Fault-tolerant quantum computers will depend crucially on the performance of the classical decoding algorithm which takes in the results of measurements and outputs corrections to the errors inferred to have occurred. Machine learning models have shown great promise as decoders for the surface code; however, this promise has not yet been substantiated for the more challenging task of decoding quantum low-density parity-check (QLDPC) codes. In this paper, we present a recurrent, transformer-based neural network designed to decode circuit-level noise on Bivariate Bicycle (BB) codes. For the [ [ 72 , 12 , 6 ] ] BB code, at a physical error rate of p = 0.1 % , our model achieves logical error rates almost 5 times lower than belief propagation with ordered statistics decoding (BP-OSD), and roughly 5 times larger than a most-likely error decoder. Moreover, while BP-OSD has a wide distribution of runtimes with significant outliers, our model has a consistent runtime and is an order-of-magnitude faster than the worst-case times from a benchmark BP-OSD implementation. On the [ [ 144 , 12 , 12 ] ] BB code, our model obtains worse logical error rates but maintains the speed advantage. These results provide initial evidence that machine learning decoders can out-perform conventional decoders on small QLDPC codes, but suggest more complex architectures and/or training procedures are necessary to scale to larger code sizes.
Fault-tolerant quantum computers will depend crucially on the performance of the classical decoding algorithm which takes in the results of measurements and outputs corrections to the errors inferred to have occurred. Machine learning models have shown great promise as decoders for the surface code; however, this promise has not yet been substantiated for the more challenging task of decoding quantum low-density parity-check (QLDPC) codes. In this paper, we present a recurrent, transformer-based neural network designed to decode circuit-level noise on Bivariate Bicycle (BB) codes. For the [ [ 72 , 12 , 6 ] ] BB code, at a physical error rate of p = 0.1 % , our model achieves logical error rates almost 5 times lower than belief propagation with ordered statistics decoding (BP-OSD), and roughly 5 times larger than a most-likely error decoder. Moreover, while BP-OSD has a wide distribution of runtimes with significant outliers, our model has a consistent runtime and is an order-of-magnitude faster than the worst-case times from a benchmark BP-OSD implementation. On the [ [ 144 , 12 , 12 ] ] BB code, our model obtains worse logical error rates but maintains the speed advantage. These results provide initial evidence that machine learning decoders can out-perform conventional decoders on small QLDPC codes, but suggest more complex architectures and/or training procedures are necessary to scale to larger code sizes.
An efficient decoder is essential for quantum error correction, and data-driven neural decoders have emerged as promising, flexible solutions. Here, we introduce a diffusion model framework to infer logical errors from syndrome measurements in quantum low-density parity-check codes. Using the bivariate bicycle code with realistic circuit-level noise, we show that masked diffusion decoders are more accurate, often faster on average, and always faster in the worst case than other state-of-the-art decoders, including belief propagation with ordered statistics decoding (BP-OSD) and autoregressive neural decoders. We show that by using fewer diffusion steps during inference one can gain significant speed at minimal cost in accuracy. By examining the factored attention from our trained neural network we find that, despite being trained solely on paired samples of syndrome-logical errors, this diffusion decoder learns the structure of the quantum codes. We also compare both masked and continuous diffusion decoders on code-capacity noise models, finding that masked diffusion decoders scale better than continuous diffusion decoders.
In this paper, we propose a new decoder, called the Multiple-Bases Belief-Propagation List Decoder (MBBP-LD), for Quantum Low-Density Parity-Check (QLDPC) codes. It extends the Multiple-Bases Belief-Propagation (MBBP) framework, originally developed for classical cyclic LDPC codes. The proposed method preserves the linear-time complexity of standard BP decoder while improving the logical error rate. To further reduce the logical error rate, a new decision rule is introduced for the post-processing list decoder, outperforming the conventional least-metric selector (LMS) criterion. For the recently developed and implemented bivariate bicycle (BB) code with parameters [[144,12,12]], our proposed MBBP-LD decoder achieves up to 40\% lower logical error rate compared to the state-of-the-art decoder for short QLDPC codes, i.e., BP with ordered-statistics decoding (BP-OSD), while retaining the linear-time complexity of the plain BP decoder. In addition, we explore a new subclass of BB codes, that we refer to as the univariate bicycle (UB) codes, specifically with lower-weight parity checks (w=6,8). This reduces the polynomial search space for the code compared to general BB codes, i.e., by reducing the search space over two polynomial components in BB codes to just a single polynomial component in UB codes. Simulations demonstrate the promising performance of these codes under various types of BP decoders.
