Get started with the Estimator primitive

The steps in this topic describe how to set up primitives, explore the options you can use to configure them, and invoke them in a program.

service = QiskitRuntimeService()
fractional_gate_backend = service.least_busy(use_fractional_gates=True)

qiskit[all]~=2.3.0
qiskit-ibm-runtime~=0.43.1

Steps to use the Estimator primitive

1. Initialize the account

Because Qiskit Runtime Estimator is a managed service, you first need to initialize your account. You can then select the QPU you want to use to calculate the expectation value.

Follow the steps in the Install and set up topic if you don't already have an account.

from qiskit_ibm_runtime import QiskitRuntimeService
 
service = QiskitRuntimeService()
backend = service.least_busy(
    operational=True, simulator=False, min_num_qubits=127
)
 
print(backend.name)

Output:

ibm_torino

2. Create a circuit and an observable

You need at least one circuit and one observable as inputs to the Estimator primitive.

from qiskit.circuit.library import qaoa_ansatz
from qiskit.quantum_info import SparsePauliOp
 
entanglement = [tuple(edge) for edge in backend.coupling_map.get_edges()]
observable = SparsePauliOp.from_sparse_list(
    [("ZZ", [i, j], 0.5) for i, j in entanglement],
    num_qubits=backend.num_qubits,
)
circuit = qaoa_ansatz(observable, reps=2)
# the circuit is parametrized, so we will define the parameter values for execution
param_values = [0.1, 0.2, 0.3, 0.4]
 
print(f">>> Observable: {observable.paulis}")

Output:

>>> Observable: ['IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...',
 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII...', ...]

The circuit and observable need to be transformed to only use instructions supported by the QPU (referred to as instruction set architecture (ISA) circuits). We'll use the transpiler to do this.

from qiskit.transpiler import generate_preset_pass_manager
 
pm = generate_preset_pass_manager(optimization_level=1, backend=backend)
isa_circuit = pm.run(circuit)
isa_observable = observable.apply_layout(isa_circuit.layout)
print(f">>> Circuit ops (ISA): {isa_circuit.count_ops()}")

Output:

>>> Circuit ops (ISA): OrderedDict([('rz', 3826), ('sx', 1601), ('cz', 968)])

3. Initialize Qiskit Runtime Estimator

When you initialize the Estimator, use the mode parameter to specify the mode you want it to run in. Possible values are batch, session, or backend objects for batch, session, and job execution mode, respectively. For more information, see Introduction to Qiskit Runtime execution modes.

from qiskit_ibm_runtime import EstimatorV2 as Estimator
 
estimator = Estimator(mode=backend)

4. Invoke the Estimator and get results

Next, invoke the run() method to calculate expectation values for the input circuits and observables. The circuit, observable, and optional parameter value sets are input as primitive unified bloc (PUB) tuples.

job = estimator.run([(isa_circuit, isa_observable, param_values)])
print(f">>> Job ID: {job.job_id()}")
print(f">>> Job Status: {job.status()}")

Output:

>>> Job ID: d5k96c4jt3vs73ds5smg

>>> Job Status: QUEUED

result = job.result()
print(f">>> {result}")
print(f"  > Expectation value: {result[0].data.evs}")
print(f"  > Metadata: {result[0].metadata}")

Output:

>>> PrimitiveResult([PubResult(data=DataBin(evs=np.ndarray(<shape=(), dtype=float64>), stds=np.ndarray(<shape=(), dtype=float64>), ensemble_standard_error=np.ndarray(<shape=(), dtype=float64>)), metadata={'shots': 4096, 'target_precision': 0.015625, 'circuit_metadata': {}, 'resilience': {}, 'num_randomizations': 32})], metadata={'dynamical_decoupling': {'enable': False, 'sequence_type': 'XX', 'extra_slack_distribution': 'middle', 'scheduling_method': 'alap'}, 'twirling': {'enable_gates': False, 'enable_measure': True, 'num_randomizations': 'auto', 'shots_per_randomization': 'auto', 'interleave_randomizations': True, 'strategy': 'active-accum'}, 'resilience': {'measure_mitigation': True, 'zne_mitigation': False, 'pec_mitigation': False}, 'version': 2})
  > Expectation value: 25.8930784649363
  > Metadata: {'shots': 4096, 'target_precision': 0.015625, 'circuit_metadata': {}, 'resilience': {}, 'num_randomizations': 32}

Get started with the Estimator backend primitive

Unlike provider-specific primitives, backend primitives are generic implementations that can be used with an arbitrary backend object, as long as it implements the Backend interface.

The Estimator primitive can be run with any provider by using qiskit.primitives.BackendEstimatorV2.

Some providers implement primitives natively. See the Qiskit Ecosystem page for details.

Example:

from qiskit.primitives import BackendEstimatorV2
from <some_qiskit_provider> import QiskitProvider
 
provider = QiskitProvider()
backend = provider.get_backend('backend_name')
estimator = BackendEstimatorV2(backend)

Similarities and differences between backend and Runtime primitives

• The inputs to and outputs from qiskit.primitives.BackendEstimatorV2 follow the same PUB format as the primitives in Qiskit Runtime. See PUBs for details. However, there can be differences in the fields of the returned metadata.

• The qiskit.primitives.BackendEstimatorV2 class offers no measurement or gate error mitigation implementations out-of-the-box, as backend primitives are designed to run locally in the user's machine.

• The backend primitive interfaces expose custom EstimatorV2 Options that are different from the Runtime implementations.

Next steps