Google uses reinforcement learning to correct quantum processor drift in real time
VraserX · x · 2026-07-26
Google researchers used reinforcement learning to tune thousands of quantum-control parameters while a computation was still running.
- On the Willow processor, the method improved logical stability by 3.5× under artificially introduced drift.
- After expert calibration, it reduced logical error rates by another 20%.
- The work was published in Nature, and the accompanying figure frames it as a step toward a quantum computer that can learn from its own errors.
More from Infra
- AWS on agent prompt caching: explicit + implicit patterns to slash token costs — AWS ML Blog · 2026-10-27
- Thieves spoofing GPS steal $100M+ in data center cargo since June — Ars Technica AI · 2026-10-27
- Lam Research: weekly token usage up ~250x in 18 months — Beth_Kindig · 2026-10-27
- AWS shows how to build enterprise agentic retrieval on Bedrock with CDK — AWS ML Blog · 2026-10-27
- Encrypted GPT-2 speedrun challenges FHE inference: best is 3.031 s/token, goal sub-1s — tensorqt · 2026-10-27
- Open-source Latch compiles PyTorch into training megakernels, 2.75x faster on H100 — sguada · 2026-10-27