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:0907.2833 · Physical Review A

Learning and testing algorithms for the Clifford group

Richard A. Low

Published Nov 11, 2009• 42 citations
arXivHTMLPDF

Abstract

Given oracle access to an unknown unitary $C$ from the Clifford group and its conjugate, we give an exact algorithm for identifying $C$ with $O(n)$ queries, which we prove is optimal. We then extend this to all levels of the Gottesman-Chuang hierarchy (also known as the ${\mathcal{C}}_{k}$ hierarchy). Further, for unitaries not in the hierarchy itself but known to be close to an element of the hierarchy, we give a method of finding this close element. We also present a Clifford testing algorithm that decides whether a given black-box unitary is close to an element of the Clifford group or far from every element of the Clifford group.

DOI
10.1103/physreva.80.052314

Related papers

Clifford Testing: Algorithms and Lower Bounds

Marcel Hinsche, Zongbo Bao, Philippe van Dordrecht +3 more

We consider the problem of Clifford testing, which asks whether a black-box n-qubit unitary is a Clifford unitary or at least ε-far from every Clifford unitary. We give the first 4-query Clifford tester, which decides this problem with probability poly(ε). This contrasts with the minimum of 6 copies required for the closely-related task of stabilizer testing. We show that our tester is tolerant, by adapting techniques from tolerant stabilizer testing to our setting. In doing so, we settle in the positive a conjecture of Bu, Gu and Jaffe, by proving a polynomial inverse theorem for a non-commutative Gowers 3-uniformity norm. We also consider the restricted setting of single-copy access, where we give an O(n)-query Clifford tester that requires no auxiliary memory qubits or adaptivity. We complement this with a lower bound, proving that any such, potentially adaptive, single-copy algorithm needs at least Ω(n1/4) queries. To obtain our results, we leverage the structure of the commutant of the Clifford group, obtaining several technical statements that may be of independent interest.

arXiv:2510.07164•ACM Symposium on Theory of Computing (STOC)•Jun 9, 2026•5 citations

The Generalized Semi-Clifford Conjecture Holds at Level 4

Maxwell Marcus, Sathyawageeswar Subramanian, Marcel Dall'Agnol

The Clifford hierarchy $\mathsf{C}_1 \subset \mathsf{C}_2 \subset \cdots$ was introduced by Gottesman and Chuang (arXiv:quant-ph/9908010) to characterize gates that admit fault-tolerant implementation by gate teleportation. Yet, despite its rich mathematical structure and the attention it has received in recent years, little is known about $\mathsf{C}_k$ for $k > 3$. Most progress has focused on identifying structural properties of restrictions of the hierarchy, such as diagonal gates and gates on systems of small dimension $d$ or with few qudits. The generalized semi-Clifford conjecture, proposed by Zeng et al. (arXiv:0712.2084), states that every gate in $\cup_k \mathsf{C}_k$ is, up to multiplication by Cliffords, the product of a permutation and a diagonal matrix. Beigi and Shor proved the case $d=2$, $k=3$ (arXiv:0810.5108) and Pllaha et al. found an alternative proof by exploiting fixed points of the conjugation map induced by (a Clifford correction of) $U \in \mathsf{C}_3$ on the span of maximal stabilizer subgroups (arXiv:2006.14040). By extending their fixed-point arguments to the group $Γ_1(U)$ generated by $U \mathsf{P} U^\dagger$ and beyond, we prove the conjecture for $k \leq 4$ and any prime dimension $d$. Our proof centers on conjugation groups $Γ_1(U), Γ_2(U), \ldots$ of $U \in \mathsf{C}_k$, which we expect to be a useful tool in the study of the Clifford hierarchy more generally. We also show a natural sufficient condition on such groups for gates in higher levels to be generalized semi-Clifford.

