Methods and systems for multi-type probabilistic quantum error mitigation
A computer implemented method, for mitigating errors in a quantum circuit comprising at least one occurrence of a quantum logic operation G. The method includes computing a set of coefficients {c p }, associated with a set of basis operations ={B p }, to obtain a quasi-probability decomposition G 0 ≈Σ p c p B p on the set of basis operations . The quasi-probability decomposition is of a target version of the quantum logic operation G, denoted G 0 . The set of basis operations {B p } forms a multi-type basis, constructed from the quantum logic operation G and elements of a set of mitigation operations . The decomposition is computed so as to reach a decomposition target, being based on at least one of a decomposition accuracy target, and a decomposition sampling overhead target. The method includes implementing the quasi-probability decomposition on the quantum processor.
1 . A computer implemented method for mitigating errors in a quantum circuit comprising at least one occurrence of a quantum logic operation G of a quantum processor, the computer implemented method comprising:
computing a set of coefficients {c p } associated with a set of basis operations {B p } to obtain a quasi-probability decomposition G 0 ≈Σ p c p B p of a target version G 0 of the quantum logic operation G on the set of basis operations {B p }, wherein:
i) the set of basis operations {B p } forms a multi-type basis, constructed from the quantum logic operation G and elements of a set of mitigation operations , wherein said set of basis operations {B p } comprises at least two basis operation types selected from a plurality of basis operation types GS i , S i G, S k G SV S m , said S i , . . . , S m being elements of the set of mitigation operations , wherein mid-circuit measurements operations are excluded from said set of mitigation operations ;
ii) said decomposition is computed so as to reach a decomposition target based on at least one of a decomposition accuracy target and a decomposition sampling overhead target;
implementing said quasi-probability decomposition on the quantum processor to estimate an outcome of a target quantum circuit in which the target version G 0 of the quantum operation G replaces the at least one occurrence of the quantum operation G.
2 . The computer implemented method according to claim 1 , wherein implementing said quasi-probability decomposition on the quantum processor comprises:
i) for each of said at least one occurrence of the quantum operation G in the quantum circuit, sampling the set of basis operations {B p } based on said set of coefficients {c p } to obtain at least one corresponding set of sampled operations; wherein the sampling is according to a probabilistic distribution defined by weights
w
p
=
❘
"\[LeftBracketingBar]"
c
p
❘
"\[RightBracketingBar]"
w
,
wherein W=Σ p |c p |
ii) running a set of sampled quantum circuits on the quantum processor wherein said set of sampled quantum circuits is determined by replacing the at least one occurrence of G in the quantum circuit by one sampled operation of said corresponding set of sampled operations so as to obtain sampled quantum circuit outcomes; wherein said replacing at least one occurrence of G in the quantum circuit is performed by replacing each occurrence of G by a sample from said corresponding set of basis operations;
iii) estimating the outcome of the target quantum circuit based on the sampled quantum circuit outcomes.
3 . The computer implemented method according to claim 1 , further comprising computing the decomposition sampling overhead, wherein computing the sampling overhead includes computing the quasi-probability norm W=Σ p |c p |.
4 . The computer implemented method according to claim 1 , wherein said target version G 0 is an ideal implementation of said quantum logic operation G.
5 . The computer implemented method according to claim 1 , wherein said target version G 0 is a noise-amplified implementation of said quantum logic operation G.
6 . The computer implemented method according to claim 1 , wherein ideally-non-unitary quantum logic operations are excluded from said set of mitigation operations .
7 . The computer implemented method according to claim 1 , wherein said S i , . . . , S m are respectively elements of the sets 1 , . . . , 5 and at least one, and preferably each, of the sets 1 , . . . , 5 includes at least about 10 n linearly independent operations, wherein n is the number of qubits on which G is applied.
8 . The method according to claim 1 , wherein the set of mitigation operations includes linearly dependent operations.
9 . The computer implemented method according to claim 1 , wherein at least some of the operations in the set of mitigation operations ideally operate non-trivially on up to a predefined number of qubits r; wherein each of the mitigation operations included in the set of mitigation operations ideally operates non-trivially on up to said predefined number of qubits r and at least one, and preferably each, of the sets 1 , . . . , 5 includes at least about
(
r
n
)
1
0
r
linearly independent operations.
10 . The computer implemented method according to claim 1 , wherein said target version G 0 of said quantum logic operation G is a layer of gates acting on non-overlapping sets of qubits.
11 . The computer implemented method according to claim 1 , wherein the quantum logic operation G is twirled.
12 . The computer implemented method according to claim 11 , wherein said target version G 0 of said quantum logic operation G comprises non-Clifford gates, and wherein said non-Clifford gates are twirled using gates that commute with G 0 .
13 . The computer implemented method according to claim 1 , wherein said set of basis operations {B p } further comprises subcircuits B p comprising two or more occurrences of the quantum logic operation G, or comprising two or more mitigation operations chosen from the set of mitigation operations ; wherein said subcircuits are of depth lower than a predefined threshold so as to be shallow.
14 . The computer implemented method according to claim 1 , wherein:
the quantum logic operation G is a sub-circuit given by a sequence of sub-operations
G
=
Ξ
K
…
Ξ
1
;
said subcircuits B p comprise of zero or more sub-operations Ξ 1 , . . . , Ξ K included in said sequence of sub-operations and zero or more operations included in the set of mitigation operations .
15 . The computer implemented method according to claim 1 , comprising optimizing a tradeoff between a decomposition accuracy and a decomposition sampling overhead, wherein said optimizing a tradeoff comprises optimizing the decomposition accuracy by solving a least-squares problem, and minimizing the sampling overhead by solving a linear program.
16 . The computer implemented method according to claim 1 , wherein said decomposition accuracy target comprises a decomposition accuracy staying below a predefined accuracy threshold, and wherein said decomposition sampling overhead target comprises a decomposition sampling overhead staying below a predefined sampling overhead threshold η.
17 . The computer implemented method according to claim 2 , wherein said estimating the outcome of said target quantum circuit comprises computing, for each sampled quantum circuit of the set of sampled quantum circuits, a corresponding sampled circuit sign equal to a sign of a product of the at least one coefficient corresponding to at least one sampled operation which replaces the at least one occurrence of G in said sampled quantum circuit; and wherein said estimating the outcome of the target quantum circuit comprises computing an average of the sampled quantum circuit outcomes, said sampled quantum circuit outcomes being weighted by said corresponding sampled circuit sign and by a quasi-probability norm W.
18 . A non-transient computer readable storage medium storing computer instructions, wherein the computer instructions are used for causing a computer to execute the computer implemented method according to claim 1 .
19 . A computer system comprising a quantum processor and a classical processor, the classical processor being configured to execute the computer executable components stored in memory, wherein the computer executable components comprise:
a decomposition component that computes a set of coefficients {c p } associated with a set of basis operations {B p } to obtain a quasi-probability decomposition G 0 ≈Σ p c p B p of a target version G 0 of the quantum logic operation G on the set of basis operations {B p }, wherein:
i) the set of basis operations {B p } forms a multi-type basis comprising at least two basis operation types selected from a plurality of basis operation types GS i , S i G, S k GS v S m , said S v , . . . , S m being elements of a set of mitigation operations , wherein mid-circuit measurements operations are excluded from said set of mitigation operations ;
ii) said decomposition is computed so as to reach a decomposition target based on at least one of a decomposition accuracy target and a decomposition sampling overhead target;
an implementation component that implements said quasi-probability decomposition on the quantum processor to estimate an outcome of a target quantum circuit in which the target version G 0 of the quantum operation G replaces the at least one occurrence of the quantum operation G.