Measuring a quantum object singles out a quantum state from a set of possible states. The process is irreversible, since the object retains no information about its premeasurement uncertainty. However, a group of physicists have devised a kind of “data recovery” process based on error correction techniques used in quantum computing. As described in *Physical Review Letters*, they measured one part of an entangled quantum system and then used the other, unmeasured part to reset everything to the preobserved state.

In a quantum computer, the unit of information is a qubit that exists in two states, “zero” and “one,” at the same time. This superposition is not directly observable, since measuring a qubit can only return either “zero” or “one.” The initial state is irretrievable once a measurement is made, making it impossible to backup (or “clone”) a qubit to compensate for errors in quantum computing. However, by entangling multiple qubits, quantum error correction creates a cross-check for spotting data corruption.

Errors and measurements induce similar changes to a quantum system. Therefore, Philipp Schindler of the University of Innsbruck in Austria and his colleagues adapted an error correction protocol to recover quantum information following a measurement. They started by encoding an arbitrary initial state on a system of three trapped calcium ions. They then temporarily excited two of the ions to energetically isolate them from a light beam that measured whether the third ion was in the “zero” or “one” state. To undo the effects of this measurement, the team re-cooled the ion and then re-imprinted the initial state using the two unmeasured ions. The final three-ion configuration matched the original at a level of around $84\%$. – *Michael Schirber*