quant-pharXiv:2609.38751•Oct 1, 2026•1 citation

Unitary Subgroup Testing

Zvika Brakerski, Devika Sharma, Guy Weissenberg

We consider the problem of $\textit{subgroup testing}$ for a quantum circuit $C$: given access to $C$, determine whether it implements a unitary that is $a$-close or $b$-far from a subgroup $\mathcal{G}$ of the unitary group. It encompasses the problem of exact testing, property testing and tolerant testing. In this work, we study these problems with the group $\mathcal{G}$ as the trivial subgroup (i.e. identity testing) or the Pauli or Clifford group and their $q$-ary extension, and a $\textit{promise}$ version of these problems where $C$ is promised to be in some subgroup of the unitaries that contains $\mathcal{G}$ (e.g. identity testing for Clifford circuits). Our main result is an equivalence between Pauli testing, Clifford testing and Identity testing. We derive the equivalence between Clifford and Identity testing by showing a structural property of the Clifford unitaries. Namely, that their (normalized) trace lies in the discrete set $\{2^{-k/2}: k \in \mathbb{N}\} \cup \{0\}$, regardless of the dimension. We also state and prove the analogous property for the $q$-ary Cliffords. This result allows us to analyze a very simple single-query identity test under the Clifford/Pauli promise. To prove the equivalence between Pauli and Identity testing, we analyze the conjugation action of a non-Pauli unitary on the Pauli group and show that its distance from the Pauli group affects the number of fixed points. We believe that these results are of interest, independent of their application to establish the equivalences. We use the equivalences to compare (and thus establish) computational hardness for the problems of Pauli and Clifford testing.

arXiv:2104.03591•Apr 8, 2021•8 citations

Learning Clifford-structured quantum unitaries and Hamiltonians

Arkopal Dutt, Dale Jacobs, John Jeang +2 more

Learning algorithms for structured quantum unitaries and Hamiltonians have primarily considered classes of processes that are local or sparse in the Pauli basis. We turn our attention to learning $n$-qubit quantum unitaries $U$ and Hamiltonians $H$, given query access to $U$ or the unitary evolution of $H$, that may be dense in the Pauli basis but still admit concise Clifford decompositions. Specifically, we consider unitaries (or Hamiltonians) of the form $U = \sum_i α_i C_i$ over Cliffords $C_i$ with bounded Clifford extent $\sum_i |α_i|$. To extract this Clifford structure, we introduce an agnostic tomography protocol for Clifford unitaries that given query access to an unknown unitary $U$ with optimal Clifford fidelity $\textsf{opt}$, outputs a Clifford unitary witnessing fidelity $\geq \textsf{opt} - \varepsilon$ for some error $\varepsilon > 0$, in time $\textsf{poly}(n,(1/\varepsilon)^{\log(1/\varepsilon)})$. We then apply this protocol to obtain tomography protocols for unitaries and Hamiltonians that have bounded Clifford extent. This extends learnability of Hamiltonians from those with sparse Pauli decompositions to those that are dense (i.e., has sparsity $Ω(2^n)$) in the Pauli basis but are Clifford structured.

quant-pharXiv:2608.09912•Aug 11, 2026•5 citations

Learning Gaussian Operations and the Matchgate Hierarchy

Josh Cudby, Sergii Strelchuk

Learning an unknown quantum process is a central task for validation of the functioning of near-term devices. The task is generally hard, requiring exponentially many measurements if no prior assumptions are made on the process. However, an interesting feature of the classically-simulable Clifford group is that unknown Clif-ford operations may be efficiently determined from a black-box implementation. We extend this result to the important class of fermionic Gaussian operations. These operations have received much attention due to their close links to fermionic linear optics. We then introduce an infinite family of unitary gates, called the Matchgate Hierarchy, with a similar structure to the Clifford Hierarchy. We show that the Clifford Hierarchy is contained within the Matchgate Hierarchy and how operations at any level of the hierarchy can be efficiently learned.

arXiv:2407.12649•IEEE International Conference on Quantum Computing and Engineering (QCE)•Sep 15, 2024•9 citations

On Clifford hierarchy testing and near-extremizers of noncommutative uniformity norms

