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

ACM/IEEE Design Automation Conference (DAC)

DyREM: Dynamically Mitigating Quantum Readout Error with Embedded Accelerator

Kai Zhou, Liqiang Lu, Hanyu Zhang, Debin Xiang, Chenning Tao, Xinkui Zhao, Size Zheng, Jianwei Yin

Published Jun 22, 2025• 0 citations
Publisher

Abstract

Quantum readout error is the most significant source of error, substantially reducing the measurement fidelity. Tensor-product-based readout error mitigation has been proposed to address this issue by approximating the mitigation matrix. However, this method inevitably encounters the dynamic generation of the mitigation matrix, leading to long latency. In this paper, we propose DyREM, a software-hardware codesign approach that mitigates readout errors with an embedded accelerator. The main innovation lies in leveraging the inherent sparsity in the nonzero probability distribution of quantum states and calculating the tensor product on an embedded accelerator. Specifically, using the output sparsity, our dataflow dynamically downsamples the original mitigation matrix, which dramatically reduces the memory requirement. Then, we design DyREM architecture that can flexibly gate the redundant computation of nonzero quantum states. Experiments demonstrate that DyREM achieves an average speedup of $9.6 \times \sim 2000 \times$ and fidelity improvements of $1.03 \times \sim 1.15 \times$ compared to state-of-the-art readout error mitigation methods.

DOI
10.1109/dac63849.2025.11132635

Related papers

SiTA: Exploiting Sparsity in Tensor-Product for Accelerating Quantum Readout Error Mitigation

Hanyu Zhang, Zhiwei Ye, Liqiang Lu +6 more

Though quantum computing theoretically provides exponential computational advantage, an end-to-end quantum speedup should consider the encoding, compilation, and many other peripheral process that rely on classical computer. Among them, the readout error mitigation turns out to be the bottleneck, requiring mitigation of noise errors to achieve high fidelity. To this end, leveraging the sparsity in the tensor-product has emerged as an appealing approach to improve computational efficiency. However, existing approaches, based on intermediate data pruning or threshold pruning, suffer from the limited accelerator and fidelity loss. In this paper, we propose SiTA, a quantum error mitigation accelerator that exploits the inherent sparsity of Hilbert space. The key insight is that the superposition states generated by quantum algorithms cover only a subset of the Hilbert space (output sparsity), and the basis state naturally exhibits bit-level sparsity (input sparsity). Therefore, we introduce a sparse dataflow that features probability-level and state-level parallelism, which effectively skips calculations related to zero values. Finally, we design its hardware architecture that supports various computational paradigms of tensor-product, enabling acceleration for readout error mitigation. Experiments show that SiTA achieves an end-to-end speedup of 4.9× to 1604× over prior mitigation methods, without sacrificing fidelity.

ACM/IEEE Design Automation Conference (DAC)•Jul 26, 2026•0 citations

SpREM: Exploiting Hamming Sparsity for Fast Quantum Readout Error Mitigation

Hanyu Zhang, Liqiang Lu, Siwei Tan +3 more

The current Noisy Intermediate-Scale Quantum (NISQ) era suffers from high quantum readout error that severely reduces the measurement fidelity. Matrix-based error mitigation has been demonstrated as a promising software-level technique, which performs matrix-vector multiplication to calibrate the probability distribution with noise. However, this approach shows poor scalability and limited fidelity improvement as the matrix size exponentially increases with the number of qubits. In this paper, we propose SpREM to exploit the inherent sparsity in the mitigation matrix. Inspired by the interaction mechanism between qubits, we identify structured sparsity patterns using Hamming distance. With this insight, we propose the Hamming-Distance Sparse Row (HDSR) compression method and its format, which can achieve higher sparsity than threshold-based pruning meanwhile exhibiting great fidelity improvement. Finally, we propose the computational dataflow of the HDSR format and implement it on hardware. Experiments demonstrate that SpREM achieves 98.9% sparsity and a 27.3× reduction in fidelity loss on the real-world quantum device, compared to threshold-based pruning. It achieves an average 11.2× ~ 36.4× speedup compared to Xilinx Vitis SPARSE library and NVIDIA A100 GPU implementations.

