A Python package designed for fast implementation of Pauli Twirling in Qiskit circuits.
- It includes a fast built-in dictionary of Pauli Twirling Sets for Qiskit two-qubit Clifford gates CX, CY, CZ, CH, DCX, CSX, CSdg, ECR, SWAP, iSWAP.
- It can generate Pauli Twirling Sets for any given two-qubit gate.
- If used with an instance of Qiskit's PassManager, it can implement randomized Pauli Twirls to a circuit containing two-qubit gates.
Pauli Twirling is a powerful error supression/mitigation technique for quantum circuits. By combining the results from many random, yet logically equivalent circuits, it converts an arbitrary noise channel into a stochastic Pauli error channel, thus suppressing off-diagonal coherent error contributions. It is implemented by sandwiching a two-qubit Clifford gate between randomly sampled single-qubit twirling gates drawn from a set of Pauli operations.
Crucially, the gates are selected such that the net operation, in the absence of noise, is unchanged. That is, the Pauli Twirling Set for a two-qubit Clifford gate is a set of four Pauli gates that when pre-pended and appended to a circuit containing the two-qubit gate, the unitary is equal to that of the single two-qubit gate alone.
- Purification of noisy entanglement and faithful teleportation via noisy channels (Bennet, et. al., 1996)
- Mixed state entanglement and quantum error correction (Bennet, et. al., 1996)
- Practical Implementations of Twirl Operations (Anwar, Short, and Jones, 2008)
- Scalable error mitigation for noisy quantum circuites produces competitive expectation values (Kim, et. al., 2021)
- Constructing smaller Pauli twirling sets for arbitrary error channels (Cai and Benjamin, 2018)
Native Pauli Twirling Sets are included for the following common two-qubit Clifford gates in Qiskit:
- CXGate (Controlled X Gate, a.k.a. CNOT Gate)
- CYGate (Controlled Y Gate)
- CZGate (Controlled Z Gate)
- CHGate (Controlled Hadamard Gate)
- DCXGate (Double Controlled X Gate, a.k.a., Double CNOT Gate)
- CSXGate (Controlled Square Root of X Gate)
- CSdgGate (Controlled S^dagger Gate)
- ECRGate (Echoed Cross-Resonance Gate)
- SWAPGate (Swap Gate)
- iSWAPGate (Imaginary Swap Gate)
Import Module:
from quantum_twirling import load_pauli_twirling_sets
Load the entire Pauli Twirling Sets dictionary:
quantum_twirling.load_pauli_twirling_sets.load_pauli_twirling_dict()
Output is the full dictionary.
Each gate's Pauli Twirling Sets can be accessed by passing the lowercase name of the gate to the dictionary,
e.g: cx_twirls = pauli_twirling_dict['cx']
Return list of CX Gate Pauli Twirling Sets:
quantum_twirling.load_pauli_twirling_sets.load_cxgate_twirls()
Return list of CY Gate Pauli Twirling Sets:
quantum_twirling.load_pauli_twirling_sets.load_cygate_twirls()
Return list of CZ Gate Pauli Twirling Sets:
quantum_twirling.load_pauli_twirling_sets.load_czgate_twirls()
Return list of CH Gate Pauli Twirling Sets:
quantum_twirling.load_pauli_twirling_sets.load_chgate_twirls()
Return list of DCX Gate Pauli Twirling Sets:
quantum_twirling.load_pauli_twirling_sets.load_dcxgate_twirls()
Return list of CSX Gate Pauli Twirling Sets:
quantum_twirling.load_pauli_twirling_sets.load_csxgate_twirls()
Return list of CSdg Gate Pauli Twirling Sets:
load_pauli_twirling_sets.load_csdggate_twirls()
Return list of ECR Gate Pauli Twirling Sets:
load_pauli_twirling_sets.load_ecrgate_twirls()
Return list of SWAP Gate Pauli Twirling Sets:
load_pauli_twirling_sets.load_swapgate_twirls()
Return list of iSWAP Gate Pauli Twirling Sets:
load_pauli_twirling_sets.load_iswapgate_twirls()
Import Module:
from quantum_twirling import generate_pauli_twirling_sets
generate_pauli_twirling_sets(TwoQubitGate())
- The
TwoQubitGate()
should to be a two-qubit instance of the Qiskit Gate class. - Finds Pauli Twirls for the given Gate; returns a list of the form:
[(gates,phase), (gates,phase), ...)]
.
Import Module:
from quantum_twirling import pauli_twirling
CLASS PauliTwirling(PassManager)
- Bases:
PassManager
(PassManager Class in Qiskit) To twirl all instances of a given two-qubit gate in a quantum circuit, attachPauliTwirling
pass to a QiskitPassManager
. Tell thePauliTwirling
pass to twirl gates of a certain type by providing a list of Pauli Twirling Sets for a given two-qubit gate. Seed the pass; Pauli Twirls are randomly applied to circuit.
pm = PassManager([PauliTwirling(gate_twirling_sets, seed=" ")])
To use the pass, call pm.run(qc)
where qc
is a quantum circuit.
All gates of type given in PassManager
will be randomly Pauli Twirled.
Subsequent calls of pm.run(qc)
will yield other circuits that are likewise randomly twirled.
import qiskit
import quantum_twirling
from quantum_twirling import load_pauli_twirling_sets
from quantum_twirling.generate_pauli_twirling_sets import generate_pauli_twirling_sets
from quantum_twirling.pauli_twirling import PauliTwirling
Load the entire existing dictionary with load_pauli_twirling_dict()
:
twirl_dict = load_pauli_twirling_sets.load_pauli_twirling_dict()
Load individual gate twirling sets with: cxgate_twirls()
, etc:
cx_twirls = load_pauli_twirling_sets.load_cxgate_twirls()
cy_twirls = load_pauli_twirling_sets.load_cygate_twirls()
cz_twirls = load_pauli_twirling_sets.load_czgate_twirls()
ch_twirls = load_pauli_twirling_sets.load_chgate_twirls()
dcx_twirls = load_pauli_twirling_sets.load_dcxgate_twirls()
csx_twirls = load_pauli_twirling_sets.load_csxgate_twirls()
csdg_twirls = load_pauli_twirling_sets.load_csdggate_twirls()
ecr_twirls = load_pauli_twirling_sets.load_ecrgate_twirls()
swap_twirls = load_pauli_twirling_sets.load_swapgate_twirls()
iswap_twirls = load_pauli_twirling_sets.load_iswapgate_twirls()
from quantum_twirling import generate_pauli_twirling_sets
twirling_sets = generate_pauli_twirling_sets(TwoQubitGate)
Import the PauliTwirling
class:
from quantum_twirling.pauli_twirling import PauliTwirling
Create a two-qubit quantum circuit to Pauli Twirl.
Note that the two-qubit gate in this circuit is a CX gate.
Thus, we can use our cx_dict
retrieved from get_pauli_twirling_dict('cx')
, earlier.
qc = QuantumCircuit(3)
qc.h(1)
qc.cx(1,0)
qc.cx(1,2)
qc.draw('mpl')
To use the pass, attach to a PassManager
.
Here, we do that and tell our PauliTwirling
pass to twirl CX gates in our cx_dict
.
pm = PassManager([PauliTwirling('cx_dict', seed=54321)])
Call pm.run
once and draw the circuit:
twirl_qc = pm.run(qc)
twirl_qc.draw('mpl')
Subsequent calls will yield other circuits that are likewise randomly twirled.