rand_design.jl

RandDesign

Pauli frame randomised experimental ensemble for an experimental design.

Fields

• d::Design: Experimental design.
• pauli_randomisation_ensemble::Vector{Vector{Vector{Vector{String}}}}: For each tuple, experiment, and randomisation, a vector of Pauli frame randomisations between each layer in the randomised circuit.
• pauli_sign_ensemble::Vector{Vector{Vector{Vector{Bool}}}}: For each tuple, experiment, and randomisation, a vector of signs for the Paulis estimated by the randomised experiment.
• covariance_pauli_sign_ensemble::Vector{Vector{Vector{Vector{Bool}}}}: For each tuple, experiment, and randomisation, a vector of signs for the covariance Paulis estimated by the randomised experiment.
• job_indices::Vector{Vector{NTuple{3, Int}}}: For each job, a vector of indices (tuple, experiment, randomisation) for each circuit in the job.
• randomisations::Vector{Int}: Number of randomisations for the experiment set corresponding to each tuple in the design.
• shot_budget::Int: Shot budget for the randomised experimental ensemble.
• experiment_shots::Int: Number of shots for each experiment in the randomised experimental ensemble.
• seed::UInt64: Random seed used to generate the randomised experimental ensemble.

generate_rand_design(d::Design, randomisations::Vector{Int}, shot_budget::Integer, experiment_shots::Integer; kwargs...)
generate_rand_design(d::Design, min_randomisations::Integer, target_shot_budget::Integer, experiment_shots::Integer; kwargs...)

Returns a RandDesign object describing a Pauli frame randomised experimental ensemble for the design d. Supply either the number of randomisations randomisations and shot budget shot_budget output by get_randomisations, as well as the number of shots for each experiment experiment_shots, or the arguments of that function, namely the minimum number of randomisations min_randomisations, target shot budget target_shot_budget, and number of shots for each experiment experiment_shots.

Arguments

• d::Design: Experimental design.
• randomisations::Vector{Int}: Number of randomisations for each tuple in the design.
• shot_budget::Integer: Shot budget for the randomised experimental ensemble.
• min_randomisations::Integer: Minimum number of randomisations for each tuple in the design.
• target_shot_budget::Integer: Target shot budget for the randomised experimental ensemble.
• experiment_shots::Integer: Number of shots for each experiment in the randomised experimental ensemble.

Keyword arguments

• job_circuit_number::Integer = 200: Number of circuits in each job.
• seed::Union{UInt64, Nothing} = nothing: Random seed for the randomisation.
• diagnostics::Bool = false: Whether to print diagnostics.
• save_data::Bool = false: Whether to save the randomised design.

get_design(d_rand::RandDesign)

Returns the design corresponding to the randomised design d_rand.

get_meas_budget(d_rand::RandDesign)

Returns the measurement budget corresponding to the randomised design d_rand.

get_randomisations(d::Design, min_randomisations::Integer, target_shot_budget::Integer, experiment_shots::Integer; p_norm::Real = 2)

Returns the number of randomisations for each tuple in the experimental design d, given a target shot budget target_shot_budget, number of shots for each experiment experiment_shots, and minimum number of randomisations min_randomisations. The number of randomisations is greedily optimised to minimise the p_norm-norm to the shot weights associated with the experimental design.

Typically, the 2-norm works better for smaller shot budgets, and the 1-norm works better for larger shot budgets. The number of randomisations corresponds to unnormalised shot weights randomisations .* d.experiment_numbers.

get_stim_qiskit_ensemble(d_rand::RandDesign, qiskit_qubit_num::Integer, qiskit_qubit_map::Vector{Int})

Returns the Stim and Qiskit circuits for the randomised ensemble d_rand.

The Qiskit circuits act on qiskit_qubit_num qubits, which in general must be at least maximum(qiskit_qubit_map) + 1, and the mapping of the qubits upon which the circuits act to the qubits in the Qiskit circuits is controlled by qiskit_qubit_map, which is a vector whose length is the number of qubits upon which the circuits act.

append_qiskit!(qiskit_circuit::Py, l::Layer, qiskit_qubit_map::Vector{Int})

Appends the gates in the layer l to the Qiskit circuit qiskit_circuit, using the mapping to Qiskit qubits given by qiskit_qubit_map.

append_qiskit_gate!(qiskit_circuit, gate_type::String, gate_targets::Vector{Vector{Int}})

Appends the gates of the type gate_type to the Qiskit circuit qiskit_circuit with the targets specified by gate_targets.

calc_mapping_signs(experiment_mappings::Vector{Mapping}, randomised_circuit::Vector{Layer})

Returns the signs of the Paulis corresponding to the mappings in experiment_mappings for the randomised circuit randomised_circuit.

get_pauli_layer(pauli_string::String, n::Int16)

Returns a layer of Pauli gates corresponding to the supplied Pauli string pauli_string on n qubits.`

get_randomised_circuit(prep_layer::Layer, tuple_circuit::Vector{Layer}, meas_layer::Layer, pauli_randomisation::Vector{String})

Returns the randomised circuit corresponding to the preparation layer prep_layer, tuple circuit tuple_circuit, measurement layer meas_layer, interspersed with the Pauli frame randomisation layers specified by pauli_randomisation.

get_stim_qiskit_circuit(randomised_circuit::Vector{Layer}, gate_probabilities::Dict{Gate, Vector{Float64}}, qiskit_qubit_num::Integer, qiskit_qubit_map::Vector{Int}, noisy_prep::Bool, noisy_meas::Bool; extra_fields::Dict{Symbol, Any} = Dict{Symbol, Any}())

Generate the Stim circuit string and Qiskit circuit for a Pauli frame randomised circuit randomised_circuit.

The Stim circuit features noise on each of the non-randomisation gates specified by gate_probabilities, as well as noisy preparations if noisy_prep is true, and noisy measurements if noisy_meas is true. Qubit coordinates are specified by extra_fields if it contains a CodeParameters object.

The Qiskit circuit acts on qiskit_qubit_num qubits, which in general must be at least maximum(qiskit_qubit_map) + 1, and the mapping of the qubits upon which the circuit acts to the qubits in the Qiskit circuit is controlled by qiskit_qubit_map, which is a vector whose length is the number of qubits upon which the circuit acts.