph Laplacian. In its strongest form, "spectral sparsification" reduces the number of edges to near-linear in the number of nodes, while approximately preserving the cut and spectral structure of the graph. In this work we demonstrate a polynomial quantum speedup for spectral sparsification and many of its applications. In particular, we give a quantum algorithm that, given a weighted graph with $n$ nodes and $m$ edges, outputs a classical description of an $ε$-spectral sparsifier in sublinear time $\tilde{O}(\sqrt{mn}/ε)$. This contrasts with the optimal classical complexity $\tilde{O}(m)$. We also prove that our quantum algorithm is optimal up to polylog-factors. The algorithm builds on a string of existing results on sparsification, graph spanners, quantum algorithms for shortest paths, and efficient constructions for $k$-wise independent random strings. Our algorithm implies a quantum speedup for solving Laplacian systems and for approximating a range of cut problems such as min cut and sparsest cut.

quant-ph↗

arXiv2022

(No) Quantum space-time tradeoff for USTCON

Undirected $st$-connectivity is important both for its applications in network problems, and for its theoretical connections with logspace complexity. Classically, a long line of work led to a time-space tradeoff of $T=\tilde{O}(n^2/S)$ for any $S$ such that $S=Ω(\log (n))$ and $S=O(n^2/m)$. Surprisingly, we show that quantumly there is no nontrivial time-space tradeoff: there is a quantum algorithm that achieves both optimal time $\tilde{O}(n)$ and space $O(\log (n))$ simultaneously. This improves on previous results, which required either $O(\log (n))$ space and $\tilde{O}(n^{1.5})$ time, or $\tilde{O}(n)$ space and time. To complement this, we show that there is a nontrivial time-space tradeoff when given a lower bound on the spectral gap of a corresponding random walk.

quant-ph↗

arXiv2022

Hamiltonian Monte Carlo for efficient Gaussian sampling: long and random steps

Hamiltonian Monte Carlo (HMC) is a Markov chain algorithm for sampling from a high-dimensional distribution with density $e^{-f(x)}$, given access to the gradient of $f$. A particular case of interest is that of a $d$-dimensional Gaussian distribution with covariance matrix $Σ$, in which case $f(x) = x^\top Σ^{-1} x$. We show that HMC can sample from a distribution that is $\varepsilon$-close in total variation distance using $\widetilde{O}(\sqrtκ d^{1/4} \log(1/\varepsilon))$ gradient queries, where $κ$ is the condition number of $Σ$. Our algorithm uses long and random integration times for the Hamiltonian dynamics. This contrasts with (and was motivated by) recent results that give an $\widetildeΩ(κd^{1/2})$ query lower bound for HMC with fixed integration times, even for the Gaussian case.

stat.ML↗

arXiv2022

Quantum XYZ Product Codes

We study a three-fold variant of the hypergraph product code construction, differing from the standard homological product of three classical codes. When instantiated with 3 classical LDPC codes, this "XYZ product" yields a non CSS quantum LDPC code which might display a large minimum distance. The simplest instance of this construction, corresponding to the product of 3 repetition codes, is a non CSS variant of the 3-dimensional toric code known as the Chamon code. The general construction was introduced in Denise Maurice's PhD thesis, but has remained poorly understood so far. The reason is that while hypergraph product codes can be analyzed with combinatorial tools, the XYZ product codes also depend crucially on the algebraic properties of the parity-check matrices of the three classical codes, making their analysis much more involved. Our main motivation for studying XYZ product codes is that the natural representatives of logical operators are two-dimensional objects. This contrasts with standard hypergraph product codes in 3 dimensions which always admit one-dimensional logical operators. In particular, specific instances of XYZ product codes with constant rate might display a minimum distance as large as $Θ(N^{2/3})$. While we do not prove this result here, we obtain the dimension of a large class of XYZ product codes, and when restricting to codes with dimension 1, we reduce the problem of computing the minimum distance to a more elementary combinatorial problem involving binary 3-tensors. We also discuss in detail some families of XYZ product codes that can be embedded in three dimensions with local interaction. Some of these codes seem to share properties with Haah's cubic codes and might be interesting candidates for self-correcting quantum memories with a logarithmic energy barrier.