ACM/IEEE Design Automation Conference (DAC)•Jun 23, 2024•4 citations

Perturbative readout-error mitigation for near-term quantum computers

Evan Peters, Andy C. Y. Li, Gabriel Perdue

Readout errors on near-term quantum computers can introduce significant error to the empirical probability distribution sampled from the output of a quantum circuit. These errors can be mitigated by classical postprocessing given the access of an experimental response matrix that describes the error associated with the measurement of each computational basis state. However, the resources required to characterize a complete response matrix and to compute the corrected probability distribution scale exponentially with the number of qubits, $n$. In this work, we modify standard matrix inversion techniques using perturbative approximations with significantly reduced complexity and bounded error when the likelihood of high-order bit-flip events is strongly suppressed. Given a characteristic error rate $q$, we discuss a method to recover the probability of the all-zeros bit string ${p}_{0}$ by sampling only a small subspace of the response matrix before inverting readout error, resulting in a relative speedup of $\text{poly}[{2}^{n}/(\begin{array}{c}n\\ w\end{array})]$, which we motivate using a simplified error model for which the approximation incurs only $O({q}^{w})$ error for some integer $w$. We then provide a generalized technique to efficiently recover full output distributions with $O({q}^{w})$ error in the perturbative limit. These approximate techniques for readout-error correction may greatly accelerate near-term quantum computing applications.

arXiv:2105.08161•Physical Review A•Jun 30, 2023•12 citations

MCMit: Hardware-Software Co-Design for Mid-Circuit Measurement Error Mitigation

Emmanouil Giortamis, Felix Gust, Aleksandra Świerkowska +7 more

Distributed Quantum Computing (DQC) and Quantum Error Correction (QEC) rely on dynamic circuits that include Mid-Circuit Measurements (MCMs) and classical feedback. These operations present a major bottleneck: MCMs suffer from high error rates that lead to real-time branching errors, while MCM and classical feedback latencies amplify decoherence errors. Current hardware controllers, qubit-state discriminators, and software error mitigation techniques fail to address these challenges holistically. We propose MCMit, a hardware-software co-design to mitigate branching and latency-induced errors. MCMit introduces a scalable, constant-latency multi-control branch instruction for faster classical feedback and two qubit-state discriminators, a transformer, and a CNN, with high accuracy even under short measurement durations. On the software side, static MCM elimination and stochastic branching complement the hardware by mitigating residual branching errors that persist despite hardware improvements. We implement MCMit on Qubic and evaluate it using experimentally extracted QPU readout traces. Our branch instruction reduces feedback latency by up to 70\%, improving circuit depths by up to $7\times$ over Qubic. Our CNN discriminator achieves up to 62\% higher accuracy for short measurement durations than the baselines, driving $1.2-9.4\times$ lower logical error rates in QEC. Last, our software mitigation improves fidelity by 18--30\% over baseline methods in DQC workloads.

quant-pharXiv:2604.25863•Apr 29, 2026•3 citations

Repetition-code-based readout error detection and correction across hardware platforms and generations

Csaba Czabán, Orsolya Kálmán, Sergey N. Filippov +1 more

Readout errors are one of the dominant sources of noise in current quantum processors, limiting both expectation-value estimation and sampling-based applications. Since they affect only the classical measurement outcomes, they can be addressed using classical coding techniques: immediately before measurement, each data qubit is redundantly encoded with ancilla qubits, and the resulting bit string is decoded either by post-selection or by majority voting. Unlike conventional readout error mitigation, which corrects only aggregate quantities such as expectation values, this approach operates on individual measurement shots and can therefore produce approximately corrected samples. We present a systematic cross-platform and cross-generation experimental evaluation of repetition-code readout error detection and correction. We benchmark the same protocol on IBM Heron r1-r3 superconducting processors and Quantinuum H1 and H2 trapped-ion processors while independently varying the code distance, hardware generation, and encoding layout. We find that both error detection and correction improve readout fidelity on every device and generation tested, even as the unencoded baseline improves substantially across successive hardware releases. At the same time, the value of additional redundancy depends strongly on the underlying hardware. On superconducting processors, the extra gate errors introduced by the encoding rapidly offset its benefits, whereas on trapped-ion processors the much lower gate error rates allow larger code distances to remain advantageous.

quant-pharXiv:2606.30606•Jun 30, 2026•0 citations

Efficient quantum readout-error mitigation for sparse measurement outcomes of near-term quantum devices

Bo Yang, Rudy Raymond, Shumpei Uno

The readout error on near-term quantum devices is one of the dominant noise factors, which can be mitigated by classical postprocessing called quantum readout error mitigation (QREM). The standard QREM applies the inverse of noise calibration matrix to the outcome probability distribution using exponential computational resources to the number of measured qubits. This becomes infeasible for the current quantum devices with tens of qubits or more. Here we propose two efficient QREM methods finishing in $O(n{s}^{2})$ time for probability distributions of $n$ qubits and $s$ shots, which mainly aim at mitigating sparse probability distributions such that only a few states are dominant. We compare the proposed methods with several recent QREM methods in the following three cases: expectation values of the GHZ state, its fidelities, and the estimation error of maximum likelihood amplitude estimation (MLAE) algorithm with a modified Grover iterator. The two cases of the GHZ state are on real IBM quantum devices, while the third is with numerical simulation. Using the proposed method, the mitigation of the 65-qubit GHZ state takes only a few seconds, and we witness the fidelity of the 29-qubit GHZ state exceeding 0.5. The proposed methods also succeed in reducing the estimation error in the MLAE algorithm, outperforming the results by other QREM methods in general.

arXiv:2201.11046•Physical Review A•Jul 18, 2022•45 citations
Quantum readout error is the most significant source of error, substantially reducing the measurement fidelity. Tensor-product-based readout error mitigation has been proposed to address this issue by approximating the mitigation matrix. However, this method inevitably encounters the dynamic generation of the mitigation matrix, leading to long latency. In this paper, we propose DyREM, a software-hardware codesign approach that mitigates readout errors with an embedded accelerator. The main innovation lies in leveraging the inherent sparsity in the nonzero probability distribution of quantum states and calculating the tensor product on an embedded accelerator. Specifically, using the output sparsity, our dataflow dynamically downsamples the original mitigation matrix, which dramatically reduces the memory requirement. Then, we design DyREM architecture that can flexibly gate the redundant computation of nonzero quantum states. Experiments demonstrate that DyREM achieves an average speedup of 9.6×∼2000×9.6 \times \sim 2000 \times9.6×∼2000× and fidelity improvements of 1.03×∼1.15×1.03 \times \sim 1.15 \times1.03×∼1.15× compared to state-of-the-art readout error mitigation methods.
Though quantum computing theoretically provides exponential computational advantage, an end-to-end quantum speedup should consider the encoding, compilation, and many other peripheral process that rely on classical computer. Among them, the readout error mitigation turns out to be the bottleneck, requiring mitigation of noise errors to achieve high fidelity. To this end, leveraging the sparsity in the tensor-product has emerged as an appealing approach to improve computational efficiency. However, existing approaches, based on intermediate data pruning or threshold pruning, suffer from the limited accelerator and fidelity loss. In this paper, we propose SiTA, a quantum error mitigation accelerator that exploits the inherent sparsity of Hilbert space. The key insight is that the superposition states generated by quantum algorithms cover only a subset of the Hilbert space (output sparsity), and the basis state naturally exhibits bit-level sparsity (input sparsity). Therefore, we introduce a sparse dataflow that features probability-level and state-level parallelism, which effectively skips calculations related to zero values. Finally, we design its hardware architecture that supports various computational paradigms of tensor-product, enabling acceleration for readout error mitigation. Experiments show that SiTA achieves an end-to-end speedup of 4.9× to 1604× over prior mitigation methods, without sacrificing fidelity.
The current Noisy Intermediate-Scale Quantum (NISQ) era suffers from high quantum readout error that severely reduces the measurement fidelity. Matrix-based error mitigation has been demonstrated as a promising software-level technique, which performs matrix-vector multiplication to calibrate the probability distribution with noise. However, this approach shows poor scalability and limited fidelity improvement as the matrix size exponentially increases with the number of qubits. In this paper, we propose SpREM to exploit the inherent sparsity in the mitigation matrix. Inspired by the interaction mechanism between qubits, we identify structured sparsity patterns using Hamming distance. With this insight, we propose the Hamming-Distance Sparse Row (HDSR) compression method and its format, which can achieve higher sparsity than threshold-based pruning meanwhile exhibiting great fidelity improvement. Finally, we propose the computational dataflow of the HDSR format and implement it on hardware. Experiments demonstrate that SpREM achieves 98.9% sparsity and a 27.3× reduction in fidelity loss on the real-world quantum device, compared to threshold-based pruning. It achieves an average 11.2× ~ 36.4× speedup compared to Xilinx Vitis SPARSE library and NVIDIA A100 GPU implementations.
