Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2609.40321

Exponential quantum speedup for $\mathbb{F}_3^n$-Subset-Sum? Or, rigorous classical algorithms for Binary-Error LWE

Robin Kothari, Tony Metger, Ryan O'Donnell, Noah Shutty, Kewen Wu

quant-phcs.CRcs.DSSubmitted Sep 30, 2026• Announced Oct 1, 2026• 1 citation
arXivHTMLPDF

Abstract

We study vector subset sum over $\mathbb{F}_3^n$: given $m$ random vectors from $\mathbb{F}_3^n$, find a nonempty subset that sums to zero; the smaller $m$, the more difficult it is to find such a subset. Chen, Liu, and Zhandry (EUROCRYPT'22) introduced an efficient quantum algorithm that solves this problem when $m\approx n^2/2$, where a naive classical algorithm would require exponential time. Subsequently, Kothari, O'Donnell, and Wu (STOC'2026) gave an efficient classical algorithm that only requires $m \approx n^2/3$ vectors, thus removing the hope for an exponential quantum advantage in this parameter regime. Using the framework of Chen, Liu, and Zhandry, we give quantum algorithms that require much fewer input vectors, renewing the possibility of an exponential quantum speedup: for any fixed $ε>0$, our quantum algorithm solves $\mathbb{F}_3$-subset sum in polynomial time with $m=ε\cdot n^2$ vectors. More generally, we establish a full sample--time tradeoff that interpolates between exponential and polynomial runtime. The main ingredient is a deterministic classical algorithm for the binary-error Learning-with-Errors problem, which is of independent cryptographic interest. For this, we rigorously establish a sample--time tradeoff that was predicted by earlier algebraic heuristics. For vector subset sums over larger fields, we also significantly improve classical algorithms in Kothari, O'Donnell, and Wu (STOC'2026).

AI Summary

The paper studies average-case vector subset sum over $\mathbb F_3^n$: given $m$ random vectors, find a nonempty subset summing to zero. It asks whether this explicit-input search problem supports an exponential quantum advantage after classical algorithms reached the previously proposed quantum regime, $m\approx n^2/3$. Its main quantum result gives runtime $(m/n)^{O_\kappa(\lceil n^2/m\rceil)}$ and failure probability $2^{-\Omega_\kappa(m)}$ for every $m\ge\kappa n$, with fixed $\kappa>\log_2 3$. Consequently, polynomial time suffices for $m=\epsilon n^2$ for every fixed $\epsilon>0$; intermediate sample counts yield quasipolynomial and subexponential runtimes. The new ingredient is a deterministic classical binary-error LWE decoder over any fixed prime field $\mathbb F_q$, $q\ge3$, with the same tradeoff above the uniqueness threshold $m>\log_{q/2}(q)n$. Recovery holds for each fixed secret and binary error vector, with probability over the random matrix. The analysis encodes decoding as quadratic equations with field equations, proves degree-$D$ surjectivity for $D=O(n^2/m)$ using symbolic-rank bounds, algebraic shifting, multiset shadows and martingale concentration, then computes the quotient algebra by linear algebra. An existing coherent quantum reduction transfers this decoder to subset sum. Additional classical results reduce larger-field average-case sample requirements to $O_q(n^{1+\lfloor(q+3)/4\rfloor})$ and worst-case requirements to $O_q(n^{q-1})$. No classical lower bound establishes an exponential separation. A claimed classical $\mathbb F_3$ threshold of approximately $0.1022n^2$, obtained through LLM-assisted gadget optimization, is stated without a complete algorithm or proof; practical cryptographic attacks are not established.

Little sign of AI-generated textAI use disclosedOverclaims
We note that this value of $\eps$ is the result of an LLM-assisted optimization of existing classical algorithmic ideas.

Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.