Zongbo, Bao, Jop Briët +3 more

We consider the problem of testing whether an unknown unitary is close to a specified level of the Clifford hierarchy. Bu, Gu, and Jaffe proposed a candidate tester for this task based on a connection with noncommutative analogues of the Gowers uniformity norms. The complexity of this tester -- whose analysis depends on a robust characterization of the near-extremizers of these norms -- was left open. We establish such a characterization for the fourth noncommutative uniformity norm and, as a consequence, obtain an efficient tester for the third level of the Clifford hierarchy. We further discuss possible routes toward resolving the problem of testing for all higher levels, highlighting the main barriers that remain.

quant-pharXiv:2605.26983•May 27, 2026•4 citations
Given oracle access to an unknown unitary CCC from the Clifford group and its conjugate, we give an exact algorithm for identifying CCC with O(n)O(n)O(n) queries, which we prove is optimal. We then extend this to all levels of the Gottesman-Chuang hierarchy (also known as the Ck{\mathcal{C}}_{k}Ck​ hierarchy). Further, for unitaries not in the hierarchy itself but known to be close to an element of the hierarchy, we give a method of finding this close element. We also present a Clifford testing algorithm that decides whether a given black-box unitary is close to an element of the Clifford group or far from every element of the Clifford group.
We consider the problem of Clifford testing, which asks whether a black-box n-qubit unitary is a Clifford unitary or at least ε-far from every Clifford unitary. We give the first 4-query Clifford tester, which decides this problem with probability poly(ε). This contrasts with the minimum of 6 copies required for the closely-related task of stabilizer testing. We show that our tester is tolerant, by adapting techniques from tolerant stabilizer testing to our setting. In doing so, we settle in the positive a conjecture of Bu, Gu and Jaffe, by proving a polynomial inverse theorem for a non-commutative Gowers 3-uniformity norm. We also consider the restricted setting of single-copy access, where we give an O(n)-query Clifford tester that requires no auxiliary memory qubits or adaptivity. We complement this with a lower bound, proving that any such, potentially adaptive, single-copy algorithm needs at least Ω(n1/4) queries. To obtain our results, we leverage the structure of the commutant of the Clifford group, obtaining several technical statements that may be of independent interest.
The Clifford hierarchy C1⊂C2⊂⋯\mathsf{C}_1 \subset \mathsf{C}_2 \subset \cdotsC1​⊂C2​⊂⋯ was introduced by Gottesman and Chuang (arXiv:quant-ph/9908010) to characterize gates that admit fault-tolerant implementation by gate teleportation. Yet, despite its rich mathematical structure and the attention it has received in recent years, little is known about Ck\mathsf{C}_kCk​ for k>3k > 3k>3. Most progress has focused on identifying structural properties of restrictions of the hierarchy, such as diagonal gates and gates on systems of small dimension ddd or with few qudits. The generalized semi-Clifford conjecture, proposed by Zeng et al. (arXiv:0712.2084), states that every gate in ∪kCk\cup_k \mathsf{C}_k∪k​Ck​ is, up to multiplication by Cliffords, the product of a permutation and a diagonal matrix. Beigi and Shor proved the case d=2d=2d=2, k=3k=3k=3 (arXiv:0810.5108) and Pllaha et al. found an alternative proof by exploiting fixed points of the conjugation map induced by (a Clifford correction of) U∈C3U \in \mathsf{C}_3U∈C3​ on the span of maximal stabilizer subgroups (arXiv:2006.14040). By extending their fixed-point arguments to the group Γ1(U)Γ_1(U)Γ1​(U) generated by UPU†U \mathsf{P} U^\daggerUPU† and beyond, we prove the conjecture for k≤4k \leq 4k≤4 and any prime dimension ddd. Our proof centers on conjugation groups Γ1(U),Γ2(U),…Γ_1(U), Γ_2(U), \ldotsΓ1​(U),Γ2​(U),… of U∈CkU \in \mathsf{C}_kU∈Ck​, which we expect to be a useful tool in the study of the Clifford hierarchy more generally. We also show a natural sufficient condition on such groups for gates in higher levels to be generalized semi-Clifford.
