TR2026-133

Net-Voltage Screening of Bivariate-Bicycle Tanner-Graph Covers


    •  Nourozi, V., Mitchell, D., Koike-Akino, T., "Net-Voltage Screening of Bivariate-Bicycle Tanner-Graph Covers", IEEE International Conference on Quantum Computing and Engineering (QCE), September 2026.
      BibTeX TR2026-133 PDF
      • @inproceedings{Nourozi2026sep4,
      • author = {Nourozi, Vahid and Mitchell, David and Koike-Akino, Toshiaki},
      • title = {{Net-Voltage Screening of Bivariate-Bicycle Tanner-Graph Covers}},
      • booktitle = {IEEE International Conference on Quantum Computing and Engineering (QCE)},
      • year = 2026,
      • month = sep,
      • url = {https://www.merl.com/publications/TR2026-133}
      • }
  • MERL Contact:
  • Research Area:

    Signal Processing

Abstract:

Finite-length bivariate-bicycle (BB) quantum LDPC codes are attractive low-overhead memory candidates, but typical design methods consider only a scalar cover factor. We introduce a directionally shifted-cover design rule that treats both the split h = ax * ay and the monomial lift offsets as optimization variables. The novelty is a short-cycle net-voltage criterion. Support-difference equality defines a base Tanner 4-cycle in the base polynomial pair, which is assigned voltages in Gh = Zax * Zay; zero voltage lets the corresponding small base Tanner cycle or closed walk return to the starting lifted vertex after one traversal, whereas nonzero high-order voltage forces it to close after r traversals, where r is the net-voltage order. This criterion provides an interpretable prefilter before exact distance verification and explains why cover direction matters at fixed length. For a (6, 6) base and h = 4, the optimized pure y-cover gives [[288, 20, 18]], improving kd ^ 2 / n over the optimized pure x-cover [[288, 16, 18]].

 

  • Related News & Events

    •  NEWS    MERL Presents Five Papers at IEEE Quantum Week 2026
      Date: September 13, 2026 - September 18, 2026
      Where: Toronto, Canada
      MERL Contact: Toshiaki Koike-Akino
      Research Areas: Applied Physics, Artificial Intelligence, Machine Learning, Optimization, Signal Processing
      Brief
      • MERL is pleased to announce that five papers have been accepted to the 2026 IEEE International Conference on Quantum Computing and Engineering (QCE), also known as IEEE Quantum Week 2026, held September 13–18, 2026, in Toronto, Canada.

        The papers highlight MERL’s recent advances in quantum computing, spanning hardware-efficient quantum state preparation, quantum low-density parity-check (QLDPC) code design, graph-cover-based code construction, machine-learning-assisted code search, and reinforcement-learning-guided quantum error correction. Together, these works address important challenges toward more efficient and reliable quantum computing systems.

        The five papers are:
        - “Near-Lower-Bound Approximate Quantum State Preparation with Hardware-Efficient Circuits” — Toshiaki Koike-Akino (TR2026-131)
        - “Reinforcement-Learning-Guided Multi-Branch Decoding of Quantum LDPC Codes” — Vahid Nourozi, Toshiaki Koike-Akino, and David Mitchell (TR2026-130)
        - “Q-Learning Base Search Voltage-Labeled Covers for Weight-Six Bivariate-Bicycle Quantum LDPC Codes” — Vahid Nourozi, David Mitchell, and Toshiaki Koike-Akino (TR2026-132)
        - “Collision-Voltage Design of Directional Covers for Bivariate Bicycle Quantum LDPC Codes” — Vahid Nourozi, David Mitchell, and Toshiaki Koike-Akino (TR2026-133)
        - “Base-Preserving APM/Voltage Lifts of Bivariate Bicycle Quantum LDPC Codes” — Vahid Nourozi, David Mitchell, and Toshiaki Koike-Akino (TR2026-129)
    •