Arizona State University Receives NSF Grants for Quantum Code Decoding Approach

Researchers from Arizona State University, The University of Texas at Arlington, and New Mexico State University are testing a new approach to decoding quantum low-density parity-check codes, a critical step toward reliable quantum computing.

The team, including Mohsen Moradi, Taejoon Kim, Rémi Chou, and David Mitchell, is moving beyond variable-node-by-variable-node reinforcement learning to a sequential belief propagation method. Belief-propagation decoding for quantum low-density parity-check codes is attractive due to its low complexity, but its performance is often limited by short cycles, degeneracy, and convergence failures. This allows for parallel updating of all nodes within a cluster, potentially accelerating the decoding process. This work is supported by four National Science Foundation grants (CNS2451268, CNS2514415, ITE2515378, and CCF-2145917) and funding from the Office of Naval Research (Grant N000142112472). QLDPC Codes and Quantum Error Correction Foundations Researchers from Arizona State University, The University of Texas at Arlington, and New Mexico State University are focusing on optimizing decoding strategies for quantum low-density parity-check (QLDPC) codes, a promising architecture for fault-tolerant quantum computing, as recent code constructions have improved the rate-distance tradeoff. The work addresses critical limitations within QLDPC codes, specifically short cycles, degeneracy, and convergence failures, which hinder the reliable extraction of information from qubits. A new approach, sequential belief propagation, diverges from previous reinforcement-learning methods that selected only a single variable node for update per step. To manage computational complexity, the researchers introduced a permutation-invariant cluster state based on a normalized histogram of local mismatch weights, followed by quantization; this makes the number of cluster states depend on the quantization resolution rather than the cluster size. They also developed the corresponding cluster-level Markov decision process, reward function, and Q-learning update. Numerical results demonstrate the clustered learned scheduling preserves most of the error-rate benefit of variable-node level learned sequential scheduling while substantially reducing the number of scheduling decisions per belief propagation iteration. This provides an attractive latency-parallelism tradeoff and offers a pathway toward more efficient and scalable quantum error correction systems. This work is supported by National Science Foundation grants CNS2451268, CNS2514415, ITE2515378, and CCF-2145917, and the Office of Naval Research under Grant N000142112472. BP Enhancements: Post-Processing Techniques for Accuracy While standard BP struggles with “short cycles, degeneracy, and convergence failures” in QLDPC codes, researchers from Arizona State University, The University of Texas at Arlington, and New Mexico State University are moving beyond simply appending auxiliary processing stages after BP completes. Instead, Mohsen Moradi, Taejoon Kim, Rémi Chou, and David Mitchell are exploring methods to enhance the message-passing process itself, specifically through intelligent scheduling of variable node (VN) updates. This design choice addresses a limitation of earlier work; the variable-node-by-variable-node nature of previous methods offered limited within-iteration parallelism, since only one VN was updated at a time. Numerical results demonstrate this clustered learned scheduling preserves error-rate benefits while reducing the number of scheduling decisions per iteration, providing “an attractive latency-parallelism tradeoff.” This work, supported by National Science Foundation grants CNS2451268, CNS2514415, ITE2515378, and CCF-2145917, and funding from the Office of Naval Research (Grant N000142112472), suggests a path toward more efficient QLDPC decoding, crucial for realizing practical, low-overhead fault-tolerant quantum architectures. This new method partitions VNs into fixed clusters, allowing the selection and parallel updating of entire groups at each step, rather than individual nodes. This representation makes the number of cluster states depend on the quantization resolution rather than the cluster size. This work builds on existing improvements to BP, including memory-enhanced BP, guided decimation, and layered schedules, all aimed at overcoming the limitations of conventional BP for QLDPC codes. While belief propagation (BP) decoding offers low complexity, its performance is often hampered by issues like short cycles and quantum degeneracy. Recent advances have demonstrated that reinforcement-learning-based sequential variable-node (VN) scheduling (RL-S) can improve BP decoding by intelligently ordering updates, but this approach offers limited within-iteration parallelism, since only one VN is updated at a time. This aims to reduce the serial decision-making process inherent in variable-node-by-variable-node methods, potentially decreasing decoding latency without sacrificing performance. Their work centers on quantum low-density parity-check (QLDPC) codes, increasingly favored for their potential in low-overhead fault-tolerant quantum architectures, and specifically addresses limitations stemming from “short cycles, degeneracy, and convergence failures” within these codes. Researchers from Arizona State University, The University of Texas at Arlington, and New Mexico State University are testing a cluster-based extension of sequential belief propagation, building on recent work showing reinforcement-learning-based sequential variable-node (VN) scheduling (RL-S) can improve BP decoding by learning state-dependent update orders. This parallelization is intended to reduce the serial decision-making required per decoding iteration. 👉 More information🗞 Learning to Decode Quantum LDPC Codes via Cluster-Based Sequential Belief Propagation✍️ Mohsen Moradi, Taejoon Kim, Rémi A. Chou and David G. M. Mitchell🧠 ArXiv: https://arxiv.org/abs/2607.20130 Stay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: