Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adaptive Fidelity Estimation for Quantum Programs with Graph-Guided Noise Awareness

Tingting Li, Ziming Zhao, Jianwei Yin

TL;DR

This work tackles the high cost of fidelity estimation for quantum programs on NISQ devices by introducing QuFid, an adaptive, noise-aware framework that operates directly on graph representations of circuits. By modeling the program as a DAG and applying a control-flow-aware random walk over the transpiled graph, QuFid integrates transpilation-induced structural deformation into a noise-propagation operator P and derives a spectral complexity C(G) that guides batch-sized adaptive measurements via P = C(G) · log(𝒟). An uncertainty-driven stopping rule using z_α, σ, and δ ensures accurate fidelity estimates with minimal samples. Experimental evaluation on 18 benchmarks on IBM Quantum backends demonstrates substantial savings in measurement cost while maintaining fidelity bias within the target δ, outperforming fixed-shot methods and learning-based predictors like QuCT and QuEst. The approach provides a principled, interpretable framework for resource-efficient quantum program testing in heterogeneous, time-varying hardware environments.

Abstract

Fidelity estimation is a critical yet resource-intensive step in testing quantum programs on noisy intermediate-scale quantum (NISQ) devices, where the required number of measurements is difficult to predefine due to hardware noise, device heterogeneity, and transpilation-induced circuit transformations. We present QuFid, an adaptive and noise-aware framework that determines measurement budgets online by leveraging circuit structure and runtime statistical feedback. QuFid models a quantum program as a directed acyclic graph (DAG) and employs a control-flow-aware random walk to characterize noise propagation along gate dependencies. Backend-specific effects are captured via transpilation-induced structural deformation metrics, which are integrated into the random-walk formulation to induce a noise-propagation operator. Circuit complexity is then quantified through the spectral characteristics of this operator, providing a principled and lightweight basis for adaptive measurement planning. Experiments on 18 quantum benchmarks executed on IBM Quantum backends show that QuFid significantly reduces measurement cost compared to fixed-shot and learning-based baselines, while consistently maintaining acceptable fidelity bias.

Adaptive Fidelity Estimation for Quantum Programs with Graph-Guided Noise Awareness

TL;DR

This work tackles the high cost of fidelity estimation for quantum programs on NISQ devices by introducing QuFid, an adaptive, noise-aware framework that operates directly on graph representations of circuits. By modeling the program as a DAG and applying a control-flow-aware random walk over the transpiled graph, QuFid integrates transpilation-induced structural deformation into a noise-propagation operator P and derives a spectral complexity C(G) that guides batch-sized adaptive measurements via P = C(G) · log(𝒟). An uncertainty-driven stopping rule using z_α, σ, and δ ensures accurate fidelity estimates with minimal samples. Experimental evaluation on 18 benchmarks on IBM Quantum backends demonstrates substantial savings in measurement cost while maintaining fidelity bias within the target δ, outperforming fixed-shot methods and learning-based predictors like QuCT and QuEst. The approach provides a principled, interpretable framework for resource-efficient quantum program testing in heterogeneous, time-varying hardware environments.

Abstract

Fidelity estimation is a critical yet resource-intensive step in testing quantum programs on noisy intermediate-scale quantum (NISQ) devices, where the required number of measurements is difficult to predefine due to hardware noise, device heterogeneity, and transpilation-induced circuit transformations. We present QuFid, an adaptive and noise-aware framework that determines measurement budgets online by leveraging circuit structure and runtime statistical feedback. QuFid models a quantum program as a directed acyclic graph (DAG) and employs a control-flow-aware random walk to characterize noise propagation along gate dependencies. Backend-specific effects are captured via transpilation-induced structural deformation metrics, which are integrated into the random-walk formulation to induce a noise-propagation operator. Circuit complexity is then quantified through the spectral characteristics of this operator, providing a principled and lightweight basis for adaptive measurement planning. Experiments on 18 quantum benchmarks executed on IBM Quantum backends show that QuFid significantly reduces measurement cost compared to fixed-shot and learning-based baselines, while consistently maintaining acceptable fidelity bias.
Paper Structure (7 sections, 5 equations, 7 figures, 2 tables, 1 algorithm)

This paper contains 7 sections, 5 equations, 7 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Methodology
  4. Evaluation
  5. Discussion and Future Work
  6. Conclusion
  7. Acknowledgments

Figures (7)

  • Figure 1: The DAG conversion and multi-directed graph of the BV algorithm quantum circuit.
  • Figure 2: Illustrative explanation of circuit analysis.
  • Figure 3: Adaptive measurement planning procedure.
  • Figure 4: Some instances of quantum circuits.
  • Figure 5: The number of iterations required for 18 circuits with 4 qubit sizes sets the fidelity bias to less than 0.01.
  • ...and 2 more figures