Reproduce 2602.19722

[refraction-ray]

1. Paper Information

• Title: Differentiable Maximum Likelihood Noise Estimation for Quantum Error Correction

• Link: https://arxiv.org/abs/2602.19722, repo reference: https://github.com/CHY-i/DMLE-QEC

• Tags: quantum error correction, tensor network, automatic differentiation

2. Reproduction Goal

Which specific figure or table from the paper needs to be reproduced?

• Target: reproduce the differentiable tensor network contraction for surface code QEC. There is one key challenge you need to first survey, the support of hyper edge in TensorNetwork and cotengra, as well as the compatibility to support hyper edge in tensorcircuit. so maybe this issue can lead to two PRs, one for hyper edge support in tc with test cases and examples, and one for reproduce the differentiable surface code decoder.

• Expected Metric/Result:

• Allow to reduce the problem size compared to the paper: Yes

3. Directory Structure Plan

To keep the repository organized, please strictly follow this folder structure:

• Proposed Folder Name: examples/reproduce_papers/<YYYY>_<keywords>/
  • Naming Convention: year_keywords
  • Example: examples/reproduce_papers/2023_quantum_transformer/

4. Metadata Definition (meta.yaml)

Please provide the content for the meta.yaml file that will be placed in the folder.

Contributors/AI: Copy and fill this block directly into the meta.yaml file.

title: ""
arxiv_id: ""
url: ""
year: 
authors: 
  - ""
  - ""
tags:
  - ""
  - ""
hardware_requirements:
  gpu: False
  min_memory: ""
description: ""
outputs:
  - target: "Figure 3(a)"
    path: result.png
    script: "main.py"

5. Implementation Requirements

For AI Assistants / Contributors: Please strictly follow the rules below.

• Docstring Standard:

  • The file MUST start with a docstring containing the paper title and the link.
    Format example:

    """
    Reproduction of "Paper Title"
    Link: https://arxiv.org/abs/xxxx.xxxxx
    
    Description:
    This script reproduces Figure 3 from the paper using TensorCircuit-NG.
    """

• Outputs:
  The reproduced data and figures should be in outputs/ subfolder.

• Subfolder structure:

  examples/reproduce_papers/<YYYY>_<Keywords>/
  ├── meta.yaml
  ├── main.py
  ├── ...       # models, data, etc.
  └── outputs/
      ├── result.png    # used for gallery
      └── ...           # other outputs if necessary
  

• Code Quality & Linter:

  • Formatter: The code must be formatted using black.
  • Linter: The code must pass pylint checks (clean code, handle errors).
  • Type Hints: Standard type is NOT REQUIRED.

• Dependencies:

  • Use tensorcircuit as the main framework.
  • Use matplotlib for plotting if necessary.

6. Verification

To be considered complete, please provide:

1. A screenshot of the generated plot or the terminal output matching the paper's result.
2. Confirmation that black and pylint have been run.

# Example verification commands
black examples/reproduce_papers/<paper_subfolder>/*.py
pylint examples/reproduce_papers/<paper_subfolder>/*.py

7. Checklist

• I have checked that this paper hasn't been reproduced in the repo yet.

• The script is self-contained and runnable.

• The docstring includes the correct arXiv/DOI link.

• black formatting applied.

• pylint check passed.

[refraction-ray]

closed via #82 #83