quant-ph↗

arXiv2022

Cut query algorithms with star contraction

We study the complexity of determining the edge connectivity of a simple graph with cut queries. We show that (i) there is a bounded-error randomized algorithm that computes edge connectivity with $O(n)$ cut queries, and (ii) there is a bounded-error quantum algorithm that computes edge connectivity with $Õ(\sqrt{n})$ cut queries. We prove these results using a new technique called "star contraction" to randomly contract edges of a graph while preserving non-trivial minimum cuts. In star contraction vertices randomly contract an edge incident on a small set of randomly chosen vertices. In contrast to the related 2-out contraction technique of Ghaffari, Nowicki, and Thorup [SODA'20], star contraction only contracts vertex-disjoint star subgraphs, which allows it to be efficiently implemented via cut queries. The $O(n)$ bound from item (i) was not known even for the simpler problem of connectivity, and improves the $O(n\log^3 n)$ bound by Rubinstein, Schramm, and Weinberg [ITCS'18]. The bound is tight under the reasonable conjecture that the randomized communication complexity of connectivity is $Ω(n\log n)$, an open question since the seminal work of Babai, Frankl, and Simon [FOCS'86]. The bound also excludes using edge connectivity on simple graphs to prove a superlinear randomized query lower bound for minimizing a symmetric submodular function. Item (ii) gives a nearly-quadratic separation with the randomized complexity and addresses an open question of Lee, Santha, and Zhang [SODA'21]. The algorithm can also be viewed as making $Õ(\sqrt{n})$ matrix-vector multiplication queries to the adjacency matrix. Finally, we demonstrate the use of star contraction outside of the cut query setting by designing a one-pass semi-streaming algorithm for computing edge connectivity in the vertex arrival setting. This contrasts with the edge arrival setting where two passes are required.

cs.DS↗

arXiv2021

Quadratic speedup for spatial search by continuous-time quantum walk

Continuous-time quantum walks provide a natural framework to tackle the fundamental problem of finding a node among a set of marked nodes in a graph, known as spatial search. Whether spatial search by continuous-time quantum walk provides a quadratic advantage over classical random walks has been an outstanding problem. Thus far, this advantage is obtained only for specific graphs or when a single node of the underlying graph is marked. In this article, we provide a new continuous-time quantum walk search algorithm that completely resolves this: our algorithm can find a marked node in any graph with any number of marked nodes, in a time that is quadratically faster than classical random walks. The overall algorithm is quite simple, requiring time evolution of the quantum walk Hamiltonian followed by a projective measurement. A key component of our algorithm is a purely analog procedure to perform operations on a state of the form $e^{-tH^2}|ψ\rangle$, for a given Hamiltonian $H$: it only requires evolving $H$ for time scaling as $\sqrt{t}$. This allows us to quadratically fast-forward the dynamics of a continuous-time classical random walk. The applications of our work thus go beyond the realm of quantum walks and can lead to new analog quantum algorithms for preparing ground states of Hamiltonians or solving optimization problems.

