Quantum

CoDecoder: Make Quantum Decoders Up to 100x Faster and Even More Accurate

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 22/10/2026, 04:13:24 UTC.

Briefings

CoDecoder is a small neural frontend that improves quantum error decoding accuracy and reduces mean decoding time by composing with existing decoders. It predicts logical outcomes, routes uncertain shots to fallbacks, and uses a logical-operator-aware head to score shifted logical operators on surface and bivariate-bicycle codes. Registered tests under circuit-level noise show composition reduces failures by 6.4-31.4% relative to correlated matching on surface codes and 22-50% relative to full-DEM Relay-BP on two BB codes, with 2.3-124-fold speed-ups. Re-implementations of the neural decoder Cascade confirm composition reduces decoding time in all four settings while maintaining accuracy.

Key points

• CoDecoder predicts each shot's logical outcome and routes uncertain shots to unchanged fallbacks to reduce mean decoding time.
• A joint decision of both decoders on routed shots keeps speed and accuracy, achieving 14-18% fewer failures on the [[72,12,6]] code compared to independent heads.

Context

Fault-tolerant quantum computers require fast and accurate error decoding. CoDecoder improves accuracy and reduces decoding time by composing with existing decoders, addressing a critical bottleneck in quantum computing.

Limitations

Registered tests were conducted under circuit-level noise; results may not generalize to other noise models or hardware configurations. Some speed-up factors are relative comparisons and depend on specific code settings and decoder implementations.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 22/10/2026, 04:57:23 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeiho5s2q43fawzhhwjzjxziqp26ewpkzy653sgmdi62ayrja3dlsrq