Andrew Eddins | Error Mitigation Landscape | QDC 2025

Andrew Eddins | Error Mitigation Landscape | QDC 2025

🎙 Andrew Eddins 👥 203K 📅 December 1, 2025 ⏱ 51 min 👁 762 📄 expert opinion 🧭 2026-08-16
Available in: English (current) Français

Keywords

error mitigationPauli propagationshaded light conespropagated noise absorptionPEC

Summary

In this talk at QDC 2025, Andrew Eddins from IBM Quantum presents two new error mitigation tools built on classical Pauli propagation simulations: shaded light cones and propagated noise absorption (PNA). He begins by motivating the continued importance of error mitigation even in the fault-tolerant era, citing recent work on combining error mitigation with error correction. He reviews probabilistic error cancellation (PEC) and explains how Pauli propagation can efficiently estimate the impact of individual noise terms on observables. Shaded light cones use these bounds to ignore low-impact errors, significantly reducing PEC sampling overhead. PNA classically folds inverse noise into a modified observable, allowing the QPU to run a standard noisy circuit with lower variance. The talk includes details on the implementation of a Pauli propagation package (qiskit-addon-pauli-propagation) and stress tests on a 2D Ising model circuit. Eddins emphasizes the synergy between HPC-based Pauli propagation and noisy QPUs to extend the reach of error-mitigated quantum experiments.

155 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable insights into advanced error mitigation techniques, presenting novel methods that address the sampling overhead of PEC. The argumentation is solid, building on established concepts like PEC and Pauli propagation, and supported by references to recent literature. The speaker clearly explains the theoretical basis and practical implementation, making a compelling case for the utility of these tools. However, the talk is largely a presentation of the speaker’s own work, and independent validation is not provided.

Scientific Rigor, Source Quality, Title Accuracy

The speaker cites several relevant papers, including work on error mitigation in the fault-tolerant era and Pauli propagation methods. The sources are appropriate and recent, though not all are explicitly listed in the description. The title accurately reflects the content, focusing on the error mitigation landscape and introducing two new tools. The talk is technically rigorous, with detailed explanations of the methods and their implementation.

158 words

Title / Content Match

The title accurately reflects the content, focusing on the error mitigation landscape and specifically introducing two new tools.

Quality & Reliability

8/10

The talk presents novel error mitigation techniques (shaded light cones, PNA) developed by IBM Quantum, with technical details and references to recent literature. The speaker is an expert from IBM Quantum, and the content is consistent with current research trends. However, the talk is a conference presentation, not peer-reviewed, and some claims (e.g., performance improvements) are not independently verified.

Key Moments

Cited Sources

  • qiskit-addon-pauli-propagation — Mentioned as the software package for Pauli propagation simulations.
  • qiskit-addon-slc — Mentioned as the software package for shaded light cones.

Concurring Sources

Dissenting Sources

Contribution & Novelties

The talk introduces two novel error mitigation techniques that leverage classical Pauli propagation to reduce the sampling overhead of PEC. Shaded light cones provide a more refined way to identify and ignore errors that have negligible impact on the observable, while PNA offers an alternative approach by classically folding inverse noise into the observable. These methods represent a significant step towards making error mitigation more practical for near-term quantum devices.

Pour aller plus loin :

103 words

Radar Profile

The radar profile shows high scores in technical level and information quality, reflecting the advanced and detailed nature of the talk. The fiabilite_globale is also high, indicating that the content is well-supported by references and expertise. The quantite_information is slightly lower, as the talk focuses on specific techniques rather than a broad overview.

Reliability 8/10

💬 No comments were provided for analysis.