# Copyright Amazon.com Inc. or its affiliates. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"). You
# may not use this file except in compliance with the License. A copy of
# the License is located at
#
# http://aws.amazon.com/apache2.0/
#
# or in the "license" file accompanying this file. This file is
# distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF
# ANY KIND, either express or implied. See the License for the specific
# language governing permissions and limitations under the License.
import itertools
from collections.abc import Iterable
from typing import Any, ClassVar, Union
import numpy as np
import braket.ir.jaqcd as ir
from braket.circuits import circuit
from braket.circuits.free_parameter import FreeParameter
from braket.circuits.free_parameter_expression import FreeParameterExpression
from braket.circuits.gates import format_complex
from braket.circuits.instruction import Instruction
from braket.circuits.noise import (
DampingNoise,
GeneralizedAmplitudeDampingNoise,
MultiQubitPauliNoise,
Noise,
PauliNoise,
SingleProbabilisticNoise,
SingleProbabilisticNoise_34,
SingleProbabilisticNoise_1516,
)
from braket.circuits.quantum_operator_helpers import (
is_cptp,
verify_quantum_operator_matrix_dimensions,
)
from braket.circuits.serialization import OpenQASMSerializationProperties
from braket.registers.qubit import QubitInput
from braket.registers.qubit_set import QubitSet, QubitSetInput
"""
To add a new Noise implementation:
1. Implement the class and extend `Noise`
2. Add a method with the `@circuit.subroutine(register=True)` decorator. Method name
will be added into the `Circuit` class. This method is the default way clients add
this noise to a circuit.
3. Register the class with the `Noise` class via `Noise.register_noise()`.
"""
[docs]
class BitFlip(SingleProbabilisticNoise):
r"""Bit flip noise channel which transforms a density matrix :math:`\\rho` according to:
.. math:: \\rho \\Rightarrow (1-p) \\rho + p X \\rho X^{\\dagger}
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
X = \\left(
\\begin{matrix}
0 & 1 \\\\
1 & 0
\\end{matrix}
\\right)
p = \\text{probability}
This noise channel is shown as `BF` in circuit diagrams.
"""
def __init__(self, probability: Union[FreeParameterExpression, float]):
super().__init__(
probability=probability,
qubit_count=None,
ascii_symbols=[_ascii_representation("BF", [probability])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.BitFlip.construct(target=target[0], probability=self.probability)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return f"#pragma braket noise bit_flip({self.probability}) {target_qubit}"
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.sqrt(1 - self.probability) * np.eye(2, dtype=complex)
K1 = np.sqrt(self.probability) * np.array([[0.0, 1.0], [1.0, 0.0]], dtype=complex)
return [K0, K1]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def bit_flip(target: QubitSetInput, probability: float) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s)
probability (float): Probability of bit flipping.
Returns:
Iterable[Instruction]: `Iterable` of BitFlip instructions.
Examples:
>>> circ = Circuit().bit_flip(0, probability=0.1)
"""
return [
Instruction(Noise.BitFlip(probability=probability), target=qubit)
for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs: Union[FreeParameter, str]) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Args:
**kwargs (Union[FreeParameter, str]): Arbitrary keyword arguments.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return BitFlip(probability=_substitute_value(self._probability, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return BitFlip(probability=_parameter_from_dict(noise["probability"]))
Noise.register_noise(BitFlip)
[docs]
class PhaseFlip(SingleProbabilisticNoise):
r"""Phase flip noise channel which transforms a density matrix :math:`\\rho` according to:
.. math:: \\rho \\Rightarrow (1-p) \\rho + p X \\rho X^{\\dagger}
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
Z = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & -1
\\end{matrix}
\\right)
p = \\text{probability}
This noise channel is shown as `PF` in circuit diagrams.
"""
def __init__(self, probability: Union[FreeParameterExpression, float]):
super().__init__(
probability=probability,
qubit_count=None,
ascii_symbols=[_ascii_representation("PF", [probability])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.PhaseFlip.construct(target=target[0], probability=self.probability)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return f"#pragma braket noise phase_flip({self.probability}) {target_qubit}"
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.sqrt(1 - self.probability) * np.eye(2, dtype=complex)
K1 = np.sqrt(self.probability) * np.array([[1.0, 0.0], [0.0, -1.0]], dtype=complex)
return [K0, K1]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def phase_flip(target: QubitSetInput, probability: float) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s)
probability (float): Probability of phase flipping.