Quantum low density parity check (qLDPC) codes, particularly bivariate bicycle (BB) codes, achieve competitive fault tolerance thresholds while offering substantially higher encoding rates than planar surface codes. However, their intrinsically long-range stabilizer structure makes them difficult to implement on monolithic devices with nearest neighbor connectivity and limited qubit capacity. In this work, we study the realization of a BB code in a modular multiprocessor architecture, where quantum processors are interconnected through shared Bell pairs. We consider processors with all to all internal connectivity, which is feasible on trapped ion and neutral atom platforms, enabling flexible local gate execution while inter-processor (nonlocal) gates are mediated by shared entanglement. We describe a star network architecture that can realize this distributed setting. We partition the qubits of the [[144,12,12]] BB code across 4, 6, and 12 quantum processors and analyze the resulting logical error rates and pseudo-threshold performance under circuit level noise by varying the number of processors and a scaling factor that captures the additional noise associated with nonlocal operations. We use Monte Carlo simulations with BP+OSD decoding and extend the previously known BB code ansatz to the distributed setting. Our results provide architectural insight and design considerations for distributed BB codes in modular quantum computing architectures.
We present Astra, a novel and scalable decoder using graph neural networks. In general, Quantum Low Density Parity Check (QLDPC) decoding is based on Belief Propagation (BP, a variant of message-passing) and requires time intensive post-processing methods such as Ordered Statistics Decoding (OSD). Our decoder works on the Tanner graph, similarly to BP. Without using any post-processing, Astra achieves higher thresholds and better Logical Error Rates (LER) compared to BPOSD, both for surface codes trained up to distance 11 and Bivariate Bicycle (BB) codes trained up to distance 18. Moreover, we can successfully extrapolate the decoding functionality: we decode high distances (surface code up to distance 25 and BB code up to distance 34) by using decoders trained on lower distances. Extrapolated Astra achieves better LER than BPOSD for BB codes. Astra(+OSD) achieves orders of magnitude lower logical error rates for BB codes compared to BP(+OSD).
Fault-tolerant quantum computers need fast and accurate error decoding. We present CoDecoder, a small neural frontend that composes with an existing decoder to improve accuracy and reduce mean decoding time. It predicts each shot's logical outcome, retains confident predictions and routes the remaining shots to the unchanged fallback. A logical-operator-aware output head scores shifted logical operators, combining equivalent logical lines on surface codes and deciding logical qubits jointly on bivariate-bicycle (BB) codes. Fused GPU inference and efficient routing reduce the overhead paid on every shot. In registered tests under circuit-level noise, composition reduces failures by 6.4-31.4% relative to correlated matching on surface codes of distance 7-17 and by 22-50% relative to full-DEM Relay-BP on two BB codes, with 2.3-124-fold speed-ups across these comparisons. With our re-implementations of the neural decoder Cascade, composition reduces decoding time in all four settings. A joint decision of both decoders on the routed shots keeps this speed and makes the composition more accurate than Cascade in all four, with 14-18% fewer failures on the [[72,12,6]] code. On the [[72,12,6]] BB code, a conventional head with independent logical outputs fails 3.8-5.0 times as often as our joint class score under matched training conditions. Per-shot analysis shows how composition gains accuracy. On retained shots, the frontend corrects more fallback errors than it introduces, even when it is less accurate overall.
Belief Propagation combined with Ordered Statistics Decoding (BP OSD) is an open-source decoding tool used widely within the community due to its versatility and ease of application to Quantum Low Density Parity Check (QLDPC) codes. However, the degeneracy of quantum codes has been observed to degrade the performance of BP OSD, particularly leading to error floors for low physical error rates. Here, we investigate the low-weight error configurations responsible for such failures in Bivariate Bicycle codes. We show that, at code capacity, BP OSD fails on errors whose weight is significantly smaller than d/2, arising from global symmetries of the decoding graph that prevent the convergence of parallel BP. Furthermore, our analysis reveals new pitfalls that can arise when analysing the performance of BP OSD in the rare events regime, such as loss of ergodicity in splitting methods and numerical instabilities in the decoder. We show how to mitigate these failure mechanisms by modifying the BP update rules and the decoder implementation, leading to substantial improvement in logical error rate.