We consider the problem of subgroup testing\textit{subgroup testing}subgroup testing for a quantum circuit CCC: given access to CCC, determine whether it implements a unitary that is aaa-close or bbb-far from a subgroup G\mathcal{G}G of the unitary group. It encompasses the problem of exact testing, property testing and tolerant testing. In this work, we study these problems with the group G\mathcal{G}G as the trivial subgroup (i.e. identity testing) or the Pauli or Clifford group and their qqq-ary extension, and a promise\textit{promise}promise version of these problems where CCC is promised to be in some subgroup of the unitaries that contains G\mathcal{G}G (e.g. identity testing for Clifford circuits). Our main result is an equivalence between Pauli testing, Clifford testing and Identity testing. We derive the equivalence between Clifford and Identity testing by showing a structural property of the Clifford unitaries. Namely, that their (normalized) trace lies in the discrete set {2−k/2:k∈N}∪{0}\{2^{-k/2}: k \in \mathbb{N}\} \cup \{0\}{2−k/2:k∈N}∪{0}, regardless of the dimension. We also state and prove the analogous property for the qqq-ary Cliffords. This result allows us to analyze a very simple single-query identity test under the Clifford/Pauli promise. To prove the equivalence between Pauli and Identity testing, we analyze the conjugation action of a non-Pauli unitary on the Pauli group and show that its distance from the Pauli group affects the number of fixed points. We believe that these results are of interest, independent of their application to establish the equivalences. We use the equivalences to compare (and thus establish) computational hardness for the problems of Pauli and Clifford testing.
Learning algorithms for structured quantum unitaries and Hamiltonians have primarily considered classes of processes that are local or sparse in the Pauli basis. We turn our attention to learning nnn-qubit quantum unitaries UUU and Hamiltonians HHH, given query access to UUU or the unitary evolution of HHH, that may be dense in the Pauli basis but still admit concise Clifford decompositions. Specifically, we consider unitaries (or Hamiltonians) of the form U=∑iαiCiU = \sum_i α_i C_iU=∑i​αi​Ci​ over Cliffords CiC_iCi​ with bounded Clifford extent ∑i∣αi∣\sum_i |α_i|∑i​∣αi​∣. To extract this Clifford structure, we introduce an agnostic tomography protocol for Clifford unitaries that given query access to an unknown unitary UUU with optimal Clifford fidelity opt\textsf{opt}opt, outputs a Clifford unitary witnessing fidelity ≥opt−ε\geq \textsf{opt} - \varepsilon≥opt−ε for some error ε>0\varepsilon > 0ε>0, in time poly(n,(1/ε)log⁡(1/ε))\textsf{poly}(n,(1/\varepsilon)^{\log(1/\varepsilon)})poly(n,(1/ε)log(1/ε)). We then apply this protocol to obtain tomography protocols for unitaries and Hamiltonians that have bounded Clifford extent. This extends learnability of Hamiltonians from those with sparse Pauli decompositions to those that are dense (i.e., has sparsity Ω(2n)Ω(2^n)Ω(2n)) in the Pauli basis but are Clifford structured.
Learning an unknown quantum process is a central task for validation of the functioning of near-term devices. The task is generally hard, requiring exponentially many measurements if no prior assumptions are made on the process. However, an interesting feature of the classically-simulable Clifford group is that unknown Clif-ford operations may be efficiently determined from a black-box implementation. We extend this result to the important class of fermionic Gaussian operations. These operations have received much attention due to their close links to fermionic linear optics. We then introduce an infinite family of unitary gates, called the Matchgate Hierarchy, with a similar structure to the Clifford Hierarchy. We show that the Clifford Hierarchy is contained within the Matchgate Hierarchy and how operations at any level of the hierarchy can be efficiently learned.
We consider the problem of testing whether an unknown unitary is close to a specified level of the Clifford hierarchy. Bu, Gu, and Jaffe proposed a candidate tester for this task based on a connection with noncommutative analogues of the Gowers uniformity norms. The complexity of this tester -- whose analysis depends on a robust characterization of the near-extremizers of these norms -- was left open. We establish such a characterization for the fourth noncommutative uniformity norm and, as a consequence, obtain an efficient tester for the third level of the Clifford hierarchy. We further discuss possible routes toward resolving the problem of testing for all higher levels, highlighting the main barriers that remain.