Related papers

On learning linear functions from subset and its applications in quantum computing

Gábor Ivanyos, Anupam Prakash, Miklós Sántha

Let $\mathbb{F}_q$ be the finite field of size $q$ and let $\ell: \mathbb{F}_q^n \to \mathbb{F}_q$ be a linear function. We introduce the {\em Learning From Subset} problem LFS$(q,n,d)$ of learning $\ell$, given samples $u \in \mathbb{F}_q^n$ from a special distribution depending on $\ell$: the probability of sampling $u$ is a function of $\ell(u)$ and is non zero for at most $d$ values of $\ell(u)$. We provide a randomized algorithm for LFS$(q,n,d)$ with sample complexity $(n+d)^{O(d)}$ and running time polynomial in $\log q$ and $(n+d)^{O(d)}$. Our algorithm generalizes and improves upon previous results \cite{Friedl, Ivanyos} that had provided algorithms for LFS$(q,n,q-1)$ with running time $(n+q)^{O(q)}$. We further present applications of our result to the {\em Hidden Multiple Shift} problem HMS$(q,n,r)$ in quantum computation where the goal is to determine the hidden shift $s$ given oracle access to $r$ shifted copies of an injective function $f: \mathbb{Z}_q^n \to \{0, 1\}^l$, that is we can make queries of the form $f_s(x,h) = f(x-hs)$ where $h$ can assume $r$ possible values. We reduce HMS$(q,n,r)$ to LFS$(q,n, q-r+1)$ to obtain a polynomial time algorithm for HMS$(q,n,r)$ when $q=n^{O(1)}$ is prime and $q-r=O(1)$. The best known algorithms \cite{CD07, Friedl} for HMS$(q,n,r)$ with these parameters require exponential time.

arXiv:1806.09660•Jun 25, 2018•4 citations

Towards An Implementation of the Subset-sum Problem on the IBM Quantum Experience

David Gunter, Toks Adedoyin

In seeking out an algorithm to test out the capability of the IBM Quantum Experience quantum computer, we were given a review paper covering various algorithms for solving the subset-sum problem, including both classical and quantum algorithms. The paper went on to present a novel algorithm that beat the previous best algorithm known at the time. The complex nature of the algorithm made it difficult to see a path for implementation on the Quantum Experience machine and the exponential cost - only slightly better than the best classical algorithm - left us looking for a different approach for solving this problem. We present here a new quantum algorithm for solving the subset-sum problem that for many cases should lead to O(poly(n))-time to solution. The work is reminiscent of the verification procedure used in a polynomial-time algorithm for the quantum Arthur-Merlin games presented elsewhere, where the use of a quantum binary search to find a maximum eigenvalue in the final output stage has been adapted to the subset-sum problem as in another paper.

arXiv:1912.03254•Dec 4, 2019•1 citation

An Exponential Separation Between Quantum and Quantum-Inspired Classical Algorithms for Linear Systems

Allan Grønlund, Kasper Green Larsen

Achieving a provable exponential quantum speedup for an important machine learning task has been a central research goal since the seminal HHL quantum algorithm for solving linear systems and the subsequent quantum recommender systems algorithm by Kerenidis and Prakash. These algorithms were initially believed to be strong candidates for exponential speedups, but a lower bound ruling out similar classical improvements remained absent. In breakthrough work by Tang, it was demonstrated that this lack of progress in classical lower bounds was for good reasons. Concretely, she gave a classical counterpart of the quantum recommender systems algorithm, reducing the quantum advantage to a mere polynomial. Her approach is quite general and was named quantum-inspired classical algorithms. Since then, almost all the initially exponential quantum machine learning speedups have been reduced to polynomial via new quantum-inspired classical algorithms. From the current state-of-affairs, it is unclear whether we can hope for exponential quantum speedups for any natural machine learning task. In this work, we present the first such provable exponential separation between quantum and quantum-inspired classical algorithms for the basic problem of solving a linear system when the input matrix is well-conditioned and has sparse rows and columns.

arXiv:2411.02087•Nov 4, 2024•4 citations

No Exponential Quantum Speedup for SIS∞ Anymore

Robin Kothari, Ryan O’Donnell, Kewen Wu

In 2021, Chen, Liu, and Zhandry presented an efficient quantum algorithm for the average-case ℓ∞-Short Integer Solution (SIS∞) problem, in a parameter range outside the normal range of cryptographic interest, but still with no known efficient classical algorithm. This was particularly exciting since SIS∞ is a simple problem without structure, and their algorithmic techniques were different from those used in prior exponential quantum speedups.

arXiv:2510.07515•ACM Symposium on Theory of Computing (STOC)•Jun 9, 2026•13 citations

On learning linear functions from subset and its applications in quantum computing

