Crate ruqu_neural_decoder 

Neural Quantum Error Decoder (NQED)

This crate implements a neural-network-based quantum error decoder that combines Graph Neural Networks (GNN) with Mamba state-space models for efficient syndrome decoding.

Architecture

The NQED pipeline consists of:

1. Syndrome Graph Construction: Converts syndrome bitmaps to graph structures
2. GNN Encoding: Multi-layer graph attention for syndrome representation
3. Mamba Decoder: State-space model for sequential decoding
4. Feature Fusion: Integrates min-cut structural information

Quick Start

use ruqu_neural_decoder::{NeuralDecoder, DecoderConfig};

let config = DecoderConfig::default();
let mut decoder = NeuralDecoder::new(config);

// Create syndrome from measurements
let syndrome = vec![true, false, true, false, false];
let correction = decoder.decode(&syndrome)?;

Re-exports

pub use error::NeuralDecoderError;

pub use error::Result;

pub use graph::DetectorGraph;

pub use graph::GraphBuilder;

pub use graph::Node;

pub use graph::Edge;

pub use gnn::GNNEncoder;

pub use gnn::GNNConfig;

pub use gnn::AttentionLayer;

pub use mamba::MambaDecoder;

pub use mamba::MambaConfig;

pub use mamba::MambaState;

pub use fusion::FeatureFusion;

pub use fusion::FusionConfig;

Modules

error

Error types for the Neural Quantum Error Decoder

fusion

Feature Fusion for Neural Quantum Error Decoding

gnn

GNN Encoder for Syndrome Graphs

graph

Syndrome Graph Construction

mamba

Mamba State-Space Decoder

Structs

Correction

Correction output from the decoder

DecoderConfig

Configuration for the neural decoder

NeuralDecoder

Neural Quantum Error Decoder