CTMP mitigator over corrects for 8Q stabilizer

[nonhermitian]

Informations

• Qiskit Ignis version: latest
• Python version:
• Operating system:

What is the current behavior?

For an 8Q X-stabilizer on a GHZ, the CTMP over corrects from the expected value of 1.0 to about 1.33 on ibmq_dublin. Standard tensored mitigation corrects to ~0.95.

Steps to reproduce the problem

import numpy as np
from qiskit import *
from qiskit.providers.aer import QasmSimulator, noise

from qiskit.ignis.mitigation import (expectation_value, 
                                     expval_meas_mitigator_circuits,
                                     ExpvalMeasMitigatorFitter)

IBMQ.load_account()
pro= IBMQ.get_provider(project='internal-test')
sim = QasmSimulator()
backend = pro.backends.ibmq_dublin

# Test Circuit Measure <XXXXXXXX> stabilizer
# Eight qubit stabilizer
qubits=[1, 4, 7, 10, 12, 2, 3, 5]

qc = QuantumCircuit(8, 8)
qc.h(0)
qc.cx(0, 1)
qc.cx(1, 2)
qc.cx(2, 3)
qc.cx(3, 4)
qc.cx(0, 5)
qc.cx(5, 6)
qc.cx(6, 7)
qc.h(range(8))
qc.measure(range(8), range(8))

shots = 8192
#Get ideal result and raw counts from backend
result_target = execute(qc, sim, shots=shots, initial_layout=qubits).result()
result_noise = execute(qc, backend, shots=shots, initial_layout=qubits).result()
counts_target = result_target.get_counts(0)
counts_noise = result_noise.get_counts(0)

circuits1, metadata1 = expval_meas_mitigator_circuits(len(qubits), method='tensored')

job_cal1 = execute(circuits1, backend, shots=8192, initial_layout=qubits)
result_cal1 = job_cal1.result()

ctmp_mitigator = ExpvalMeasMitigatorFitter(result_cal1, metadata1).fit()
print('CTMP mitigated:', expectation_value(counts_noise, meas_mitigator=ctmp_mitigator)[0])

gives

CTMP mitigated: 1.3377458081929379

What is the expected behavior?

Suggested solutions