Returns:
Iterable[Instruction]: `Iterable` of PhaseFlip instructions.
Examples:
>>> circ = Circuit().phase_flip(0, probability=0.1)
"""
return [
Instruction(Noise.PhaseFlip(probability=probability), target=qubit)
for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs: Union[FreeParameter, str]) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Args:
**kwargs (Union[FreeParameter, str]): Arbitrary keyword arguments.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return PhaseFlip(probability=_substitute_value(self._probability, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return PhaseFlip(probability=_parameter_from_dict(noise["probability"]))
Noise.register_noise(PhaseFlip)
[docs]
class PauliChannel(PauliNoise):
r"""Pauli noise channel which transforms a density matrix :math:`\\rho` according to:
.. math::
\\rho \\Rightarrow (1-probX-probY-probZ) \\rho
+ probX X \\rho X^{\\dagger}
+ probY Y \\rho Y^{\\dagger}
+ probZ Z \\rho Z^{\\dagger}
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
X = \\left(
\\begin{matrix}
0 & 1 \\\\
1 & 0
\\end{matrix}
\\right)
Y = \\left(
\\begin{matrix}
0 & -i \\\\
i & 0
\\end{matrix}
\\right)
Z = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & -1
\\end{matrix}
\\right)
This noise channel is shown as `PC` in circuit diagrams.
"""
def __init__(
self,
probX: Union[FreeParameterExpression, float],
probY: Union[FreeParameterExpression, float],
probZ: Union[FreeParameterExpression, float],
):
"""Creates PauliChannel noise.
Args:
probX (Union[FreeParameterExpression, float]): X rotation probability.
probY (Union[FreeParameterExpression, float]): Y rotation probability.
probZ (Union[FreeParameterExpression, float]): Z rotation probability.
"""
super().__init__(
probX=probX,
probY=probY,
probZ=probZ,
qubit_count=None,
ascii_symbols=[_ascii_representation("PC", [probX, probY, probZ])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.PauliChannel.construct(
target=target[0], probX=self.probX, probY=self.probY, probZ=self.probZ
)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return (
f"#pragma braket noise pauli_channel"
f"({self.probX}, {self.probY}, {self.probZ}) {target_qubit}"
)
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.sqrt(1 - self.probX - self.probY - self.probZ) * np.eye(2, dtype=complex)
K1 = np.sqrt(self.probX) * np.array([[0.0, 1.0], [1.0, 0.0]], dtype=complex)
K2 = np.sqrt(self.probY) * 1j * np.array([[0.0, -1.0], [1.0, 0.0]], dtype=complex)
K3 = np.sqrt(self.probZ) * np.array([[1.0, 0.0], [0.0, -1.0]], dtype=complex)
return [K0, K1, K2, K3]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def pauli_channel(
target: QubitSetInput, probX: float, probY: float, probZ: float
) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s)
probability list[float]: Probabilities for the Pauli X, Y and Z noise
happening in the Kraus channel.
probX (float): X rotation probability.
probY (float): Y rotation probability.
probZ (float): Z rotation probability.
Returns:
Iterable[Instruction]: `Iterable` of PauliChannel instructions.
Examples:
>>> circ = Circuit().pauli_channel(0, probX=0.1, probY=0.2, probZ=0.3)
"""
return [
Instruction(Noise.PauliChannel(probX=probX, probY=probY, probZ=probZ), target=qubit)
for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
probX = _substitute_value(self.probX, **kwargs)
probY = _substitute_value(self.probY, **kwargs)
probZ = _substitute_value(self.probZ, **kwargs)
return PauliChannel(probX=probX, probY=probY, probZ=probZ)
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return PauliChannel(
probX=_parameter_from_dict(noise["probX"]),
probY=_parameter_from_dict(noise["probY"]),
probZ=_parameter_from_dict(noise["probZ"]),
)
Noise.register_noise(PauliChannel)
[docs]
class Depolarizing(SingleProbabilisticNoise_34):
r"""Depolarizing noise channel which transforms a density matrix :math:`\\rho` according to:
.. math::
\\rho \\Rightarrow (1-p) \\rho
+ p/3 X \\rho X^{\\dagger}
+ p/3 Y \\rho Y^{\\dagger}
+ p/3 Z \\rho Z^{\\dagger}
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
X = \\left(
\\begin{matrix}
0 & 1 \\\\
1 & 0
\\end{matrix}
\\right)
Y = \\left(
\\begin{matrix}
0 & -i \\\\
i & 0
\\end{matrix}
\\right)
Z = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & -1
\\end{matrix}
\\right)
p = \\text{probability}
This noise channel is shown as `DEPO` in circuit diagrams.
