Note

This is the documentation for the current state of the development branch of Qiskit Experiments. The documentation or APIs here can change prior to being released.

ProcessTomographyAnalysis¶

class ProcessTomographyAnalysis[source]¶

Quantum state and process tomography experiment analysis.

Warning

The API for tomography fitters is still under development so may change in future releases.

Overview

Fitter Functions

Built-in fitter functions may be selected using the following string labels, refer to the corresponding functions documentation for additional details on the fitters.

"linear_inversion": linear_inversion() (Default)
"cvxpy_linear_lstsq": cvxpy_linear_lstsq()
"cvxpy_gaussian_lstsq": cvxpy_gaussian_lstsq()

PSD Rescaling

For fitters that do not constrain the reconstructed state to be positive-semidefinite (PSD) we construct the maximum-likelihood nearest PSD state under the assumption of Gaussian measurement noise using the rescaling method in Reference [1]. For fitters that already support PSD constraints this option can be disabled by setting rescale_positive=False.

References

[1] John A. Smolin, Jay M. Gambetta, Graeme Smith, Maximum Likelihood, Minimum Effort, Phys. Rev. Lett. 108, 070502 (2012), doi: 10.1103/PhysRevLett.108.070502 (open)

Analysis options

These are the keyword arguments of the run() method.

Options

Defined in the class ProcessTomographyAnalysis:
- measurement_basis (MeasurementBasis)
  
  Default value: Instance of PauliMeasurementBasis
  
  The measurement MeasurementBasis to use for tomographic process reconstruction.
- preparation_basis (PreparationBasis)
  
  Default value: Instance of PauliPreparationBasis
  
  The preparation PreparationBasis to use for tomographic process reconstruction.
- fitter (str or Callable)
  
  Default value: "linear_inversion"
  
  The fitter function to use for reconstruction. This can be a string to select one of the built-in fitters, or a callable to supply a custom fitter function. See the Fitter Functions section for additional information.
- fitter_options (dict)
  
  Default value: {}
  
  Any addition kwarg options to be supplied to the fitter function. For documentation of available kwargs refer to the fitter function documentation.
- rescale_positive (bool)
  
  Default value: True
  
  If True rescale the state returned by the fitter to be positive-semidefinite. See the PSD Rescaling section for additional information (Default: True).
- rescale_trace (bool)
  
  Default value: True
  
  If True rescale the state returned by the fitter have either trace 1 for DensityMatrix, or trace dim for Choi matrices (Default: True).
- measurement_qubits (Sequence[int])
  
  Default value: None
  
  Optional, the physical qubits with tomographic measurements. If not specified will be set to [0, ..., N-1] for N-qubit tomographic measurements.
- preparation_qubits (Sequence[int])
  
  Default value: None
  
  Optional, the physical qubits with tomographic preparations. If not specified will be set to [0, ..., N-1] for N-qubit tomographic preparations.
- target (str or QuantumChannel or Operator)
  
  Default value: None
  
  Optional, Set a custom target quantum channel for computing the process_fidelity() of the fitted process against (Default: None).
- conditional_circuit_clbits (list[int])
  
  Default value: None
  
  Optional, the clbit indices in the source circuit to be conditioned on when reconstructing the channel. Enabling this will return a list of reconstructed channel components conditional on the values of these clbit values. The integer value of the conditioning clbits is stored in state analysis result extra field “conditional_circuit_outcome”.
- conditional_measurement_indices (list[int])
  
  Default value: None
  
  Optional, indices of tomography measurement qubits to used for conditional state reconstruction. Enabling this will return a list of reconstructed channel components conditioned on the remaining tomographic bases conditional on the basis index, and outcome value for these measurements. The conditional measurement basis index and integer value of the measurement outcome is stored in state analysis result extra fields “conditional_measurement_index” and “conditional_measurement_outcome” respectively.
- conditional_preparation_indices (list[int])
  
  Default value: None
  
  Optional, indices of tomography preparation qubits to used for conditional state reconstruction. Enabling this will return a list of reconstructed channel components conditioned on the remaining tomographic bases conditional on the basis index. The conditional preparation basis index is stored in state analysis result extra fields “conditional_preparation_index”.
Defined in the class TomographyAnalysis:
- target_bootstrap_samples (int)
  
  Default value: 0
  
  Optional, number of outcome re-samples to draw from measurement data for each basis for computing a bootstrapped standard error of fidelity with the target state. If 0 no bootstrapping will be performed and the target fidelity will not include a standard error (Default: 0).
- target_bootstrap_seed (int | None | Generator)
  
  Default value: None
  
  Optional, RNG seed or Generator to use for bootstrapping data for boostrapped fidelity standard error calculation (Default: None).
- extra (Dict[str, Any])
  
  Default value: {}
  
  Extra metadata dictionary attached to analysis results.
Defined in the class BaseAnalysis:
- figure_names (str or List[str])
  
  Default value: None
  
  Identifier of figures that appear in the experiment data to sort figures by name.

Note

Fitters starting with "cvxpy_*" require the optional CVXPY Python package to be installed.