quant-ph↗

arXiv2021

Finding the KT partition of a weighted graph in near-linear time

In a breakthrough work, Kawarabayashi and Thorup (J.~ACM'19) gave a near-linear time deterministic algorithm for minimum cut in a simple graph $G = (V,E)$. A key component is finding the $(1+\varepsilon)$-KT partition of $G$, the coarsest partition $\{P_1, \ldots, P_k\}$ of $V$ such that for every non-trivial $(1+\varepsilon)$-near minimum cut with sides $\{S, \bar{S}\}$ it holds that $P_i$ is contained in either $S$ or $\bar{S}$, for $i=1, \ldots, k$. Here we give a near-linear time randomized algorithm to find the $(1+\varepsilon)$-KT partition of a weighted graph. Our algorithm is quite different from that of Kawarabayashi and Thorup and builds on Karger's framework of tree-respecting cuts (J.~ACM'00). We describe applications of the algorithm. (i) The algorithm makes progress towards a more efficient algorithm for constructing the polygon representation of the set of near-minimum cuts in a graph. This is a generalization of the cactus representation initially described by Benczúr (FOCS'95). (ii) We improve the time complexity of a recent quantum algorithm for minimum cut in a simple graph in the adjacency list model from $\widetilde O(n^{3/2})$ to $\widetilde O(\sqrt{mn})$. (iii) We describe a new type of randomized algorithm for minimum cut in simple graphs with complexity $O(m + n \log^6 n)$. For slightly dense graphs this matches the complexity of the current best $O(m + n \log^2 n)$ algorithm which uses a different approach based on random contractions. The key technical contribution of our work is the following. Given a weighted graph $G$ with $m$ edges and a spanning tree $T$, consider the graph $H$ whose nodes are the edges of $T$, and where there is an edge between two nodes of $H$ iff the corresponding 2-respecting cut of $T$ is a non-trivial near-minimum cut of $G$. We give a $O(m \log^4 n)$ time deterministic algorithm to compute a spanning forest of $H$.

cs.DS↗

arXiv2021

Quantum complexity of minimum cut

The minimum cut problem in an undirected and weighted graph $G$ is to find the minimum total weight of a set of edges whose removal disconnects $G$. We completely characterize the quantum query and time complexity of the minimum cut problem in the adjacency matrix model. If $G$ has $n$ vertices and edge weights at least $1$ and at most $τ$, we give a quantum algorithm to solve the minimum cut problem using $\tilde O(n^{3/2}\sqrtτ)$ queries and time. Moreover, for every integer $1 \le τ\le n$ we give an example of a graph $G$ with edge weights $1$ and $τ$ such that solving the minimum cut problem on $G$ requires $Ω(n^{3/2}\sqrtτ)$ many queries to the adjacency matrix of $G$. These results contrast with the classical randomized case where $Ω(n^2)$ queries to the adjacency matrix are needed in the worst case even to decide if an unweighted graph is connected or not. In the adjacency array model, when $G$ has $m$ edges the classical randomized complexity of the minimum cut problem is $\tilde Θ(m)$. We show that the quantum query and time complexity are $\tilde O(\sqrt{mnτ})$ and $\tilde O(\sqrt{mnτ} + n^{3/2})$, respectively, where again the edge weights are between $1$ and $τ$. For dense graphs we give lower bounds on the quantum query complexity of $Ω(n^{3/2})$ for $τ> 1$ and $Ω(τn)$ for any $1 \leq τ\leq n$. Our query algorithm uses a quantum algorithm for graph sparsification by Apers and de Wolf (FOCS 2020) and results on the structure of near-minimum cuts by Kawarabayashi and Thorup (STOC 2015) and Rubinstein, Schramm and Weinberg (ITCS 2018). Our time efficient implementation builds on Karger's tree packing technique (STOC 1996).

quant-ph↗

arXiv2021

Testing properties of signed graphs

In graph property testing the task is to distinguish whether a graph satisfies a given property or is "far" from having that property, preferably with a sublinear query and time complexity. In this work we initiate the study of property testing in signed graphs, where every edge has either a positive or a negative sign. We show that there exist sublinear algorithms for testing three key properties of signed graphs: balance (or 2-clusterability), clusterability and signed triangle freeness. We consider both the dense graph model, where we can query the (signed) adjacency matrix of a signed graph, and the bounded-degree model, where we can query for the neighbors of a node and the sign of the connecting edge. Our algorithms use a variety of tools from graph property testing, as well as reductions from one setting to the other. Our main technical contribution is a sublinear algorithm for testing clusterability in the bounded-degree model. This contrasts with the property of k-clusterability which is not testable with a sublinear number of queries. The tester builds on the seminal work of Goldreich and Ron for testing bipartiteness.