"""
def __init__(self, probability: Union[FreeParameterExpression, float]):
super().__init__(
probability=probability,
qubit_count=None,
ascii_symbols=[_ascii_representation("DEPO", [probability])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.Depolarizing.construct(target=target[0], probability=self.probability)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return f"#pragma braket noise depolarizing({self.probability}) {target_qubit}"
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.sqrt(1 - self.probability) * np.eye(2, dtype=complex)
K1 = np.sqrt(self.probability / 3) * np.array([[0.0, 1.0], [1.0, 0.0]], dtype=complex)
K2 = np.sqrt(self.probability / 3) * 1j * np.array([[0.0, -1.0], [1.0, 0.0]], dtype=complex)
K3 = np.sqrt(self.probability / 3) * np.array([[1.0, 0.0], [0.0, -1.0]], dtype=complex)
return [K0, K1, K2, K3]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def depolarizing(target: QubitSetInput, probability: float) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s)
probability (float): Probability of depolarizing.
Returns:
Iterable[Instruction]: `Iterable` of Depolarizing instructions.
Examples:
>>> circ = Circuit().depolarizing(0, probability=0.1)
"""
return [
Instruction(Noise.Depolarizing(probability=probability), target=qubit)
for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return Depolarizing(probability=_substitute_value(self._probability, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return Depolarizing(probability=_parameter_from_dict(noise["probability"]))
Noise.register_noise(Depolarizing)
[docs]
class TwoQubitDepolarizing(SingleProbabilisticNoise_1516):
r"""Two-Qubit Depolarizing noise channel which transforms a
density matrix :math:`\\rho` according to:
.. math::
\\rho \\Rightarrow (1-p) \\rho + p/15 (
IX \\rho IX^{\\dagger} + IY \\rho IY^{\\dagger} + IZ \\rho IZ^{\\dagger}
+ XI \\rho XI^{\\dagger} + XX \\rho XX^{\\dagger} + XY \\rho XY^{\\dagger}
+ XZ \\rho XZ^{\\dagger} + YI \\rho YI^{\\dagger} + YX \\rho YX^{\\dagger}
+ YY \\rho YY^{\\dagger} + YZ \\rho YZ^{\\dagger} + ZI \\rho ZI^{\\dagger}
+ ZX \\rho ZX^{\\dagger} + ZY \\rho ZY^{\\dagger} + ZZ \\rho ZZ^{\\dagger})
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
X = \\left(
\\begin{matrix}
0 & 1 \\\\
1 & 0
\\end{matrix}
\\right)
Y = \\left(
\\begin{matrix}
0 & -i \\\\
i & 0
\\end{matrix}
\\right)
Z = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & -1
\\end{matrix}
\\right)
p = \\text{probability}
This noise channel is shown as `DEPO` in circuit diagrams.
"""
def __init__(self, probability: Union[FreeParameterExpression, float]):
super().__init__(
probability=probability,
qubit_count=None,
ascii_symbols=[_ascii_representation("DEPO", [probability])] * 2,
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.TwoQubitDepolarizing.construct(
targets=[target[0], target[1]], probability=self.probability
)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit_0 = serialization_properties.format_target(int(target[0]))
target_qubit_1 = serialization_properties.format_target(int(target[1]))
return (
f"#pragma braket noise two_qubit_depolarizing({self.probability}) "
f"{target_qubit_0}, {target_qubit_1}"
)
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
SI = np.array([[1.0, 0.0], [0.0, 1.0]], dtype=complex)
SX = np.array([[0.0, 1.0], [1.0, 0.0]], dtype=complex)
SY = np.array([[0.0, -1.0j], [1.0j, 0.0]], dtype=complex)
SZ = np.array([[1.0, 0.0], [0.0, -1.0]], dtype=complex)
K_list_single = [SI, SX, SY, SZ]
K_list = [np.kron(i, j) for i in K_list_single for j in K_list_single]
K_list[0] *= np.sqrt(1 - self._probability)
K_list[1:] = [np.sqrt(self._probability / 15) * i for i in K_list[1:]]
return K_list
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 2
[docs]
@staticmethod
@circuit.subroutine(register=True)
def two_qubit_depolarizing(
target1: QubitInput, target2: QubitInput, probability: float
) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target1 (QubitInput): Target qubit 1.
target2 (QubitInput): Target qubit 2.
probability (float): Probability of two-qubit depolarizing.
Returns:
Iterable[Instruction]: `Iterable` of Depolarizing instructions.
Examples:
>>> circ = Circuit().two_qubit_depolarizing(0, 1, probability=0.1)
"""
return [
Instruction(
Noise.TwoQubitDepolarizing(probability=probability), target=[target1, target2]
)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return TwoQubitDepolarizing(probability=_substitute_value(self._probability, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return TwoQubitDepolarizing(probability=_parameter_from_dict(noise["probability"]))
Noise.register_noise(TwoQubitDepolarizing)
[docs]
class TwoQubitDephasing(SingleProbabilisticNoise_34):
r"""Two-Qubit Dephasing noise channel which transforms a
density matrix :math:`\\rho` according to:
.. math::
\\rho \\Rightarrow (1-p) \\rho + p/3 (
IZ \\rho IZ^{\\dagger} + ZI \\rho ZI^{\\dagger} + ZZ \\rho ZZ^{\\dagger})
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
Z = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & -1
\\end{matrix}
\\right)
p = \\text{probability}
This noise channel is shown as `DEPH` in circuit diagrams.
"""
def __init__(self, probability: Union[FreeParameterExpression, float]):
super().__init__(
probability=probability,
qubit_count=None,
ascii_symbols=[_ascii_representation("DEPH", [probability])] * 2,
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.TwoQubitDephasing.construct(
targets=[target[0], target[1]], probability=self.probability
)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit_0 = serialization_properties.format_target(int(target[0]))
target_qubit_1 = serialization_properties.format_target(int(target[1]))
return (
f"#pragma braket noise two_qubit_dephasing({self.probability}) "
f"{target_qubit_0}, {target_qubit_1}"
)
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
SI = np.array([[1.0, 0.0], [0.0, 1.0]], dtype=complex)
SZ = np.array([[1.0, 0.0], [0.0, -1.0]], dtype=complex)
K0 = np.sqrt(1 - self._probability) * np.kron(SI, SI)
K1 = np.sqrt(self._probability / 3) * np.kron(SI, SZ)
K2 = np.sqrt(self._probability / 3) * np.kron(SZ, SI)
K3 = np.sqrt(self._probability / 3) * np.kron(SZ, SZ)
return [K0, K1, K2, K3]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 2
[docs]
@staticmethod
@circuit.subroutine(register=True)
def two_qubit_dephasing(
target1: QubitInput, target2: QubitInput, probability: float
) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target1 (QubitInput): Target qubit 1.
target2 (QubitInput): Target qubit 2.
probability (float): Probability of two-qubit dephasing.
Returns:
Iterable[Instruction]: `Iterable` of Dephasing instructions.
