Google Quantum AI's Willow Teaches Itself to Fix Errors Without Stopping
Google's Willow processor now uses reinforcement learning to self-calibrate mid-computation, achieving a 3.5x stability improvement and record-low logical error rates published in Nature.
- Google published a Nature paper showing a quantum computer can self-calibrate mid-computation using reinforcement learning, without stopping.
- The RL agent continuously adjusts 1,000+ analog control parameters using QEC error-detection events as its reward signal.
- Logical stability improved 3.5x under hardware drift; logical error rates dropped 20% beyond expert human calibration.
- Record logical error rates achieved: fewer than 1 error per 1,000 QEC cycles on the surface code with Willow.
- Simulations up to 449 qubits show training cost does not grow with system size, suggesting the approach scales.
- The work challenges the assumption that calibration and computation must be separate, pointing toward fully autonomous quantum processors.
Quantum computers have a calibration problem. Because they are fundamentally analog machines, their control parameters — the frequencies, amplitudes, and phases of the signals that drive the qubits — drift constantly. Quantum error correction only works when errors stay rare enough, which requires continuously adapting those parameters to shifting hardware conditions. The current solution is to halt the computation and recalibrate, but useful quantum algorithms will need to run for days or months. Stopping to retune every few hours makes that impossible.
Google Quantum AI just published a fix. In a paper in Nature titled "Reinforcement learning control of quantum error correction," the team demonstrated that a quantum computer can tune itself in real time without ever pausing the computation.
Using error data twice
Quantum error correction (QEC) works by encoding one logical qubit across many physical qubits and running parity checks to detect failures. Those checks produce a stream of binary signals called error detection events — a continuous readout of where and when errors occur. Until now, that data went to a decoder (like Google's own AlphaQubit) to determine what corrections to apply. The decoder answers "what went wrong?" but never asks "why?"
Google's approach gives the QEC process a second job. Error detection events still drive the decoder, but they also feed a reinforcement learning agent that watches the error stream, learns which control parameters to adjust to reduce it, and keeps tuning — all while the computation runs. The agent answers the "why?" in real time.