Boolean-Circuits-for-Differential-Privacy

The project develops a randomized response in boolean circuits solved via SAT problem solver engine, Z3.

Structure

In BooleanCircuit folder, there are three core classes.

BC.py
BC_Stat.py
RRclass.py

BC.py contains functions for generation.

Example

from BC import BcSequence

// seq is array for boolean circuits info, like [0,0]
bcSequence = BcSequence( seq )
// return representation of boolean circuit without any constrains.
bcSequence.get_out_bc() 
// return representation of boolean circuit with output != input constrain.
bcSequence.getbc()
// return representation of boolean circuit with output != input, input == true constrains.
bcSequence.get_t_bc() 
// return representation of boolean circuit with output != input, input == false constrain.
bcSequence.get_f_bc() 

// input_of_bc = 0 or 1, random is arrary of random bits. here length of boolean circuit is 2 (length of seq), like [0, 1].
output = bcSequence.get_output_with_random(input_of_bc, random) 
// randomizer is instance of Randomize class, inputs is like [0, 0, 0, 1]
output = bcSequence.get_outputs(inputs_of_bc, randomizer)

BC_Stat is the father class for BC.py containing iteration models function and probability calculations.

Example

// find_solution will save count
bcSequence.find_solution(bcSequence.getbc())
// after iteration, it could get running time
bcSequence.gettime()
bcSequence.find_probability()

More examples could found in BC for DP.ipynb (ipthyon notebook file)