Gábor Ivanyos, Anupam Prakash, Miklós Sántha

Let F_{q} be the finite field of size q and let l: F_{q}^{n} -> F_{q} be a linear function. We introduce the Learning From Subset problem LFS(q,n,d) of learning l, given samples u in F_{q}^{n} from a special distribution depending on l: the probability of sampling u is a function of l(u) and is non zero for at most d values of l(u). We provide a randomized algorithm for LFS(q,n,d) with sample complexity (n+d)^{O(d)} and running time polynomial in log q and (n+d)^{O(d)}. Our algorithm generalizes and improves upon previous results [Friedl et al., 2014; Gabor Ivanyos, 2008] that had provided algorithms for LFS(q,n,q-1) with running time (n+q)^{O(q)}. We further present applications of our result to the Hidden Multiple Shift problem HMS(q,n,r) in quantum computation where the goal is to determine the hidden shift s given oracle access to r shifted copies of an injective function f: Z_{q}^{n} -> {0, 1}^{l}, that is we can make queries of the form f_{s}(x,h) = f(x-hs) where h can assume r possible values. We reduce HMS(q,n,r) to LFS(q,n, q-r+1) to obtain a polynomial time algorithm for HMS(q,n,r) when q=n^{O(1)} is prime and q-r=O(1). The best known algorithms [Andrew M. Childs and Wim van Dam, 2007; Friedl et al., 2014] for HMS(q,n,r) with these parameters require exponential time.

•Jan 1, 2018•0 citations

Tangible Quantum Speedup in Learning-With-Errors Problem

Wooyeong Song, Youngrong Lim, Kabgyun Jeong +4 more

Very recently, one of the most biggest agenda issues is to provide the proof of quantum computational speedup, particularly with a prospect for near-term uses. However, many quantum algorithms are beyond the reach of noisy intermediate-scale quantum (NISQ) realization. This is mainly because of the requirement of excessively large superposition and massive quantum circuit. Hence, we propose a (say) ``NISQ-compatible'' algorithm for one of the crucial problems in computation and modern cryptography, the learning-with-errors (LWE) problem. We base an approach on the divide-and-conquer, wherein a large core process is subdivided into smaller subprocesses. For a specific error distribution and problem condition, it is shown that the proposed quantum LWE algorithm allows the use of exponentially less-superposed quantum samples and operation overheads are reduced, while achieving polynomial sample complexity.