Examples:
>>> circ = Circuit().two_qubit_dephasing(0, 1, probability=0.1)
"""
return [
Instruction(Noise.TwoQubitDephasing(probability=probability), target=[target1, target2])
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return TwoQubitDephasing(probability=_substitute_value(self._probability, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return TwoQubitDephasing(probability=_parameter_from_dict(noise["probability"]))
Noise.register_noise(TwoQubitDephasing)
[docs]
class TwoQubitPauliChannel(MultiQubitPauliNoise):
r"""Two-Qubit Pauli noise channel which transforms a
density matrix :math:`\\rho` according to:
.. math::
\\rho \\Rightarrow (1-p) \\rho +
p_{IX} IX \\rho IX^{\\dagger} +
p_{IY} IY \\rho IY^{\\dagger} +
p_{IZ} IZ \\rho IZ^{\\dagger} +
p_{XI} XI \\rho XI^{\\dagger} +
p_{XX} XX \\rho XX^{\\dagger} +
p_{XY} XY \\rho XY^{\\dagger} +
p_{XZ} XZ \\rho XZ^{\\dagger} +
p_{YI} YI \\rho YI^{\\dagger} +
p_{YX} YX \\rho YX^{\\dagger} +
p_{YY} YY \\rho YY^{\\dagger} +
p_{YZ} YZ \\rho YZ^{\\dagger} +
p_{ZI} ZI \\rho ZI^{\\dagger} +
p_{ZX} ZX \\rho ZX^{\\dagger} +
p_{ZY} ZY \\rho ZY^{\\dagger} +
p_{ZZ} ZZ \\rho ZZ^{\\dagger})
where
.. math::
I = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & 1
\\end{matrix}
\\right)
X = \\left(
\\begin{matrix}
0 & 1 \\\\
1 & 0
\\end{matrix}
\\right)
Y = \\left(
\\begin{matrix}
0 & -i \\\\
i & 0
\\end{matrix}
\\right)
Z = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & -1
\\end{matrix}
\\right)
p = \\text{sum of all probabilities}
This noise channel is shown as `PC_2({"pauli_string": probability})` in circuit diagrams.
"""
_paulis: ClassVar = {
"I": np.array([[1.0, 0.0], [0.0, 1.0]], dtype=complex),
"X": np.array([[0.0, 1.0], [1.0, 0.0]], dtype=complex),
"Y": np.array([[0.0, -1.0j], [1.0j, 0.0]], dtype=complex),
"Z": np.array([[1.0, 0.0], [0.0, -1.0]], dtype=complex),
}
_tensor_products_strings = itertools.product(_paulis.keys(), repeat=2)
_names_list: ClassVar = ["".join(x) for x in _tensor_products_strings]
def __init__(self, probabilities: dict[str, float]):
super().__init__(
probabilities=probabilities,
qubit_count=None,
ascii_symbols=[
f"PC2({probabilities})",
f"PC2({probabilities})",
],
)
self._matrix = None
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
if self._matrix is not None:
return self._matrix
total_prob = sum(self._probabilities.values())
K_list = [np.sqrt(1 - total_prob) * np.identity(4)] # "II" element
for pstring in self._names_list[1:]: # ignore "II"
if pstring in self._probabilities:
mat = np.sqrt(self._probabilities[pstring]) * np.kron(
self._paulis[pstring[0]], self._paulis[pstring[1]]
)
K_list.append(mat)
else:
K_list.append(np.zeros((4, 4)))
self._matrix = K_list
return self._matrix
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.MultiQubitPauliChannel.construct(
targets=[target[0], target[1]], probabilities=self.probabilities
)
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 2
[docs]
@staticmethod
@circuit.subroutine(register=True)
def two_qubit_pauli_channel(
target1: QubitInput, target2: QubitInput, probabilities: dict[str, float]
) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target1 (QubitInput): Target qubit 1.
target2 (QubitInput): Target qubit 2.
probabilities (dict[str, float]): Probability of two-qubit Pauli channel.
Returns:
Iterable[Instruction]: `Iterable` of Depolarizing instructions.
Examples:
>>> circ = Circuit().two_qubit_pauli_channel(0, 1, {"XX": 0.1})
"""
return [
Instruction(
Noise.TwoQubitPauliChannel(probabilities=probabilities),
target=[target1, target2],
)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
probabilities = {
pauli_string: _substitute_value(prob, **kwargs)
for pauli_string, prob in self._probabilities.items()
}
return TwoQubitPauliChannel(probabilities=probabilities)
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
probabilities = {
pauli_string: _parameter_from_dict(prob)
for pauli_string, prob in noise["probabilities"].items()
}
return TwoQubitPauliChannel(probabilities=probabilities)
Noise.register_noise(TwoQubitPauliChannel)
[docs]
class AmplitudeDamping(DampingNoise):
r"""AmplitudeDamping noise channel which transforms a density matrix :math:`\\rho` according to:
.. math:: \\rho \\Rightarrow E_0 \\rho E_0^{\\dagger} + E_1 \\rho E_1^{\\dagger}
where
.. math::
E_0 = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & \\sqrt{1-\\gamma}
\\end{matrix}
\\right)
E_1 = \\left(
\\begin{matrix}
0 & \\sqrt{\\gamma} \\\\
0 & 0
\\end{matrix}
\\right)
This noise channel is shown as `AD` in circuit diagrams.