Readout errors on near-term quantum computers can introduce significant error to the empirical probability distribution sampled from the output of a quantum circuit. These errors can be mitigated by classical postprocessing given the access of an experimental response matrix that describes the error associated with the measurement of each computational basis state. However, the resources required to characterize a complete response matrix and to compute the corrected probability distribution scale exponentially with the number of qubits, nnn. In this work, we modify standard matrix inversion techniques using perturbative approximations with significantly reduced complexity and bounded error when the likelihood of high-order bit-flip events is strongly suppressed. Given a characteristic error rate qqq, we discuss a method to recover the probability of the all-zeros bit string p0{p}_{0}p0​ by sampling only a small subspace of the response matrix before inverting readout error, resulting in a relative speedup of poly[2n/(nw)]\text{poly}[{2}^{n}/(\begin{array}{c}n\\ w\end{array})]poly[2n/(nw​)], which we motivate using a simplified error model for which the approximation incurs only O(qw)O({q}^{w})O(qw) error for some integer www. We then provide a generalized technique to efficiently recover full output distributions with O(qw)O({q}^{w})O(qw) error in the perturbative limit. These approximate techniques for readout-error correction may greatly accelerate near-term quantum computing applications.
Distributed Quantum Computing (DQC) and Quantum Error Correction (QEC) rely on dynamic circuits that include Mid-Circuit Measurements (MCMs) and classical feedback. These operations present a major bottleneck: MCMs suffer from high error rates that lead to real-time branching errors, while MCM and classical feedback latencies amplify decoherence errors. Current hardware controllers, qubit-state discriminators, and software error mitigation techniques fail to address these challenges holistically. We propose MCMit, a hardware-software co-design to mitigate branching and latency-induced errors. MCMit introduces a scalable, constant-latency multi-control branch instruction for faster classical feedback and two qubit-state discriminators, a transformer, and a CNN, with high accuracy even under short measurement durations. On the software side, static MCM elimination and stochastic branching complement the hardware by mitigating residual branching errors that persist despite hardware improvements. We implement MCMit on Qubic and evaluate it using experimentally extracted QPU readout traces. Our branch instruction reduces feedback latency by up to 70\%, improving circuit depths by up to 7×7\times7× over Qubic. Our CNN discriminator achieves up to 62\% higher accuracy for short measurement durations than the baselines, driving 1.2−9.4×1.2-9.4\times1.2−9.4× lower logical error rates in QEC. Last, our software mitigation improves fidelity by 18--30\% over baseline methods in DQC workloads.
Readout errors are one of the dominant sources of noise in current quantum processors, limiting both expectation-value estimation and sampling-based applications. Since they affect only the classical measurement outcomes, they can be addressed using classical coding techniques: immediately before measurement, each data qubit is redundantly encoded with ancilla qubits, and the resulting bit string is decoded either by post-selection or by majority voting. Unlike conventional readout error mitigation, which corrects only aggregate quantities such as expectation values, this approach operates on individual measurement shots and can therefore produce approximately corrected samples. We present a systematic cross-platform and cross-generation experimental evaluation of repetition-code readout error detection and correction. We benchmark the same protocol on IBM Heron r1-r3 superconducting processors and Quantinuum H1 and H2 trapped-ion processors while independently varying the code distance, hardware generation, and encoding layout. We find that both error detection and correction improve readout fidelity on every device and generation tested, even as the unencoded baseline improves substantially across successive hardware releases. At the same time, the value of additional redundancy depends strongly on the underlying hardware. On superconducting processors, the extra gate errors introduced by the encoding rapidly offset its benefits, whereas on trapped-ion processors the much lower gate error rates allow larger code distances to remain advantageous.
The readout error on near-term quantum devices is one of the dominant noise factors, which can be mitigated by classical postprocessing called quantum readout error mitigation (QREM). The standard QREM applies the inverse of noise calibration matrix to the outcome probability distribution using exponential computational resources to the number of measured qubits. This becomes infeasible for the current quantum devices with tens of qubits or more. Here we propose two efficient QREM methods finishing in O(ns2)O(n{s}^{2})O(ns2) time for probability distributions of nnn qubits and sss shots, which mainly aim at mitigating sparse probability distributions such that only a few states are dominant. We compare the proposed methods with several recent QREM methods in the following three cases: expectation values of the GHZ state, its fidelities, and the estimation error of maximum likelihood amplitude estimation (MLAE) algorithm with a modified Grover iterator. The two cases of the GHZ state are on real IBM quantum devices, while the third is with numerical simulation. Using the proposed method, the mitigation of the 65-qubit GHZ state takes only a few seconds, and we witness the fidelity of the 29-qubit GHZ state exceeding 0.5. The proposed methods also succeed in reducing the estimation error in the MLAE algorithm, outperforming the results by other QREM methods in general.