Statistical Study Extends Analysis of Google 2019 Quantum Supremacy Experiment and Digital Error Model
Summarize this article with:
The challenge of verifying claims of quantum supremacy demands rigorous statistical analysis, and a new study by Gil Kalai, Tomer Shoham, and Carsten Voelkmann delves deeper into this critical area. Building upon previous investigations into Google’s 2019 quantum computation experiment, the researchers refine their statistical approach using more detailed data provided by the Google team. This work extends the analysis of predictions based on Google’s digital error model, offering a more nuanced understanding of the experiment’s results and contributing to the ongoing effort to establish verifiable benchmarks for quantum computation. By scrutinising the statistical underpinnings of these complex experiments, the team advances the field’s ability to confidently assess progress towards practical quantum technologies. This paper further studies Google’s 2019 quantum supremacy claim, with particular attention to concerns regarding the predictive power of the fidelity estimation described in previous work. The research also details a preliminary analysis of other recent Noisy Intermediate-Scale Quantum (NISQ) experiments. A central component of the quantum supremacy claims was a prediction of the fidelity of samples produced by the quantum circuit, based on fidelities, or probabilities of errors, for the individual components of the computer. The reported gap between prediction and actual results appeared statistically surprising, and this discrepancy forms a key focus of the investigation. Researchers applied statistical tools to analyze data from these experiments, emphasizing the need for rigorous methods to validate claims of quantum advantage. The study highlights a need for increased data sharing and transparency within the field to facilitate independent verification and accelerate progress. The analysis reveals that while significant progress has been made, independent validation and comprehensive data are crucial for confirming claims of computational breakthroughs.
Fidelity Formula Fails to Predict Circuit Performance This work revisits and extends analysis of the Google 2019 “quantum supremacy” experiment, focusing on the predictive power of their fidelity formula. Researchers obtained detailed data from Google and retrieved two-gate fidelities from published figures to further investigate concerns about the original claims. Analysis confirms substantial deviations between reported values and computations based on the provided data, casting doubt on the conclusiveness of the initial assertions, even for systems with 10 to 20 qubits. The core of the Google claim rested on predicting circuit fidelity based on individual component error rates. The formula calculates predicted fidelity by multiplying the probabilities of success for each individual operation and qubit readout. Using data from the 2019 experiment, researchers found average error rates for one-gate operations, two-gate operations, and qubit readout. A simplified approximation of the formula using these averages yielded a predicted fidelity, while a further simplification, combining one- and two-gate contributions, resulted in another estimate. However, concerns arose regarding the statistical surprise of the formula’s success, suggesting potential optimization influencing the results. Google initially justified the predictive power by asserting accuracy in estimating individual error probabilities and statistical independence of errors. This work challenges these assumptions, demonstrating that the accuracy of individual readout errors appears considerably higher than initially reported. Researchers retrieved individual readout and one-gate errors from a published data file, revealing discrepancies that further support their concerns about the original claims.
Error Model Inconsistencies Challenge Supremacy Claims This research presents a detailed re-examination of the methodology underpinning claims of quantum supremacy, specifically focusing on the 2019 Google experiment and extending the analysis to several additional quantum computing experiments.
The team’s investigation of Google’s error model revealed inconsistencies when applied to newly provided data, failing to reproduce the originally reported values and demonstrating reduced predictive quality. Furthermore, combining data from the Google paper with two-gate error rates did not align with either reported or empirically measured fidelities, raising concerns about the accuracy of the initial estimations. The study also identified discrepancies in the supplied readout data, which contradicted claims of stable error rates and the absence of systematic errors, and failed to adequately explain the behaviour of fidelities in patched circuits. The authors conclude that, based on the current evidence, claims of conclusive quantum supremacy, even within the 10 to 20 qubit range, require further validation and more robust supporting data. 👉 More information 🗞 Further Statistical Study of NISQ Experiments 🧠 ArXiv: https://arxiv.org/abs/2512.10722 Tags: