Skip to main content

​
Execution Primitives

As well as designing quantum programs, the Classiq quantum model contains classical instructions for the execution process.
When designing a model, you must specify classical instructions for the execution process to take part.

​
Sample

The sample execution primitive instructs the execution process to sample the state of the quantum program.
Upon execution, the results of a program with a sample primitive are of type ExecutionDetails, describing the measurement results of the quantum program. Ways to access these measurements:
  • The counts attribute allows access to the measurement results of all qubits. The qubit order of each state in counts is indicated by the counts_lsb_right Boolean flag.
  • The parsed_counts attribute contains parsed states according to the arithmetic information of the output registers.
  • The parsed_counts_of_outputs method allows access to the parsed counts of specific given outputs. It receives either a single output name or a tuple of output names.
  • The counts_of_qubits method allows access to results of specific qubits. The order of qubits in the measurement result is determined by their order in the qubits argument of the method.
  • The counts_of_output method is similar to counts_of_qubits, but receives an output name as an argument. Note it may only be used if the generated model has outputs.
  • The counts_of_multiple_outputs is similar to counts_of_output. It receives a tuple of output names, and returns the counts of all specified outputs, keyed by a tuple of states matching the requested outputs.
  • The counts_by_qubit_order method allows access to the counts attribute in the required qubit order.
  • The num_shots attribute is the sum of all of the resulting count fields.
The ExecutionDetails object also contains a dataframe property that displays the relevant information in a pandas dataframe for easy manipulation.

​
VQE

The vqe execution primitive instructs the execution process to perform the Variational Quantum Eigensolver (VQE) algorithm. Given a parametric quantum program (an ansatz) and an Hamiltonian, the algorithm tries to minimize the expectation value of the Hamiltonian with respect to the resulting quantum states of the quantum program. The vqe primitive accepts these parameters:
  • hamiltonian: The Hamiltonian with which to optimize.
  • initial_point: The initial parameter assignment. Default: None.
  • maximize: If True, maximizes the expectation value instead of minimizing it.
  • optimizer: The kind of optimizer to use: COBYLA, SPSA, L_BFGS_B, NELDER_MEAD, ADAM, or SLSQP.
  • max_iteration: The maximum number of optimizer iterations.
  • tolerance: The final accuracy in the optimization. Default: 0.
  • step_size: The step size for numerically calculating the gradient. Default: 0.
  • skip_compute_variance: If True, the optimizer does not compute the variance of the ansatz. Default: False.
  • alpha_cvar: The parameter for the CVaR[1] summarizing method. Default: 1.
The following example defines a quantum model with a single RX gate with the parameter θ\theta, and executes the vqe primitive with 0.3Z as the Hamiltonian.
The results of a program with a vqe primitive are of type VQESolverResult, which describes the algorithm results. It contains this information:
  • optimal_parameters: The optimal parameters found by the algorithm.
  • energy: The expectation value of the ansatz with the optimal parameters (the minimum/maximum eigenvalue of the Hamiltonian).
  • optimized_circuit_sample_results: The results of sampling the ansatz with the optimal parameters. If executed with state vector simulation, the inner field state_vector is the eigenstate of the Hamiltonian.
  • time: The execution time of the algorithm (in seconds).
  • num_shots: The number of shots used in each iteration.
  • intermediate_results: List of per-iteration results.
  • convergence_graph_str: A string representing the energy convergence graph (shown in the IDE).

​
IQAE

The iqae execution primitive instructs the execution process to perform the Iterative Quantum Amplitude Estimation algorithm [2]. Given AA such that A∣0⟩n∣0⟩=1−a∣ψ0⟩n∣0⟩+a∣ψ1⟩n∣1⟩A|0\rangle_n|0\rangle = \sqrt{1-a}|\psi_0\rangle_n|0\rangle + \sqrt{a}|\psi_1\rangle_n|1\rangle, the algorithm tries to estimate aa by iteratively sampling QkAQ^kA, where Q=AS0A†Sψ0Q=AS_0A^{\dagger}S_{\psi_0} and kk is an integer variable.
The iqae primitive assumes you have correctly defined the quantum model; i.e., QkAQ^kA, where kk is specified by adding power="k" to the function parameters of the desirable function. In addition, the only output port should be the last qubit.
There are two parameters to the iqae primitive: epsilon specifies the target accuracy, and alpha specifies the confidence level (meaning the precision probability is 1−α1 - \alpha). The following example defines A=RY(θ)A = RY(\theta) and Q=RY(2θ)Q = RY(2\theta). The estimation result should be a=sin⁡2(θ2)a = \sin^2(\frac{\theta}{2}), as A∣0⟩=cos⁡θ2∣0⟩+sin⁡θ2∣1⟩A|0\rangle = \cos\frac{\theta}{2}|0\rangle + \sin\frac{\theta}{2}|1\rangle. The results of a program with an iqae primitive are of type IQAEResult, which describes the algorithm results. It contains this information:
  • estimation: The estimated value of aa.
  • confidence_interval: The confidence interval for the value of aa.
  • iterations_data: List of per-iteration information. Each item contains:
    • grover_iterations: The value of kk for this iteration.
    • sample_results: The results of sampling QkAQ^kA in this iteration.
  • warnings: List of warnings yielded throughout the algorithm execution, such as reaching the maximum number of iterations.

​
References

[1] Barkoutsos, P. K. et al., Improving variational quantum optimization using CVaR, Quantum 4, 256 (2019). [2] Grinko, D., Gacon, J., Zoufal, C. et al., Iterative quantum amplitude estimation, npj Quantum Inf 7, 52 (2021).