stim_noise#

stim_noise(stim_name, *parameters_and_qubits, pauli_string=None)[source]#

Applies a StimNoiseGate to the given qubits.

For a list of supported error instructions, please check Stims gate reference.

Some of the most common errors are listed below:

Name	Description
`DEPOLARIZE1`	Single qubit depolarizing noise. This channel applies one of the Pauli errors X, Y, Z with probability \(p/3\).
`DEPOLARIZE2`	Two qubit depolarizing noise. This channel applies one of the 15 non-identity two-qubit Pauli errors (IX, IY, …, ZZ) with probability \(p/15\).
`X_ERROR`	Single qubit Pauli-X error (Bit flip). Applies X with probability \(p\).
`Y_ERROR`	Single qubit Pauli-Y error. Applies Y with probability \(p\).
`Z_ERROR`	Single qubit Pauli-Z error (Phase flip). Applies Z with probability \(p\).
`PAULI_CHANNEL_1`	Custom single qubit Pauli channel. Takes 3 arguments (px, py, pz) specifying the probabilities of applying X, Y, and Z errors respectively.
`PAULI_CHANNEL_2`	Custom two qubit Pauli channel. Takes 15 arguments specifying the probabilities of applying each of the 15 non-identity two-qubit Pauli errors.
`E` (or `CORRELATED_ERROR`)	Correlated Pauli error on multiple qubits (requires `pauli_string` argument). Applies the specified Pauli string with probability \(p\).
`ELSE_CORRELATED_ERROR`	Similar to `CORRELATED_ERROR` but only applies if the previous error instruction did NOT apply an error. This allows constructing more complex conditional error models.

Warning

Every noisy operation described here behaves as a purely unitary identity gate, unless the compilation target is indeed Stim (see to_stim()). This means for instance that to_qiskit() converts the noisy operations to trivial identity gates. The same applies to the behavior of the Qrisp simulator. In other words - the noisy operations will only behave noisy if pushed through the Stim compiler.

Parameters:

stim_namestr: The name of the Stim error gate (e.g. DEPOLARIZE1, X_ERROR, CORRELATED_ERROR).
*parametersfloat: The parameters of the error channel (e.g. error probability). Further details about the semantics of the parameters can be found in the Stims gate reference.
*qubitsQubit: The qubits to apply the error channel to.
pauli_stringstr, optional: A string of Pauli operators (e.g. XX, IZ, Y) characterizing the error. This is required for correlated errors (e.g. E, CORRELATED_ERROR).

Examples

We construct a noisy Bell-pair using the extract_stim() decorator.

from qrisp import *
from qrisp.misc.stim_tools import stim_noise

@extract_stim
def generate_noisy_bell_pair():

    qv = QuantumVariable(2)

    h(qv[0])
    cx(qv[0], qv[1])

    # Add single qubit noise
    stim_noise("X_ERROR", 0.1, qv[0])
    stim_noise("X_ERROR", 0.1, qv[1])

    # Add correlated multi-qubit noise
    stim_noise("E", 0.1, qv[0], qv[1], pauli_string = "XX")

    return measure(qv)

# Generate result indices and stim circuit
res_indices, stim_circuit = generate_noisy_bell_pair()

# Compile sampler and sample
sampler = stim_circuit.compile_sampler()
all_samples = sampler.sample(1000)

# Extract results through slicing
samples = all_samples[:, res_indices]

print(samples)

# Yields:
# array([[False,  True],
#        [False,  True],
#        [ True,  True],
#        ...,
#        [ True,  True],
#        [ True,  True],
#        [False, False]], shape=(1000, 2))