This is a super class for all data-based interpreting methods. Implemented are the following methods:

  • Deep Learning Important Features (DeepLift)

  • Layer-wise Relevance Propagation (LRP)

  • Gradient-based methods:

    • Vanilla gradients including 'Gradients x Input' (Gradient)

    • Smoothed gradients including 'SmoothGrad x Input' (SmoothGrad)

Public fields

data

The passed data as a torch tensor in the given data type (dtype) to be interpreted with the selected method.

converter

An instance of the R6 class Converter.

dtype

The data type for the calculations. Either 'float' for torch::torch_float or 'double' for torch::torch_double.

channels_first

The format of the given date, i.e. channels on last dimension (FALSE) or after the batch dimension (TRUE). If the data has no channels, the default value TRUE is used.

ignore_last_act

A boolean value to include the last activation into all the calculations, or not (default: TRUE). In some cases, the last activation leads to a saturation problem.

result

The methods result of the given data as a torch tensor of size (batch_size, dim_in, dim_out) in the given data type (dtype).

output_idx

This vector determines for which outputs the method will be applied. By default (NULL), all outputs (but limited to the first 10) are considered.

Methods


Method new()

Create a new instance of this super class.

Usage

InterpretingMethod$new(
  converter,
  data,
  channels_first = TRUE,
  output_idx = NULL,
  ignore_last_act = TRUE,
  dtype = "float"
)

Arguments

converter

An instance of the R6 class Converter.

data

The data for which this method is to be applied. It has to be an array or array-like format of size (batch_size, dim_in).

channels_first

The format of the given data, i.e. channels on last dimension (FALSE) or after the batch dimension (TRUE). If the data has no channels, use the default value TRUE.

output_idx

This vector determines for which output indices the method will be applied. By default (NULL), all outputs (but limited to the first 10) are considered.

ignore_last_act

A boolean value to include the last activation into all the calculations, or not (default: TRUE). In some cases, the last activation leads to a saturation problem.

dtype

dtype The data type for the calculations. Use either 'float' for torch::torch_float or 'double' for torch::torch_double.


Method get_result()

This function returns the result of this method for the given data either as an array ('array'), a torch tensor ('torch.tensor', or 'torch_tensor') of size (batch_size, dim_in, dim_out) or as a data.frame ('data.frame').

Usage

InterpretingMethod$get_result(type = "array")

Arguments

type

The data type of the result. Use one of 'array', 'torch.tensor', 'torch_tensor' or 'data.frame' (default: 'array').

Returns

The result of this method for the given data in the chosen type.


Method clone()

The objects of this class are cloneable with this method.

Usage

InterpretingMethod$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.