A scalable and real-time neural decoder for topological quantum codes

2025-12-08

AI Summary

Quantum computation requires extremely low logical error rates, yet physical qubits suffer from high noise, necessitating quantum error correction (QEC) for fault tolerance. Existing decoders struggle to simultaneously achieve high accuracy, real-time latency, and scalability—particularly for resource-efficient codes such as color codes. This paper introduces AlphaQubit 2, a neural decoder built upon a deep neural network architecture that integrates realistic noise modeling and reinforcement learning–based training, and is optimized for inference acceleration on commercial hardware. AlphaQubit 2 is the first decoder to concurrently attain near-optimal correction performance, microsecond-scale single-cycle decoding (≤1 μs per cycle for surface codes), and strong scalability across both surface and color codes. For color codes, it achieves decoding speeds several orders of magnitude faster than state-of-the-art high-accuracy methods, while significantly outperforming mainstream real-time decoders in accuracy.

Technology Category

• Machine Learning: Quantum Machine Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements

Abstract

Fault-tolerant quantum computing will require error rates far below those achievable with physical qubits. Quantum error correction (QEC) bridges this gap, but depends on decoders being simultaneously fast, accurate, and scalable. This combination of requirements has not yet been met by a machine-learning decoder, nor by any decoder for promising resource-efficient codes such as the colour code. Here we introduce AlphaQubit 2, a neural-network decoder that achieves near-optimal logical error rates for both surface and colour codes at large scales under realistic noise. For the colour code, it is orders of magnitude faster than other high-accuracy decoders. For the surface code, we demonstrate real-time decoding faster than 1 microsecond per cycle up to distance 11 on current commercial accelerators with better accuracy than leading real-time decoders. These results support the practical application of a wider class of promising QEC codes, and establish a credible path towards high-accuracy, real-time neural decoding at the scales required for fault-tolerant quantum computation.

Problem

• Develops a fast, accurate neural decoder for quantum error correction
• Enables real-time decoding for surface and color codes at large scales
• Bridges the gap between theoretical QEC requirements and practical implementation

Innovation

• Neural network decoder for surface and color codes
• Real-time decoding faster than 1 microsecond per cycle
• Near-optimal logical error rates at large scales