/Modified-IQAE

Code used for experiments in Modified IQAE paper

Primary LanguagePython

Modified Iterative Quantum Amplitude Estimation is Asymptotically Optimal

Install dependencies and run python modified_iqae.py. Edit config.yaml to modify experiment parameters. Running the script saves a folder that contains the input config file, a .csv file with estimations, and optionally the plots included in the paper.

Config parameters

All of the following keys must be present in the config with associated values.

  • runs (int): number of runs for the experiment.
  • shots (int): $N_{\text{shots}}$ in the paper, number of circuit trials per quantum execution. Not applicable to the 'classical' backend.
  • a_resolution (int): the resolution of the amplitude, which is the number of qubits used to express the amplitude in the circuit. Only accepts powers of 2.
  • a_step (int, List[int]): the size of the step taken between tested amplitudes in units of a_resolution. An integer a_step tests all amplitudes on $[0,1]$. A list of step values tests those specific amplitudes.
  • alpha (int): $\alpha$ as defined in the paper, representing the confidence level or maximum failure rate of the algorithm.
  • epsilons (List[float]): sequence of maximum error tolerances $\epsilon$.
  • confint_method {'chernoff, 'beta', 'all'}:

Optional parameters

Unless specified otherwise in the config, these keys will use their default values.

  • experiment_name (str): Default: <datetime>_modified-iqae.
  • results_path (str): Default: ./results/<experiment_name>.
  • simulator (str): specifies which Qiskit backend to use. Can also pass in the string 'classical' to simulate circuit measurements using a Bernoulli distribution. Default: aer_simulator.
  • compare (bool): whether to add the original (Qiskit builtin) IQAE to the experiment results. Default: True.
  • noise (float): adds a value to each tested amplitude $a$. This value is sampled from the uniform distribution on the closed interval $[a-$ noise $, a+$ noise $]$. If necessary, the endpoints are shifted to fit within the interval $[0,1]$. Default: 0.0.
  • plots (bool): whether to save plots. Default: True.
  • verbose (bool): prints intermediate algorithm values if True. Default: False. Use with caution for high numbers of experiments.

Cite this work

If you find our work useful, please consider using the following citation:

@misc{fukuzawa2022,
  title = {Modified {{Iterative Quantum Amplitude Estimation}} Is {{Asymptotically Optimal}}},
  author = {Fukuzawa, Shion and Ho, Christopher and Irani, Sandy and Zion, Jasen},
  year = {2022},
  month = aug,
  number = {arXiv:2208.14612},
  eprint = {2208.14612},
  eprinttype = {arxiv},
  primaryclass = {quant-ph},
  keywords = {Quantum Physics},
}