masking module

Module for the generation of masks for coupling cells

get_bin(x, n=0)[source]

Get the binary representation of x.

Parameters

  • x (int) –

  • n (int) – Minimum number of digits. If x needs less digits in binary, the rest is filled with zeros.

Returns

Return type

list of binary digits

iflow_strategy(d, repetitions=1)[source]

Generate a list of masks using the strategy of Gao et al. arXiv:2001.05486. Each dimension is numbered and the number is converted into the binary representation. Then, one bit after another, starting with the most significant bit, is converted into a boolean. This determines wether the dimension is transformed in this cell or not. In order to be able to test for all correlations, for each mask, the inverse mask is used too.

The minimal amount of cells will be used to represent all possible correlations.

Parameters

d (int) – number of dimensions

Returns

list of masks

Return type

list of list of bool

maximal_masking_strategy(d, repetitions=1)[source]

Generate a list of masks using the maximally complex option: (not really but within the reasonable options)

Each variable is transformed using all the others as parameters, looping over all variables. The size of the model is therefore linear with the number of dimensions

Parameters

  • d (int) – number of dimensions

  • repetitions (int) – number of repetitions

n_ary_mask(d, n, offset)[source]

Create a n-ary mask with n entries (a list of bool with each nth entry True)

Parameters

  • d (int) – numbers of entries in the mask

  • n (int) – period of the mask

  • offset (int) – offset of the True entries

Returns

True/False mask with each nth entry True

Return type

list of bool

n_ary_mask_strategy(d, n=2, repetitions=1)[source]

Generate an list of masks using the n-ary periodic strategy: every nth entry is True and each next mask is offset by one position with respect to the previous one. We loop for a given number of repetitions through all possible masks.

Parameters

  • d (int) – dimensionality of the space

  • n (int) – periodicity of the mask

  • repetitions (int) – number of iterations through all possible masks

Returns

list of masks

Return type

list of list of bool