•Aug 17, 2019•0 citations
Exponential quantum speedup for F3n\mathbb{F}_3^nF3n​-Subset-Sum? Or, rigorous classical algorithms for Binary-Error LWE
We study vector subset sum over F3n\mathbb{F}_3^nF3n​: given mmm random vectors from F3n\mathbb{F}_3^nF3n​, find a nonempty subset that sums to zero; the smaller mmm, the more difficult it is to find such a subset. Chen, Liu, and Zhandry (EUROCRYPT'22) introduced an efficient quantum algorithm that solves this problem when m≈n2/2m\approx n^2/2m≈n2/2, where a naive classical algorithm would require exponential time. Subsequently, Kothari, O'Donnell, and Wu (STOC'2026) gave an efficient classical algorithm that only requires m≈n2/3m \approx n^2/3m≈n2/3 vectors, thus removing the hope for an exponential quantum advantage in this parameter regime.
Using the framework of Chen, Liu, and Zhandry, we give quantum algorithms that require much fewer input vectors, renewing the possibility of an exponential quantum speedup: for any fixed ε>0ε>0ε>0, our quantum algorithm solves F3\mathbb{F}_3F3​-subset sum in polynomial time with m=ε⋅n2m=ε\cdot n^2m=ε⋅n2 vectors. More generally, we establish a full sample--time tradeoff that interpolates between exponential and polynomial runtime.
The main ingredient is a deterministic classical algorithm for the binary-error Learning-with-Errors problem, which is of independent cryptographic interest. For this, we rigorously establish a sample--time tradeoff that was predicted by earlier algebraic heuristics. For vector subset sums over larger fields, we also significantly improve classical algorithms in Kothari, O'Donnell, and Wu (STOC'2026).
The paper studies average-case vector subset sum over F3n\mathbb F_3^nF3n​: given mmm random vectors, find a nonempty subset summing to zero. It asks whether this explicit-input search problem supports an exponential quantum advantage after classical algorithms reached the previously proposed quantum regime, m≈n2/3m\approx n^2/3m≈n2/3.
Its main quantum result gives runtime (m/n)Oκ(⌈n2/m⌉)(m/n)^{O_\kappa(\lceil n^2/m\rceil)}(m/n)Oκ​(⌈n2/m⌉) and failure probability 2−Ωκ(m)2^{-\Omega_\kappa(m)}2−Ωκ​(m) for every m≥κnm\ge\kappa nm≥κn, with fixed κ>log⁡23\kappa>\log_2 3κ>log2​3. Consequently, polynomial time suffices for m=ϵn2m=\epsilon n^2m=ϵn2 for every fixed ϵ>0\epsilon>0ϵ>0; intermediate sample counts yield quasipolynomial and subexponential runtimes.
The new ingredient is a deterministic classical binary-error LWE decoder over any fixed prime field Fq\mathbb F_qFq​, q≥3q\ge3q≥3, with the same tradeoff above the uniqueness threshold m>log⁡q/2(q)nm>\log_{q/2}(q)nm>logq/2​(q)n. Recovery holds for each fixed secret and binary error vector, with probability over the random matrix. The analysis encodes decoding as quadratic equations with field equations, proves degree-DDD surjectivity for D=O(n2/m)D=O(n^2/m)D=O(n2/m) using symbolic-rank bounds, algebraic shifting, multiset shadows and martingale concentration, then computes the quotient algebra by linear algebra. An existing coherent quantum reduction transfers this decoder to subset sum.
Additional classical results reduce larger-field average-case sample requirements to Oq(n1+⌊(q+3)/4⌋)O_q(n^{1+\lfloor(q+3)/4\rfloor})Oq​(n1+⌊(q+3)/4⌋) and worst-case requirements to Oq(nq−1)O_q(n^{q-1})Oq​(nq−1). No classical lower bound establishes an exponential separation. A claimed classical F3\mathbb F_3F3​ threshold of approximately 0.1022n20.1022n^20.1022n2, obtained through LLM-assisted gadget optimization, is stated without a complete algorithm or proof; practical cryptographic attacks are not established.
Let Fq\mathbb{F}_qFq​ be the finite field of size qqq and let ℓ:Fqn→Fq\ell: \mathbb{F}_q^n \to \mathbb{F}_qℓ:Fqn​→Fq​ be a linear function. We introduce the {\em Learning From Subset} problem LFS(q,n,d)(q,n,d)(q,n,d) of learning ℓ\ellℓ, given samples u∈Fqnu \in \mathbb{F}_q^nu∈Fqn​ from a special distribution depending on ℓ\ellℓ: the probability of sampling uuu is a function of ℓ(u)\ell(u)ℓ(u) and is non zero for at most ddd values of ℓ(u)\ell(u)ℓ(u). We provide a randomized algorithm for LFS(q,n,d)(q,n,d)(q,n,d) with sample complexity (n+d)O(d)(n+d)^{O(d)}(n+d)O(d) and running time polynomial in log⁡q\log qlogq and (n+d)O(d)(n+d)^{O(d)}(n+d)O(d). Our algorithm generalizes and improves upon previous results \cite{Friedl, Ivanyos} that had provided algorithms for LFS(q,n,q−1)(q,n,q-1)(q,n,q−1) with running time (n+q)O(q)(n+q)^{O(q)}(n+q)O(q). We further present applications of our result to the {\em Hidden Multiple Shift} problem HMS(q,n,r)(q,n,r)(q,n,r) in quantum computation where the goal is to determine the hidden shift sss given oracle access to rrr shifted copies of an injective function f:Zqn→{0,1}lf: \mathbb{Z}_q^n \to \{0, 1\}^lf:Zqn​→{0,1}l, that is we can make queries of the form fs(x,h)=f(x−hs)f_s(x,h) = f(x-hs)fs​(x,h)=f(x−hs) where hhh can assume rrr possible values. We reduce HMS(q,n,r)(q,n,r)(q,n,r) to LFS(q,n,q−r+1)(q,n, q-r+1)(q,n,q−r+1) to obtain a polynomial time algorithm for HMS(q,n,r)(q,n,r)(q,n,r) when q=nO(1)q=n^{O(1)}q=nO(1) is prime and q−r=O(1)q-r=O(1)q−r=O(1). The best known algorithms \cite{CD07, Friedl} for HMS(q,n,r)(q,n,r)(q,n,r) with these parameters require exponential time.
