Diamond Norm

The diamond norm is a distance measure on quantum channels — completely positive, trace-preserving maps — that captures the distinguishability of two quantum processes when they are used as subroutines in larger circuits. Unlike the operator norm, which measures the worst-case difference on pure states, the diamond norm extends the maximization to all possible input states of a channel and its purifying environment, making it the appropriate metric for unitary approximation and quantum process tomography. It was introduced in the context of quantum information theory to resolve a subtlety: two channels that are close in operator norm may be perfectly distinguishable when composed with an entangled ancilla, and the diamond norm corrects for this by requiring complete positivity.

The diamond norm is closely related to the distinguishability of quantum states under the trace norm, and it inherits many of the same properties: it is stable under tensor products, satisfies a triangle inequality, and provides an operational interpretation in terms of the optimal success probability of a quantum hypothesis test. For unitary channels, the diamond norm reduces to a simpler expression, but for general channels it requires optimization over all possible input states. In quantum error correction, the diamond norm is used to quantify the accuracy of a recovered state after error correction, and in quantum computing it serves as the standard figure of merit for gate fidelity.