cs.DS↗

arXiv2020

Expansion Testing using Quantum Fast-Forwarding and Seed Sets

Expansion testing aims to decide whether an $n$-node graph has expansion at least $Φ$, or is far from any such graph. We propose a quantum expansion tester with complexity $\widetilde{O}(n^{1/3}Φ^{-1})$. This accelerates the $\widetilde{O}(n^{1/2}Φ^{-2})$ classical tester by Goldreich and Ron [Algorithmica '02], and combines the $\widetilde{O}(n^{1/3}Φ^{-2})$ and $\widetilde{O}(n^{1/2}Φ^{-1})$ quantum speedups by Ambainis, Childs and Liu [RANDOM '11] and Apers and Sarlette [QIC '19], respectively. The latter approach builds on a quantum fast-forwarding scheme, which we improve upon by initially growing a seed set in the graph. To grow this seed set we use a so-called evolving set process from the graph clustering literature, which allows to grow an appropriately local seed set.

quant-ph↗

arXiv2019

A Unified Framework of Quantum Walk Search

The main results on quantum walk search are scattered over different, incomparable frameworks, most notably the hitting time framework, originally by Szegedy, the electric network framework by Belovs, and the MNRS framework by Magniez, Nayak, Roland and Santha. As a result, a number of pieces are currently missing. For instance, the electric network framework allows quantum walks to start from an arbitrary initial state, but it only detects marked elements. In recent work by Ambainis et al., this problem was resolved for the more restricted hitting time framework, in which quantum walks must start from the stationary distribution. We present a new quantum walk search framework that unifies and strengthens these frameworks. This leads to a number of new results. For instance, the new framework not only detects, but finds marked elements in the electric network setting. The new framework also allows one to interpolate between the hitting time framework, which minimizes the number of walk steps, and the MNRS framework, which minimizes the number of times elements are checked for being marked. This allows for a more natural tradeoff between resources. Whereas the original frameworks only rely on quantum walks and phase estimation, our new algorithm makes use of a technique called quantum fast-forwarding, similar to the recent results by Ambainis et al. As a final result we show how in certain cases we can simplify this more involved algorithm to merely applying the quantum walk operator some number of times. This answers an open question of Ambainis et al.

quant-ph↗

arXiv2019

Quantum Walk Sampling by Growing Seed Sets

This work describes a new algorithm for creating a superposition over the edge set of a graph, encoding a quantum sample of the random walk stationary distribution. The algorithm requires a number of quantum walk steps scaling as $\widetilde{O}(m^{1/3} δ^{-1/3})$, with $m$ the number of edges and $δ$ the random walk spectral gap. This improves on existing strategies by initially growing a classical seed set in the graph, from which a quantum walk is then run. The algorithm leads to a number of improvements: (i) it provides a new bound on the setup cost of quantum walk search algorithms, (ii) it yields a new algorithm for $st$-connectivity, and (iii) it allows to create a superposition over the isomorphisms of an $n$-node graph in time $\widetilde{O}(2^{n/3})$, surpassing the $Ω(2^{n/2})$ barrier set by index erasure.

quant-ph↗

arXiv2019

Quantum Fast-Forwarding: Markov Chains and Graph Property Testing

We introduce a new tool for quantum algorithms called quantum fast-forwarding (QFF). The tool uses quantum walks as a means to quadratically fast-forward a reversible Markov chain. More specifically, with $P$ the Markov chain transition matrix and $D = \sqrt{P\circ P^T}$ its discriminant matrix ($D=P$ if $P$ is symmetric), we construct a quantum walk algorithm that for any quantum state $|v\rangle$ and integer $t$ returns a quantum state $ε$-close to the state $D^t|v\rangle/\|D^t|v\rangle\|$. The algorithm uses $O\Big(\|D^t|v\rangle\|^{-1}\sqrt{t\log(ε\|D^t|v\rangle\|)^{-1}}\Big)$ expected quantum walk steps and $O(\|D^t|v\rangle\|^{-1})$ expected reflections around $|v\rangle$. This shows that quantum walks can accelerate the transient dynamics of Markov chains, complementing the line of results that proves the acceleration of their limit behavior. We show that this tool leads to speedups on random walk algorithms in a very natural way. Specifically we consider random walk algorithms for testing the graph expansion and clusterability, and show that we can quadratically improve the dependency of the classical property testers on the random walk runtime. Moreover, our quantum algorithm exponentially improves the space complexity of the classical tester to logarithmic. As a subroutine of independent interest, we use QFF for determining whether a given pair of nodes lies in the same cluster or in separate clusters. This solves a robust version of $s$-$t$ connectivity, relevant in a learning context for classifying objects among a set of examples. The different algorithms crucially rely on the quantum speedup of the transient behavior of random walks.