In seeking out an algorithm to test out the capability of the IBM Quantum Experience quantum computer, we were given a review paper covering various algorithms for solving the subset-sum problem, including both classical and quantum algorithms. The paper went on to present a novel algorithm that beat the previous best algorithm known at the time. The complex nature of the algorithm made it difficult to see a path for implementation on the Quantum Experience machine and the exponential cost - only slightly better than the best classical algorithm - left us looking for a different approach for solving this problem. We present here a new quantum algorithm for solving the subset-sum problem that for many cases should lead to O(poly(n))-time to solution. The work is reminiscent of the verification procedure used in a polynomial-time algorithm for the quantum Arthur-Merlin games presented elsewhere, where the use of a quantum binary search to find a maximum eigenvalue in the final output stage has been adapted to the subset-sum problem as in another paper.
Achieving a provable exponential quantum speedup for an important machine learning task has been a central research goal since the seminal HHL quantum algorithm for solving linear systems and the subsequent quantum recommender systems algorithm by Kerenidis and Prakash. These algorithms were initially believed to be strong candidates for exponential speedups, but a lower bound ruling out similar classical improvements remained absent. In breakthrough work by Tang, it was demonstrated that this lack of progress in classical lower bounds was for good reasons. Concretely, she gave a classical counterpart of the quantum recommender systems algorithm, reducing the quantum advantage to a mere polynomial. Her approach is quite general and was named quantum-inspired classical algorithms. Since then, almost all the initially exponential quantum machine learning speedups have been reduced to polynomial via new quantum-inspired classical algorithms. From the current state-of-affairs, it is unclear whether we can hope for exponential quantum speedups for any natural machine learning task. In this work, we present the first such provable exponential separation between quantum and quantum-inspired classical algorithms for the basic problem of solving a linear system when the input matrix is well-conditioned and has sparse rows and columns.
Let F_{q} be the finite field of size q and let l: F_{q}^{n} -> F_{q} be a linear function. We introduce the Learning From Subset problem LFS(q,n,d) of learning l, given samples u in F_{q}^{n} from a special distribution depending on l: the probability of sampling u is a function of l(u) and is non zero for at most d values of l(u). We provide a randomized algorithm for LFS(q,n,d) with sample complexity (n+d)^{O(d)} and running time polynomial in log q and (n+d)^{O(d)}. Our algorithm generalizes and improves upon previous results [Friedl et al., 2014; Gabor Ivanyos, 2008] that had provided algorithms for LFS(q,n,q-1) with running time (n+q)^{O(q)}. We further present applications of our result to the Hidden Multiple Shift problem HMS(q,n,r) in quantum computation where the goal is to determine the hidden shift s given oracle access to r shifted copies of an injective function f: Z_{q}^{n} -> {0, 1}^{l}, that is we can make queries of the form f_{s}(x,h) = f(x-hs) where h can assume r possible values. We reduce HMS(q,n,r) to LFS(q,n, q-r+1) to obtain a polynomial time algorithm for HMS(q,n,r) when q=n^{O(1)} is prime and q-r=O(1). The best known algorithms [Andrew M. Childs and Wim van Dam, 2007; Friedl et al., 2014] for HMS(q,n,r) with these parameters require exponential time.
Very recently, one of the most biggest agenda issues is to provide the proof of quantum computational speedup, particularly with a prospect for near-term uses. However, many quantum algorithms are beyond the reach of noisy intermediate-scale quantum (NISQ) realization. This is mainly because of the requirement of excessively large superposition and massive quantum circuit. Hence, we propose a (say) ``NISQ-compatible'' algorithm for one of the crucial problems in computation and modern cryptography, the learning-with-errors (LWE) problem. We base an approach on the divide-and-conquer, wherein a large core process is subdivided into smaller subprocesses. For a specific error distribution and problem condition, it is shown that the proposed quantum LWE algorithm allows the use of exponentially less-superposed quantum samples and operation overheads are reduced, while achieving polynomial sample complexity.