A quantum computer is a finicky instrument. Each measurement of its state is like photographing a fast-moving object: the shorter the exposure, the higher the chance of getting a blurry shot. But if you skillfully process a series of such snapshots, you can reconstruct a clear picture.
That's exactly what the researchers did. They deliberately shortened the duration of the measurement pulses on an ion-based quantum computer, like reducing the camera's shutter speed. This increased the probability of errors, but the whole experiment became significantly faster. Then they applied the "robust shadows" protocol: alternating calibration and estimation stages, and adding random operations before each measurement to symmetrize the errors, making them predictable and correctable.
The protocol was tested on three types of quantum states: a random local state, a standard state for the QAOA optimization algorithm, and a state with Pauli correlations. In all cases, they managed to compensate for the increased errors and obtain reliable results. This paves the way for faster quantum computing without loss of accuracy.
Quantum measurements often destroy the state of the system, so physicists use "shadow tomography"—statistical reconstruction of properties from many incomplete measurements, like reconstructing a face from many blurry photographs.