← sonali goel

|disorder⟩

real-time inference of time-evolving disorder in quantum devices.

Four-panel animation: the observed charge-stability diagram, the true hidden disorder, the model's inferred disorder, and the charge-stability diagram re-simulated from that inference, all tracking as the disorder drifts over time
left to right: observed charge-stability diagram · true (hidden) disorder · the model's inferred disorder, best of 64 samples · the diagram re-simulated from that inference (the r value updates live)

Spin qubits and the gates that control them operate in an environment shaped by charge disorder: stray trapped charges in the material that distort device behavior and drift over time. The disorder itself cannot be measured directly. What we observe instead is the device's charge stability diagram (CSD), whose structure is determined by the underlying disorder.

In the animation, the true disorder (second panel) evolves continuously while the model only sees the streaming charge stability diagram (first panel). From these measurements, it infers the hidden disorder landscape, with the third panel showing its best prediction from 64 sampled candidates. The final panel provides the validation: a digital twin re-simulates the charge stability diagram using the inferred disorder, and the correlation with the real measurement updates live as both evolve. Real-time inference of hidden disorder is a step toward quantum devices that can continuously monitor and recalibrate themselves as operating conditions change.