"""
def __init__(self, gamma: Union[FreeParameterExpression, float]):
super().__init__(
gamma=gamma,
qubit_count=None,
ascii_symbols=[_ascii_representation("AD", [gamma])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.AmplitudeDamping.construct(target=target[0], gamma=self.gamma)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return f"#pragma braket noise amplitude_damping({self.gamma}) {target_qubit}"
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.array([[1.0, 0.0], [0.0, np.sqrt(1 - self.gamma)]], dtype=complex)
K1 = np.array([[0.0, np.sqrt(self.gamma)], [0.0, 0.0]], dtype=complex)
return [K0, K1]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def amplitude_damping(target: QubitSetInput, gamma: float) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s).
gamma (float): decaying rate of the amplitude damping channel.
Returns:
Iterable[Instruction]: `Iterable` of AmplitudeDamping instructions.
Examples:
>>> circ = Circuit().amplitude_damping(0, gamma=0.1)
"""
return [
Instruction(Noise.AmplitudeDamping(gamma=gamma), target=qubit)
for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return AmplitudeDamping(gamma=_substitute_value(self._gamma, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return AmplitudeDamping(gamma=_parameter_from_dict(noise["gamma"]))
Noise.register_noise(AmplitudeDamping)
[docs]
class GeneralizedAmplitudeDamping(GeneralizedAmplitudeDampingNoise):
r"""Generalized AmplitudeDamping noise channel which transforms a
density matrix :math:`\\rho` according to:
.. math:: \\rho \\Rightarrow E_0 \\rho E_0^{\\dagger} + E_1 \\rho E_1^{\\dagger}
+ E_2 \\rho E_2^{\\dagger} + E_3 \\rho E_3^{\\dagger}
where
.. math::
E_0 = \\sqrt(probability)\\left(
\\begin{matrix}
1 & 0 \\\\
0 & \\sqrt{1-\\gamma}
\\end{matrix}
\\right)
E_1 = \\sqrt(probability)\\left(
\\begin{matrix}
0 & \\sqrt{\\gamma} \\\\
0 & 0
\\end{matrix}
\\right)
E_2 = \\sqrt(1-probability)\\left(
\\begin{matrix}
\\sqrt{1-\\gamma} & 0 \\\\
0 & 1
\\end{matrix}
\\right)
E_3 = \\sqrt(1-probability)\\left(
\\begin{matrix}
0 & 0 \\\\
\\sqrt{\\gamma} & 0
\\end{matrix}
\\right)
This noise channel is shown as `GAD` in circuit diagrams.
"""
def __init__(
self,
gamma: Union[FreeParameterExpression, float],
probability: Union[FreeParameterExpression, float],
):
super().__init__(
gamma=gamma,
probability=probability,
qubit_count=None,
ascii_symbols=[_ascii_representation("GAD", [gamma, probability])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.GeneralizedAmplitudeDamping.construct(
target=target[0], gamma=self.gamma, probability=self.probability
)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return (
"#pragma braket noise generalized_amplitude_damping("
f"{self.gamma}, {self.probability}) {target_qubit}"
)
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.sqrt(self.probability) * np.array(
[[1.0, 0.0], [0.0, np.sqrt(1 - self.gamma)]], dtype=complex
)
K1 = np.sqrt(self.probability) * np.array(
[[0.0, np.sqrt(self.gamma)], [0.0, 0.0]], dtype=complex
)
K2 = np.sqrt(1 - self.probability) * np.array([[np.sqrt(1 - self.gamma), 0.0], [0.0, 1.0]])
K3 = np.sqrt(1 - self.probability) * np.array([[0.0, 0.0], [np.sqrt(self.gamma), 0.0]])
return [K0, K1, K2, K3]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def generalized_amplitude_damping(
target: QubitSetInput, gamma: float, probability: float
) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s).
gamma (float): The damping rate of the amplitude damping channel.
probability(float): Probability of the system being excited by the environment.
Returns:
Iterable[Instruction]: `Iterable` of GeneralizedAmplitudeDamping instructions.
