Coogee '26 Talks - Basudha Srivastava (PsiQuantum)

Coogee '26 Talks - Basudha Srivastava (PsiQuantum)

🎙 Basudha Srivastava 👥 137 📅 February 19, 2026 ⏱ 30 min 👁 210 📄 original study 🧭 2026-08-16
Available in: English (current) Français

Keywords

concatenated codesbelief propagationquantum error correctiondecodingXYZ square code

Summary

Basudha Srivastava presents her work on sequential decoding of concatenated quantum codes, focusing on the XYZ square code. She begins by explaining concatenated codes and their historical significance, then introduces belief propagation as a decoding method. The XYZ square code is described as a non-CSS topological code that can be viewed as a concatenation of a higher-level code and a lower-level 2-qubit code. The decoding strategy involves first decoding the lower-level links, then transforming the error probabilities, and finally decoding the higher-level code using matching or belief-matching decoders. Results show that this approach achieves good thresholds, especially for biased noise, and is computationally efficient. The talk also discusses challenges with phenomenological noise and the occasional superiority of matching over belief matching. The presentation concludes with a brief note on hiring at PsiQuantum.

132 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable insights into a novel decoding approach for concatenated codes, demonstrating that sequential decoding with belief propagation can achieve near-optimal performance with lower complexity. The argumentation is solid, supported by simulation results and comparisons with optimal decoding. The speaker clearly explains the intuition behind the method and addresses potential limitations, such as the issue with wrong priors in belief matching. The work contributes to the growing interest in concatenated codes for fault-tolerant quantum computing.

Scientific Rigor, Source Quality, Title Accuracy

The talk is scientifically rigorous, with clear explanations of the code structure and decoding algorithms. The speaker references relevant prior work, including the matching codes by James Wen and the paper on optimal decoding of concatenated codes. The title accurately reflects the content. The presentation is well-structured, and the speaker is transparent about the limitations and open questions. The sources cited are appropriate and support the claims made.

160 words

Title / Content Match

The title accurately reflects the content: a technical talk on sequential decoding of concatenated codes.

Quality & Reliability

8/10

The talk presents original research on decoding concatenated quantum codes, with technical depth and references to prior work. The methodology is clearly explained, and the results are presented with appropriate caveats. However, the talk is a conference presentation, not a peer-reviewed paper, and some details are omitted for brevity.

Key Moments

Cited Sources

  • Matching codes — Referenced as the basis for the XYZ square code
  • Optimal decoding of concatenated codes — Referenced for the theoretical foundation of sequential decoding

Concurring Sources

  • Quantum error correction — General background on quantum error correction, supporting the importance of decoding.
  • Belief propagation — Explains the algorithm used for decoding, consistent with the talk's description.

Contribution & Novelties

The talk presents a novel decoding scheme for concatenated codes, specifically the XYZ square code, that leverages belief propagation and sequential decoding. This approach achieves near-optimal thresholds with lower computational cost, particularly for biased noise. The work contributes to the revival of concatenated codes as a viable option for fault-tolerant quantum computing.

Pour aller plus loin :

94 words

Radar Profile

The radar profile shows high scores in technical level and information quality, indicating a specialized and well-presented talk. The lower score in information quantity reflects the focused scope of the presentation, while the overall high scores suggest a valuable contribution to the field.

Reliability 8/10

💬 No comments were provided for analysis.