dense_ev

dense_ev implements expectation value measurements in Qiskit using optimal dense grouping arXiv:quant-ph/2305.11847. For an $m$-qubit operator that includes all Pauli strings in its decomposition, it provides a speedup of $2^m$ compared to naive evaluation of all strings, and $(3/2)^m$ compared to grouping based on qubit-wise commuting (QWC) families.

Installation

Create a virtual environment to sandbox the installation (optional):

python3 -m venv test-env && source ./test-env/bin/activate

Install from GitHub:
```bash
pip install git+ssh://git@github.com/atlytle/dense-ev.git

A test function comparing results using abelian and dense grouping with exact results is available:

from dense_ev.test_estimator import test_random

m = 3  # n_qubits
run_unit_tests(m)

Qiskit version compatibility

2024-01-07: Support for dense Pauli grouping in the Aer Estimator primitive merged to main branch. See usage example below.
2024-05-30: Functionality using Opflow and other deprecated classes has been removed. Package requirements changed from qiskit < 0.43.0 to qiskit (v1.0). Added qiskit_aer requirement.

Usage

Estimator support

The Aer implementation of Estimator is extended to incorporate dense Pauli grouping, and can be invoked using the keyword argument abelian_grouping="DENSE". See test_estimator.py for a more complete listing.

from dense_ev.estimator_from_aer import Estimator

...

# Dense result.
nshots = 200000
approx = False
grouping = "DENSE"
run_options = {"shots": nshots, "seed": seed}
estimator = Estimator(
    run_options=run_options, abelian_grouping=grouping, approximation=approx)
result_dense = estimator.run(state, H, shots=nshots).result().values

...