Willow retuning was tested under injected drift

Willow retuning was tested under injected drift

Reported: Sivak and colleagues report reinforcement-learning control that reduces variation in logical error rates under deliberately injected drift.

Supported observation: Control steering improves stability 2.4-fold against a fixed policy; adding decoder steering raises this to 3.5-fold. Stability here refers to the standard deviation of logical error rates.

Conditions: The hardware experiment uses short quantum-memory runs on Willow with states prepared again for each shot. The decoder-steering method also estimates logical errors. The drift comparison uses control parameters fixed at their initial calibration.

Not established: This does not demonstrate uninterrupted execution of a long algorithm. The reported decoder steering has real-time scaling limitations; large-code scaling and continuous steering are studied in simulation.

Next evidence: Longer hardware runs with scalable feedback and explicit accounting for exploration-induced errors.

Reinforcement learning control of quantum error correction · Peer-reviewed primary research paper · 2026-07-08
https://www.nature.com/articles/s41586-026-10759-2
Source checked 2026-09-26

Revision: claim-willow-control-steering-r1 · 2026-09-26
Review: Source and editorial check. Attributed primary-source summary; no independent scientific review.
Independent replication: not assessed.

Cited revision: https://qubitwire.com/evidence/willow-control-steering?revision=claim-willow-control-steering-r1

Current corrections: https://qubitwire.com/evidence/willow-control-steering