qiskit.ignis.mitigation.CTMPExpvalMeasMitigator¶

class CTMPExpvalMeasMitigator(generators, rates, num_qubits=None, seed=None)[source]¶

N-qubit CTMP measurement error mitigator.

This class can be used with the qiskit.ignis.mitigation.expectation_value() function to apply measurement error mitigation of N-qubit measurement errors caused by one and two-body error generators. Expectation values can also be computed directly using the expectation_value() method.

For measurement mitigation to be applied the mitigator should be calibrated using the qiskit.ignis.mitigation.expval_meas_mitigator_circuits() function and qiskit.ignis.mitigation.ExpvalMeasMitigatorFitter class with the 'CTMP' mitigation method.

Initialize a TensorMeasurementMitigator

__init__(generators, rates, num_qubits=None, seed=None)[source]¶: Initialize a TensorMeasurementMitigator

Methods

`__init__`(generators, rates[, num_qubits, seed])	Initialize a TensorMeasurementMitigator
`assignment_fidelity`([qubits])	Return the measurement assignment fidelity on the specified qubits.
`assignment_matrix`([qubits])	Return the measurement assignment matrix for specified qubits.
`expectation_value`(counts[, diagonal, …])	Compute the mitigated expectation value of a diagonal observable.
`generator_matrix`([qubits])	Return the generator matrix on the specified qubits.
`mitigation_matrix`([qubits])	Return the measurement mitigation matrix for the specified qubits.
`mitigation_overhead`([qubits])	Return the mitigation overhead for expectation value estimation.
`noise_strength`([qubits])	Return the noise strength \(gamma\) on the specified qubits
`plot_assignment_matrix`([qubits, ax])	Matrix plot of the readout error assignment matrix.
`plot_mitigation_matrix`([qubits, ax])	Matrix plot of the readout error mitigation matrix.
`required_shots`(delta[, qubits])	Return the number of shots required for expectation value estimation.
`seed`([value])	Set the seed for the quantum state RNG.
`stddev_upper_bound`([shots, qubits])	Return an upper bound on standard deviation of expval estimator.

assignment_fidelity(qubits=None)¶

Return the measurement assignment fidelity on the specified qubits.

The assignment fidelity on N-qubits is defined as \(\sum_{x\in\{0, 1\}^n} P(x|x) / 2^n\), where \(P(x|x) = \rangle x|A|x\langle\), and \(A\) is the assignment_matrix().

Paramètres: qubits (Optional[List[int]]) – Optional, qubits being measured for operator expval.
Renvoie: the assignment fidelity.
Type renvoyé: float

assignment_matrix(qubits=None)[source]¶

Return the measurement assignment matrix for specified qubits.

The assignment matrix is the stochastic matrix \(A\) which assigns a noisy measurement probability distribution to an ideal input measurement distribution: \(P(i|j) = \langle i | A | j \rangle\).

Paramètres: qubits (Optional[List[int]]) – Optional, qubits being measured for operator expval.
Renvoie: the assignment matrix A.
Type renvoyé: np.ndarray

expectation_value(counts, diagonal=None, qubits=None, clbits=None)[source]¶

Compute the mitigated expectation value of a diagonal observable.

This computes the mitigated estimator of \(\langle O \rangle = \mbox{Tr}[\rho. O]\) of a diagonal observable \(O = \sum_{x\in\{0, 1\}^n} O(x)|x\rangle\!\langle x|\).

Paramètres

counts (Dict) – counts object
diagonal (Optional[ndarray]) – Optional, the vector of diagonal values for summing the expectation value. If None the the default value is \([1, -1]^\otimes n\).
qubits (Optional[List[int]]) – Optional, the measured physical qubits the count bitstrings correspond to. If None qubits are assumed to be \([0, ..., n-1]\).
clbits (Optional[List[int]]) – Optional, if not None marginalize counts to the specified bits.

Renvoie

the expectation value and standard deviation.

Type renvoyé

(float, float)

Additional Information:

The diagonal observable \(O\) is input using the diagonal kwarg as a list or Numpy array \([O(0), ..., O(2^n -1)]\). If no diagonal is specified the diagonal of the Pauli operator :math`O = mbox{diag}(Z^{otimes n}) = [1, -1]^{otimes n}` is used.

The clbits kwarg is used to marginalize the input counts dictionary over the specified bit-values, and the qubits kwarg is used to specify which physical qubits these bit-values correspond to as circuit.measure(qubits, clbits).

generator_matrix(qubits=None)[source]¶

Return the generator matrix on the specified qubits.

The generator matrix \(G\) is given by \(\sum_i r_i G_i\) where the sum is taken over all \(G_i\) acting on the specified qubits subset.

Paramètres: qubits (Optional[List[int]]) – Optional, qubit subset for the generators.
Renvoie: the generator matrix \(G\).
Type renvoyé: sps.coo_matrix

mitigation_matrix(qubits=None)[source]¶

Return the measurement mitigation matrix for the specified qubits.

The mitigation matrix \(A^{-1}\) is defined as the inverse of the assignment_matrix() \(A\).

Paramètres: qubits (Optional[List[int]]) – Optional, qubits being measured for operator expval.
Renvoie: the measurement error mitigation matrix \(A^{-1}\).
Type renvoyé: np.ndarray

mitigation_overhead(qubits=None)¶

Return the mitigation overhead for expectation value estimation.

This is the multiplicative factor of extra shots required for estimating a mitigated expectation value with the same accuracy as an unmitigated expectation value.

Paramètres: qubits (Optional[List[int]]) – Optional, qubits being measured for operator expval.
Renvoie: the mitigation overhead factor.
Type renvoyé: int

noise_strength(qubits=None)[source]¶

Return the noise strength \(gamma\) on the specified qubits

Type renvoyé: float

plot_assignment_matrix(qubits=None, ax=None)¶

Matrix plot of the readout error assignment matrix.

Paramètres

qubits (list(int)) – Optional, qubits being measured for operator expval.
ax (axes) – Optional. Axes object to add plot to.

Renvoie

the figure axes object.

Type renvoyé

plt.axes

Lève

ImportError – if matplotlib is not installed.

plot_mitigation_matrix(qubits=None, ax=None)¶

Matrix plot of the readout error mitigation matrix.

Paramètres

qubits (list(int)) – Optional, qubits being measured for operator expval.
ax (plt.axes) – Optional. Axes object to add plot to.

Renvoie

the figure axes object.

Type renvoyé

plt.axes

Lève

ImportError – if matplotlib is not installed.

required_shots(delta, qubits=None)¶

Return the number of shots required for expectation value estimation.

This is the number of shots required so that \(|\langle O \rangle_{est} - \langle O \rangle_{true}| < \delta\) with high probability (at least 2/3) and is given by \(4\delta^2 \Gamma^2\) where \(\Gamma^2\) is the mitigation_overhead().

Paramètres

delta (float) – Error tolerance for expectation value estimator.
qubits (Optional[List[int]]) – Optional, qubits being measured for operator expval.

Renvoie

the required shots.

Type renvoyé

int

seed(value=None)[source]¶: Set the seed for the quantum state RNG.

stddev_upper_bound(shots=1, qubits=None)¶

Return an upper bound on standard deviation of expval estimator.

Paramètres

shots (int) – Number of shots used for expectation value measurement.
qubits (Optional[List[int]]) – qubits being measured for operator expval.

Renvoie

the standard deviation upper bound.

Type renvoyé

float