quant-ph↗

arXiv2017

Simulation of Quantum Walks and Fast Mixing with Classical Processes

We compare discrete-time quantum walks on graphs to their natural classical equivalents, which we argue are lifted Markov chains, that is, classical Markov chains with added memory. We show that these can simulate quantum walks, allowing us to answer an open question on how the graph topology ultimately bounds their mixing performance, and that of any stochastic local evolution. The results highlight that speedups in mixing and transport phenomena are not necessarily diagnostic of quantum effects, although superdiffusive spreading is more prominent with quantum walks.

quant-ph↗

arXiv2017

Bounding the convergence time of local probabilistic evolution

Isoperimetric inequalities form a very intuitive yet powerful characterization of the connectedness of a state space, that has proven successful in obtaining convergence bounds. Since the seventies they form an essential tool in differential geometry, graph theory and Markov chain analysis. In this paper we use isoperimetric inequalities to construct a bound on the convergence time of any local probabilistic evolution that leaves its limit distribution invariant. We illustrate how this general result leads to new bounds on convergence times beyond the explicit Markovian setting, among others on quantum dynamics.

cs.DM↗

arXiv2017

Detecting topological transitions in two dimensions by Hamiltonian evolution

We show that the evolution of two-component particles governed by a two-dimensional spin-orbit lattice Hamiltonian can reveal transitions between topological phases. A kink in the mean width of the particle distribution signals the closing of the band gap, a prerequisite for a quantum phase transition between topological phases. Furthermore, for realistic and experimentally motivated Hamiltonians the density profile in topologically non-trivial phases displays characteristic rings in the vicinity of the origin that are absent in trivial phases. The results are expected to have immediate application to systems of ultracold atoms and photonic lattices.

cond-mat.quant-gas↗

arXiv2017

Lifting Markov Chains To Mix Faster: Limits and Opportunities

Lifted Markov chains are Markov chains on graphs with added local "memory" and can be used to mix towards a target distribution faster than their memoryless counterparts. Upper and lower bounds on the achievable performance have been provided under specific assumptions. In this paper, we analyze which assumptions and constraints are relevant for mixing, and how changing these assumptions affects those bounds. Explicitly, we find that requesting mixing on either the original or the full lifted graph, and allowing for reducible lifted chains or not, have no essential influence on mixing time bounds. On the other hand, allowing for suitable initialization of the lifted dynamics and/or imposing invariance of the target distribution for any initialization do significantly affect the convergence performance. The achievable convergence speed for a lifted chain goes from diameter-time to no acceleration over a standard Markov chain, with conductance bounds limiting the effectiveness of the intermediate cases. In addition, we show that the relevance of imposing ergodic flows depends on the other criteria. The presented analysis allows us to clarify in which scenarios designing lifted dynamics can lead to better mixing, and provide a flexible framework to compare lifted walks with other acceleration methods.

math.OC↗

arXiv2015

Accelerating Consensus by Spectral Clustering and Polynomial Filters

It is known that polynomial filtering can accelerate the convergence towards average consensus on an undirected network. In this paper the gain of a second-order filtering is investigated. A set of graphs is determined for which consensus can be attained in finite time, and a preconditioner is proposed to adapt the undirected weights of any given graph to achieve fastest convergence with the polynomial filter. The corresponding cost function differs from the traditional spectral gap, as it favors grouping the eigenvalues in two clusters. A possible loss of robustness of the polynomial filter is also highlighted.

eess.SY↗