A measurement concept defining how outcomes are modeled and how state descriptions are updated after an outcome is recorded. It governs outcome probabilities, information extraction, and the disturbance introduced by the measurement interaction. It does not yield reliable inference without adequate calibration, sufficient data, and appropriate estimation procedures. It supports reconstruction and validation of state and process descriptions from experimental statistics. The concept is generally stable, though practical implementations and estimation methods evolve over time.

Principle

Process tomography rests on the linearity of quantum channels acting on density matrices and on the Choi–Jamiołkowski isomorphism that identifies channels with bipartite operators. A tomographically complete set of input states and output measurements yields linear equations whose solution (with positivity and trace constraints) gives the channel estimate.

Demonstration

To characterize a single-qubit noisy gate, prepare a basis of four linearly independent input states (e.g., |0⟩,|1⟩,|+⟩,|+i⟩), apply the gate, perform state tomography on outputs, and combine the data to reconstruct the channel’s Choi matrix. Enforce complete positivity and (if appropriate) trace preservation in the fitting step.

Misapplication

Assuming a time-independent, Markovian channel when the device exhibits temporal drift or non-Markovian memory; neglecting SPAM errors and interpreting the tomographic map as a device-intrinsic channel without accounting for preparation and measurement imperfections.

Consequence

Proper process tomography yields a detailed map of how arbitrary inputs are transformed, useful for error budgeting, gate design, and verifying models used in quantum error correction. It also quantifies nonideal features (decoherence rates, systematic rotations) that guide improvements.

Reversal

If one flips the notion and treats the reconstructed process as an exact dynamical law rather than an estimator, one risks ignoring finite-sample uncertainty and context-dependence; conversely, viewing processes purely as black-box statistical summaries (without operator structure) loses predictive structure for arbitrary inputs.

Boundary

Scope: reconstruction of completely positive maps (quantum channels) or their Choi representations from controlled preparations and measurements. Excludes Hamiltonian parameter estimation methods tailored to continuous-time dynamics, protocols designed only to extract specific error rates (randomized benchmarking), and attempts to ascribe unique system-only dynamics when SPAM or environment correlations dominate.

Semantic Tension

Tension between full process tomography and scalable, partial characterization methods (randomized benchmarking, gate set tomography, direct fidelity estimation). Another tension is between ancilla-assisted tomography (which can reduce resource requirements) and prepare-and-measure approaches.

Synthesis

Process tomography is the experimental-statistical pipeline that, by preparing a tomographically complete set of inputs, measuring outputs, and solving the resulting linear inverse problem under physical constraints, reconstructs an operational representation (superoperator/Choi) of the unknown quantum process together with quantified uncertainty and model assumptions.