Optimized Clifford Noise Reduction: Theory, Simulations and Experiments
Edwin Tham, Nicolas Delfosse · 2025
Read Optimized Clifford Noise Reduction: Theory, Simulations and Experiments by Edwin Tham, Nicolas Delfosse (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-13356.txt.
Introduction
Edwin Tham and Nicolas Delfosse improve in 2025 a partial error correction scheme for quantum computers that run Clifford circuits. The scheme, called CliNR, runs the circuit by consuming a prepared resource state, and errors are caught by measuring a chosen sequence of checks.
The catch is picking a check sequence that keeps the final error rate low. The authors first search for one directly with a tabu search method. Then they build a cheaper stand-in score for the error rate, find the sequence that minimizes the stand-in, and polish it against the true rate. Simulations and experiments test the gains.
Catalog record
| CREATOR | Edwin Tham, Nicolas Delfosse |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-13356
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Optimized Clifford Noise Reduction: Theory, Simulations and Experiments Edwin Tham, Nicolas Delfosse Abstract We propose several optimizations of the CliNR partial error correction scheme which implements Clifford circuits by consuming a resource state. Errors are corrected by measuring a sequence of Pauli operators that we refer to as the verification sequence. We first propose a global optimization algorithm searching for a verification sequence resulting in a low logical error rate using tabu search. Then, we introduce a proxy for the logical error rate which is easier to evaluate and we design a two-step optimization algorithm. First, a verification sequence minimizing the proxy is computed, then this sequence is refined by reintroducing the logical error rate. Finally, we identify a large group of automorphisms of the search space which preserve the proxy and we use this symmetry to reduce the size of the search space. This results in a 168 × 168\times (respectively 20,160 × 20,160\times ) reduction of the size of the search space for the optimization of verification sequences with three (respectively four) Pauli operators. Our numerical simulations for 20-qubit Clifford circuits with size 400 400 under the ion chain model show that our optimization algorithms improve the performance of CliNR by 25 % 25\% and that the two-step optimization achieves the same results as the global optimization with 64 % 64\% fewer evaluations of the logical error rate. Finally, we perform experiments on a 36-qubit trapped ion quantum computer, without mid-circuit measurements, showing that the CZNR variant of CliNR is at breakeven. 1 Introduction Large-scale quantum computing applications likely require a fault-tolerant quantum computer in order to correct faults occurring during the computation before they spread to all the qubits [ Shor(1996) , Preskill(2025) ] . However, fault tolerance comes with a massive overhead in term of qubit and gate count. Partial error correction is likely to play an important role in the near-term as it consumes fewer qubits and fewer gates. For example, the recent CliNR scheme has a qubit overhead of only about 3\times whereas standard quantum error correction codes generally consume tens or hundreds of physical qubits per logical qubit [ Delfosse and Tham(2024) ] . Even though it is not designed to be fault-tolerant, and it does not reduce the logical error rate as much as fault-tolerant quantum error correction, previous simulations show that this scheme could be useful to reduce the logical error rate of quantum circuits in near-term noise regimes. With CliNR, the Clifford circuit to implement is decomposed into a sequence of t subcircuits and each subcircuit C is implemented in three steps as follows. 1. Resource state preparation. A resource state |\psi\rangle is prepared on ancilla qubits. This state a stabilizer state [ Gottesman(1997) ] . 2. Resource state verification. The resource state |\psi\rangle is tested by measuring some of its stabilizers. By definition, these measurements should return a trivial outcome. If not, we know that a fault has occurred, whereupon we reset the qubits and restart the preparation of |\psi\rangle . 3. Resource state injection. If all measurement outcomes are trivial, the subcircuit C is applied to an n -qubit input state by consuming the resource state |\psi\rangle . This approach to implement quantum gates through resources state has been extensively used before CliNR. Shor used resource states to perform error correction and logical Toffoli gates in his seminal paper introducing fault-tolerant quantum computation [ Shor(1996) ] . Steane and Knill proposed quantum error correction protocols based on different resource states [ Steane(1997) , Knill(2005) ] . Magic states are resource states that can be consumed to implement non-Clifford gates [ Bravyi and Kitaev(2005) ] . Gate teleportation provides a general framework to this approach and it is a core ingredient in CliNR [ Gottesman and Chuang(1999) ] . Measurement-based quantum computation [ Raussendorf and Briegel(2001) , Briegel et al.(2009)Briegel, Browne, Dür, Raussendorf, and Van den Nest ] and fusion-based quantum computation [ Bartolucci et al.(2023)Bartolucci, Birchall, Bombin, Cable, Dawson, Gimeno-Segovia, Johnston, Kieling, Nickerson, Pant, et al. ] also rely on resource states to perform quantum error correction and fault-tolerant logical operations. These schemes generally require a long sequence of stabilizer measurements for resource state verification to achieve fault tolerance or to reach the full code distance. Here, we focus on CliNR and our goal to improve the logical error rate of CliNR implementation of a circuit using a relatively short sequence of stabilizer measurements, which keeps the overhead low. We refer to the stabilizers selected for the resource state verification as the verification sequence . Our main goal is to address the following problem. The CliNR optimization problem. Given a noise model and a Clifford circuit C , find a verification sequence that minimizes the logical error rate of the CliNR implementation of C . The main challenge with the CliNR optimization problem is the combinatorial explosion of the number of possible verification sequences. The number of choices for a n -qubit circuit is \displaystyle\prod_{i=0}^{r-1}\left(2^{2n}-2^{i}\right) (1) where r is the verification sequence size. Therein, we assume that the r stabilizers of the verification sequence are independent. For n=10 and r=2 this number is already larger than 10^{12} and it reaches more than 10^{36} when n=20 and r=3 which is a relevant regime for CliNR based on previous simulations. If the circuit is sufficiently small, we can enumerate all possible verification sequences. Each verification sequence can then be evaluated through Monte-Carlo simulation estimates of the logical error rate of C under the corresponding CliNR implementation. This provides a solution to the CliNR optimization problem; though, this exhaustive approach is clearly limited to small circuits. In [ Delfosse and Tham(2024) ] , the verification sequence was designed either by selecting stabilizers uniformly at random or by selecting low-weight stabilizers because they can be implemented with fewer gates. It is natural to expect that a more careful selection of the stabilizers, tailored to the exact noise model and structure of the circuit, would lead to a reduction of the logical error rate achieved with CliNR. To go beyond these naive optimizations, we introduce a heuristic optimization algorithm ( Algorithm 1 ) based on tabu search [ Glover(1986) , Glover(1989) ] . We chose tabu search for its simplicity, but this subroutine can be replaced by other metaheuristic optimization algorithm such as hill climbing, simulated annealing or more involved strategies. We start from a randomly selected verification sequence. Then, at each iteration we evaluate a set of neighboring verification sequences by replacing one of the stabilizers within that sequence by a random stabilizer of the resource state. The neighboring verification sequence that reduces the logical error rate the most is accepted and added to a tabu list – this prevents the same verification sequence from being revisited over a set number of iterations, and helps the algorithm avoid local minima. The most resource intensive subroutine in Algorithm 1 is the cost function evaluation, that is the estimation of the logical error rate of CliNR for a given verification sequence. To speed up this subroutine, we introduce a proxy cost function, that is an approximation of the logical error rate and we propose a two-step optimization algorithm ( Algorithm 3 ). First, we