Examples:
>>> circ = Circuit().generalized_amplitude_damping(0, gamma=0.1, probability = 0.9)
"""
return [
Instruction(
Noise.GeneralizedAmplitudeDamping(gamma=gamma, probability=probability),
target=qubit,
)
for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
gamma = _substitute_value(self._gamma, **kwargs)
probability = _substitute_value(self._probability, **kwargs)
return GeneralizedAmplitudeDamping(gamma=gamma, probability=probability)
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return GeneralizedAmplitudeDamping(
gamma=_parameter_from_dict(noise["gamma"]),
probability=_parameter_from_dict(noise["probability"]),
)
Noise.register_noise(GeneralizedAmplitudeDamping)
[docs]
class PhaseDamping(DampingNoise):
r"""Phase damping noise channel which transforms a density matrix :math:`\\rho` according to:
.. math:: \\rho \\Rightarrow E_0 \\rho E_0^{\\dagger} + E_1 \\rho E_1^{\\dagger}
where
.. math::
E_0 = \\left(
\\begin{matrix}
1 & 0 \\\\
0 & \\sqrt{1-gamma}
\\end{matrix}
\\right)
E_1 = \\left(
\\begin{matrix}
0 & 0 \\\\
0 & \\sqrt{gamma}
\\end{matrix}
\\right)
p = \\text{probability}
This noise channel is shown as `PD` in circuit diagrams.
"""
def __init__(self, gamma: Union[FreeParameterExpression, float]):
super().__init__(
gamma=gamma,
qubit_count=None,
ascii_symbols=[_ascii_representation("PD", [gamma])],
)
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.PhaseDamping.construct(target=target[0], gamma=self.gamma)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
target_qubit = serialization_properties.format_target(int(target[0]))
return f"#pragma braket noise phase_damping({self.gamma}) {target_qubit}"
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
K0 = np.array([[1.0, 0.0], [0.0, np.sqrt(1 - self.gamma)]], dtype=complex)
K1 = np.array([[0.0, 0.0], [0.0, np.sqrt(self.gamma)]], dtype=complex)
return [K0, K1]
[docs]
@staticmethod
def fixed_qubit_count() -> int:
return 1
[docs]
@staticmethod
@circuit.subroutine(register=True)
def phase_damping(target: QubitSetInput, gamma: float) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
target (QubitSetInput): Target qubit(s)
gamma (float): Probability of phase damping.
Returns:
Iterable[Instruction]: `Iterable` of PhaseDamping instructions.
Examples:
>>> circ = Circuit().phase_damping(0, gamma=0.1)
"""
return [
Instruction(Noise.PhaseDamping(gamma=gamma), target=qubit) for qubit in QubitSet(target)
]
[docs]
def bind_values(self, **kwargs) -> Noise:
"""Takes in parameters and attempts to assign them to values.
Returns:
Noise: A new Noise object of the same type with the requested
parameters bound.
"""
return PhaseDamping(gamma=_substitute_value(self._gamma, **kwargs))
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
return PhaseDamping(gamma=_parameter_from_dict(noise["gamma"]))
Noise.register_noise(PhaseDamping)
[docs]
class Kraus(Noise):
"""User-defined noise channel that uses the provided matrices as Kraus operators
This noise channel is shown as `NK` in circuit diagrams.
"""
def __init__(self, matrices: Iterable[np.ndarray], display_name: str = "KR"):
"""Inits `Kraus`.
Args:
matrices (Iterable[ndarray]): A list of matrices that define a noise
channel. These matrices need to satisfy the requirement of CPTP map.
display_name (str): Name to be used for an instance of this general noise
channel for circuit diagrams. Defaults to `KR`.
Raises:
ValueError: If any matrix in `matrices` is not a two-dimensional square
matrix,
or has a dimension length which is not a positive exponent of 2,
or the `matrices` do not satisfy CPTP condition.
"""
for matrix in matrices:
verify_quantum_operator_matrix_dimensions(matrix)
if int(np.log2(matrix.shape[0])) != int(np.log2(matrices[0].shape[0])):
raise ValueError(f"all matrices in {matrices} must have the same shape")
self._matrices = [np.array(matrix, dtype=complex) for matrix in matrices]
self._display_name = display_name
qubit_count = int(np.log2(self._matrices[0].shape[0]))
if qubit_count > 2:
raise ValueError("Kraus operators with more than two qubits are not supported.")
if len(matrices) > 2 ** (2 * qubit_count):
raise ValueError("The number of Kraus operators is beyond limit.")
if not is_cptp(self._matrices):
raise ValueError(
"The input matrices do not define a completely-positive trace-preserving map."
