fieldofnodes/RobustBlindVerification.jl

Go to VeriQuEST.jl

https://github.com/fieldofnodes/VeriQuEST.jl

RobustBlindVerification.jl

Use branch: software_user_exploration to explore the following code.

Quantum emulation

Quantum computation is an exciting field. There are many research institutes and companies working hard to push the state of the art into the era beyond the noisy intermediate-scale quantum (NISQ) era. This package is designed to emulate quantum computation classically. Specifically, RobustBlindVerification.jl is an emulator of measurement based quantum computation (MBQC). Different to the gate model, MBQC relies on entangled states, projective measurements and adaptivity to perform quantum computation.

Implementation of protocol 1 from Leichtle et al (Verifying BQP Computations on Noisy Devices with Minimal Overhead).

Example

Run the following commands along with relevant input to implement the blind verification.

Load packages, if not installed, then run ] add <PackageName>, for QuEST_jl, go to https://github.com/fieldofnodes/QuEST_jl for installation instructions.

using Pkg
Pkg.activate(".")

using QuEST
using Test
using StatsBase
using Graphs
using CairoMakie
using RobustBlindVerification

Define the quantum state type, either a state vector backend (StateVector()) or a density matrix backend (DensityMatrix())

  state_type = DensityMatrix()

Using the Julia Graphs package, implement a graph.

  num_vertices = N
  graph = Graph(num_vertices)

Define the flow. As long a qubit index is inputted and a qubit index is ouputed, any function will do. Note that Julia is indexed starting from $1$, but QuEST (the quantum simulator) is in C and indexed from $0$.

The forward_flow function takes a present qubit and returns the next qubit according to the flow.

function forward_flow(vertex)
    # Define function or data structure to output new vertex index
    return forward_vertex
end

The backward_flow package takes the current vertex and returns the previous vertex.

function backward_flow(vertex)
  # Define function or data structure to output old vertex index
  return backward_vertex
end

Define the input and output vertices as tuples. For input vertices, provide the indices and the classical inputs. For no inputs, provide empty brackets ()

  input_indices = () # a tuple of indices 
  input_values = () # a tuple of input values

Provide the vertices considered the outputs of the graph.

  output_indices = (output_index_1,output_index_2,...,output_index_M)

for $M$ output vertices.

Provide the angles known to the client only. Data structures encapsulated by [] are vectors in Julia.

  secret_angles = []

Define the total number of rounds, the number of computational rounds and the test round threshold value (e.g. $n$, $d$, abd $w$ in Leichtle et al). The number of test rounds is $t = n - d$.

  total_rounds,computation_rounds = n,d
  test_rounds_theshold = w

Contain all separate variables and parametes into the following NamedTuple. Do not change this, simply name all of the above variables as seen in the code snippets.

para= (
    graph=graph,
    forward_flow = forward_flow,
    backward_flow=backward_flow,
    input_indices = input_indices,
    input_values = input_values,
    output_indices =output_indices,
    secret_angles=secret_angles,
    state_type = state_type,
    total_rounds = total_rounds,
    computation_rounds = computation_rounds,
    test_rounds_theshold = test_rounds_theshold)

To just run the verification simulator

run_verification_simulator(para)

The outcome will be a named tuple

(test_verification = test_verification_outcome, computation_verification = computation_verification_outcome, mode_outcome = mode_outcome)

where test_verification_outcomeand computation_verification_outcome is of type Ok() or Abort() based on outcomes computed according to Leichtle et al. The outcome of the computation is mode_outcome.

For an example see grover_verification_script.jl.

To do