qml.pauli¶
Overview¶
This module defines functions and classes for generating and manipulating
elements of the Pauli group. It also contains a subpackage pauli/grouping
for Pauliword partitioning functionality used in measurement optimization.
Functions¶

Performs a check if two Pauli words have the same 

Given a list of observables assumed to be valid Pauli words, determine if they are pairwise qubitwise commuting. 

Converts a binary vector of even dimension to an Observable instance. 

Transforms the Pauli word to diagonal form in the computational basis. 

Diagonalizes a list of qubitwise commutative groupings of Pauli strings. 

Diagonalizes a list of mutually qubitwise commutative Pauli words. 

Partitions a list of observables (Pauli operations and tensor products thereof) into groupings according to a binary relation (qubitwise commuting, fullycommuting, or anticommuting). 

Checks if an observable instance consists only of Pauli and Identity Operators. 

Checks if two Pauli words in the binary vector representation are qubitwise commutative. 

Converts a list of Pauli words to the binary vector representation and yields a row matrix of the binary vectors. 

Partitions then diagonalizes a list of Pauli words, facilitating simultaneous measurement of all observables within a partition. 

Partitions the \(n\)qubit Pauli group into qubitwise commuting terms. 

Decomposes a Hermitian matrix into a linear combination of Pauli operators. 

Generate the \(n\)qubit Pauli group. 

Multiply two Pauli words together and return the product as a Pauli word. 

Multiply two Pauli words together, and return both their product as a Pauli word and the global phase. 
Return the PauliSentence representation of an arithmetic operator or Hamiltonian. 


Converts a Pauli word to the binary vector representation. 

Convert a Pauli word from a tensor to its matrix representation. 

Convert a Pauli word to a string. 

Obtains the adjacency matrix for the complementary graph of the qubitwise commutativity graph for a given set of observables in the binary representation. 

Performs circuit implementation of diagonalizing unitary for a Pauli word. 

Add together identical terms in the Hamiltonian. 

Convert a string in terms of 
Classes¶

Class for partitioning a list of Pauli words according to some binary symmetric relation. 
Dictionary representing a linear combination of Pauli words, with the keys as PauliWord instances and the values correspond to coefficients. 


Immutable dictionary used to represent a Pauli Word, associating wires with their respective operators. 
PauliWord and PauliSentence¶
The singlequbit Pauli group consists of the four singlequbit Pauli operations
Identity
, PauliX
,
PauliY
, and PauliZ
. The \(n\)qubit
Pauli group is constructed by taking all possible \(N\)fold tensor products
of these elements. Elements of the \(n\)qubit Pauli group are known as
Pauli words, and have the form \(P_J = \otimes_{i=1}^{n}\sigma_i^{(J)}\),
where \(\sigma_i^{(J)}\) is one of the Pauli operators
(PauliX
, PauliY
,
PauliZ
) or identity (Identity
) acting on
the \(i^{th}\) qubit. The full \(n\)qubit Pauli group has size
\(4^n\) (neglecting the four possible global phases that may arise from
multiplication of its elements).
PauliWord
is a lightweight class which uses a dictionary
approach to represent Pauli words. A PauliWord
can be
instantiated by passing a dictionary of wires and their associated Pauli operators.
>>> from pennylane.pauli import PauliWord
>>> pw1 = qml.pauli.PauliWord({0:"X", 1:"Z"}) # [email protected]
>>> pw2 = qml.pauli.PauliWord({0:"Y", 1:"Z"}) # [email protected]
>>> pw1, pw2
(X(0) @ Z(1), Y(0) @ Z(1))
The purpose of this class is to efficiently compute products of Pauli words and obtain the matrix representation.
>>> pw1 * pw2
(Z(0), 1j)
>>> pw1.to_mat(wire_order=[0, 1])
array([[ 0, 0, 1, 0],
[ 0, 0, 0, 1],
[ 1, 0, 0, 0],
[ 0, 1, 0, 0]])
The PauliSentence
class represents linear combinations of
Pauli words. Using a similar dictionary based approach we can efficiently add, multiply
and extract the matrix of operators in this representation.
>>> ps1 = qml.pauli.PauliSentence({pw1: 1.2, pw2: 0.5j})
>>> ps2 = qml.pauli.PauliSentence({pw1: 1.2})
>>> ps1
1.2 * X(0) @ Z(1)
+ 0.5j * Y(0) @ Z(1)
>>> ps1 + ps2
0.0 * X(0) @ Z(1)
+ 0.5j * Y(0) @ Z(1)
>>> ps1 * ps2
1.44 * I
+ (0.6+0j) * Z(0)
>>> (ps1 + ps2).to_mat(wire_order=[0, 1])
array([[ 0. +0.j, 0. +0.j, 0.5+0.j, 0. +0.j],
[ 0. +0.j, 0. +0.j, 0. +0.j, 0.5+0.j],
[0.5+0.j, 0. +0.j, 0. +0.j, 0. +0.j],
[ 0. +0.j, 0.5+0.j, 0. +0.j, 0. +0.j]])
Graph colouring¶
A module for heuristic algorithms for colouring Pauli graphs.
A Pauli graph is a graph where vertices represent Pauli words and edges denote if a specified symmetric binary relation (e.g., commutation) is satisfied for the corresponding Pauli words. The graphcolouring problem is to assign a colour to each vertex such that no vertices of the same colour are connected, using the fewest number of colours (lowest “chromatic number”) as possible.
Functions¶

Performs graphcolouring using the Largest Degree First heuristic. 

Performs graphcolouring using the Recursive Largest Degree First heuristic. 
Grouping observables¶
Pauli words can be used for expressing a qubit Hamiltonian
.
A qubit Hamiltonian has the form \(H_{q} = \sum_{J} C_J P_J\) where
\(C_{J}\) are numerical coefficients, and \(P_J\) are Pauli words.
A list of Pauli words can be partitioned according to certain grouping
strategies. As an example, the group_observables()
function partitions
a list of observables (Pauli operations and tensor products thereof) into
groupings according to a binary relation (e.g., qubitwise commuting):
>>> observables = [qml.PauliY(0), qml.PauliX(0) @ qml.PauliX(1), qml.PauliZ(1)]
>>> obs_groupings = group_observables(observables)
>>> obs_groupings
[[Tensor(PauliX(wires=[0]), PauliX(wires=[1]))],
[PauliY(wires=[0]), PauliZ(wires=[1])]]
The \(C_{J}\) coefficients for each \(P_J\) Pauli word making up a Hamiltonian can also be specified along with further options, such as the Pauliword grouping method (e.g., qubitwise commuting) and the underlying graphcolouring algorithm (e.g., recursive largest first) used for creating the groups of observables:
>>> obs = [qml.PauliY(0), qml.PauliX(0) @ qml.PauliX(1), qml.PauliZ(1)]
>>> coeffs = [1.43, 4.21, 0.97]
>>> obs_groupings, coeffs_groupings = group_observables(obs, coeffs, 'qwc', 'rlf')
>>> obs_groupings
[[Tensor(PauliX(wires=[0]), PauliX(wires=[1]))],
[PauliY(wires=[0]), PauliZ(wires=[1])]]
>>> coeffs_groupings
[[4.21], [1.43, 0.97]]