)
super().__init__(qubit_count=qubit_count, ascii_symbols=[display_name] * qubit_count)
[docs]
def to_matrix(self) -> Iterable[np.ndarray]:
"""Returns a matrix representation of this noise.
Returns:
Iterable[ndarray]: A list of matrix representations of this noise.
"""
return self._matrices
def _to_jaqcd(self, target: QubitSet) -> Any:
return ir.Kraus.construct(
targets=list(target),
matrices=Kraus._transform_matrix_to_ir(self._matrices),
)
def _to_openqasm(
self, target: QubitSet, serialization_properties: OpenQASMSerializationProperties
) -> str:
matrix_list = ", ".join(
np.array2string(
matrix,
separator=", ",
formatter={"all": lambda x: format_complex(x)},
).replace("\n", "")
for matrix in self._matrices
)
qubit_list = ", ".join(
serialization_properties.format_target(int(qubit)) for qubit in target
)
return f"#pragma braket noise kraus({matrix_list}) {qubit_list}"
@staticmethod
def _transform_matrix_to_ir(matrices: Iterable[np.ndarray]) -> list:
serializable = []
for matrix in matrices:
matrix_as_list = [
[[element.real, element.imag] for element in row] for row in matrix.tolist()
]
serializable.append(matrix_as_list)
return serializable
[docs]
@staticmethod
@circuit.subroutine(register=True)
def kraus(
targets: QubitSetInput, matrices: Iterable[np.array], display_name: str = "KR"
) -> Iterable[Instruction]:
"""Registers this function into the circuit class.
Args:
targets (QubitSetInput): Target qubit(s)
matrices (Iterable[array]): Matrices that define a general noise channel.
display_name (str): The display name.
Returns:
Iterable[Instruction]: `Iterable` of Kraus instructions.
Examples:
>>> K0 = np.eye(4) * np.sqrt(0.9)
>>> K1 = np.kron([[1., 0.],[0., 1.]], [[0., 1.],[1., 0.]]) * np.sqrt(0.1)
>>> circ = Circuit().kraus([1, 0], matrices=[K0, K1])
"""
if 2 ** len(targets) != matrices[0].shape[0]:
raise ValueError(
"Dimensions of the supplied Kraus matrices are incompatible with the targets"
)
return Instruction(
Noise.Kraus(matrices=matrices, display_name=display_name), target=targets
)
[docs]
def to_dict(self) -> dict:
"""Converts this object into a dictionary representation. Not implemented at this time.
Returns:
dict: Not implemented at this time..
"""
raise NotImplementedError
[docs]
@classmethod
def from_dict(cls, noise: dict) -> Noise:
"""Converts a dictionary representation of this class into this class.
Args:
noise(dict): The dictionary representation of this noise.
Returns:
Noise: A Noise object that represents the passed in dictionary.
"""
raise NotImplementedError
Noise.register_noise(Kraus)
def _ascii_representation(
noise: str, parameters: list[Union[FreeParameterExpression, float]]
) -> str:
"""Generates a formatted ascii representation of a noise.
Args:
noise (str): The name of the noise.
parameters (list[Union[FreeParameterExpression, float]]): The parameters to the noise.
Returns:
str: The ascii representation of the noise.
"""
param_list = [
(str(param) if isinstance(param, FreeParameterExpression) else f"{param:.2g}")
for param in parameters
]
param_str = ",".join(param_list)
return f"{noise}({param_str})"
def _substitute_value(expr: float, **kwargs) -> Union[FreeParameterExpression, float]:
return expr.subs(kwargs) if isinstance(expr, FreeParameterExpression) else expr
def _parameter_from_dict(parameter: Union[dict, float]) -> Union[FreeParameter, float]:
"""Converts a parameter from a dictionary if it's a FreeParameter, otherwise returns the float.
Args:
parameter(Union[dict, float]): The parameter to convert.
Returns:
Union[FreeParameter, float]: A FreeParameter representing the parameter, if the parameter
is a dictionary, otherwise returns the float.
"""
if isinstance(parameter, dict):
return FreeParameter.from_dict(parameter)
return parameter