quantus.metrics.axiomatic.input_invariance module

This module contains the implementation of the Input Invariance metric.

final class quantus.metrics.axiomatic.input_invariance.InputInvariance(abs: bool = False, normalise: bool = False, normalise_func: Callable[[ndarray], ndarray] | None = None, normalise_func_kwargs: Dict[str, Any] | None = None, input_shift: int | float = -1, perturb_func: Callable | None = None, perturb_func_kwargs: Dict[str, Any] | None = None, return_aggregate: bool = False, aggregate_func: Callable | None = None, default_plot_func: Callable | None = None, disable_warnings: bool = False, display_progressbar: bool = False, **kwargs)

Bases: Metric[List[float]]

Implementation of Completeness test by Kindermans et al., 2017.

To test for input invariance, we add a constant shift to the input data and a mean shift to the model bias, so that the output of the original model on the original data is equal to the output of the changed model on the shifted data. The metric returns True if batch attributions stayed unchanged too. Currently only supporting constant values for the shift.

References:

Pieter-Jan Kindermans et al.: “The (Un)reliability of Saliency Methods.” Explainable AI (2019): 267-280

Attributes:

_name: The name of the metric.
_data_applicability: The data types that the metric implementation currently supports.
_models: The model types that this metric can work with.
score_direction: How to interpret the scores, whether higher/ lower values are considered better.

Attributes:

disable_warnings: A helper to avoid polluting test outputs with warnings.
display_progressbar: A helper to avoid polluting test outputs with tqdm progress bars.
get_params: List parameters of metric.

Methods

`__call__`(model, x_batch, y_batch, a_batch[, ...])	This implementation represents the main logic of the metric and makes the class object callable.
`batch_preprocess`(data_batch)	If data_batch has no a_batch, will compute explanations.
`custom_batch_preprocess`(*, model, x_batch, ...)	Implement this method if you need custom preprocessing of data or simply for creating/initialising additional attributes or assertions before a data_batch can be evaluated.
`custom_postprocess`(*, model, x_batch, ...)	Implement this method if you need custom postprocessing of results or additional attributes.
`custom_preprocess`(**kwargs)	Additional explain_func assert, as the one in prepare() won't be executed when a_batch != None.
`evaluate_batch`(model, x_batch, y_batch, ...)	Evaluates model and attributes on a single data batch and returns the batched evaluation result.
`explain_batch`(model, x_batch, y_batch)	Compute explanations, normalise and take absolute (if was configured so during metric initialization.) This method should primarily be used if you need to generate additional explanation in metrics body. It encapsulates typical for Quantus pre- and postprocessing approach. It will do few things: - call model.shape_input (if ModelInterface instance was provided) - unwrap model (if ModelInterface instance was provided) - call explain_func - expand attribution channel - (optionally) normalise a_batch - (optionally) take np.abs of a_batch.
`general_preprocess`(model, x_batch, y_batch, ...)	Prepares all necessary variables for evaluation.
`generate_batches`(data, batch_size)	Creates iterator to iterate over all batched instances in data dictionary.
`interpret_scores`()	Get an interpretation of the scores.
`plot`([plot_func, show, path_to_save])	Basic plotting functionality for Metric class.

This implementation represents the main logic of the metric and makes the class object callable. It completes instance-wise evaluation of explanations (a_batch) with respect to input data (x_batch), output labels (y_batch) and a torch or tensorflow model (model).

Calls general_preprocess() with all relevant arguments, calls () on each instance, and saves results to evaluation_scores. Calls custom_postprocess() afterwards. Finally returns evaluation_scores.

Parameters:

model: torch.nn.Module, tf.keras.Model: A torch or tensorflow model that is subject to explanation.
x_batch: np.ndarray: A np.ndarray which contains the input data that are explained.
y_batch: np.ndarray: A np.ndarray which contains the output labels that are explained.
a_batch: np.ndarray, optional: A np.ndarray which contains pre-computed attributions i.e., explanations.
s_batch: np.ndarray, optional: A np.ndarray which contains segmentation masks that matches the input.
channel_first: boolean, optional: Indicates of the image dimensions are channel first, or channel last. Inferred from the input shape if None.
explain_func: callable: Callable generating attributions.
explain_func_kwargs: dict, optional: Keyword arguments to be passed to explain_func on call.
model_predict_kwargs: dict, optional: Keyword arguments to be passed to the model’s predict method.
softmax: boolean: Indicates whether to use softmax probabilities or logits in model prediction. This is used for this __call__ only and won’t be saved as attribute. If None, self.softmax is used.
device: string: Indicated the device on which a torch.Tensor is or will be allocated: “cpu” or “gpu”.
kwargs: optional: Keyword arguments.

Returns:

evaluation_scores: list: a list of Any with the evaluation scores of the concerned batch.
Examples:

# Minimal imports. >> import quantus >> from quantus import LeNet >> import torch

# Enable GPU. >> device = torch.device(“cuda:0” if torch.cuda.is_available() else “cpu”)

# Load a pre-trained LeNet classification model (architecture at quantus/helpers/models). >> model = LeNet() >> model.load_state_dict(torch.load(“tutorials/assets/pytests/mnist_model”))

# Load MNIST datasets and make loaders. >> test_set = torchvision.datasets.MNIST(root=’./sample_data’, download=True) >> test_loader = torch.utils.data.DataLoader(test_set, batch_size=24)

# Load a batch of inputs and outputs to use for XAI evaluation. >> x_batch, y_batch = iter(test_loader).next() >> x_batch, y_batch = x_batch.cpu().numpy(), y_batch.cpu().numpy()

# Generate Saliency attributions of the test set batch of the test set. >> a_batch_saliency = Saliency(model).attribute(inputs=x_batch, target=y_batch, abs=True).sum(axis=1) >> a_batch_saliency = a_batch_saliency.cpu().numpy()

# Initialise the metric and evaluate explanations by calling the metric instance. >> metric = Metric(abs=True, normalise=False) >> scores = metric(model=model, x_batch=x_batch, y_batch=y_batch, a_batch=a_batch_saliency)

__init__(abs: bool = False, normalise: bool = False, normalise_func: Callable[[ndarray], ndarray] | None = None, normalise_func_kwargs: Dict[str, Any] | None = None, input_shift: int | float = -1, perturb_func: Callable | None = None, perturb_func_kwargs: Dict[str, Any] | None = None, return_aggregate: bool = False, aggregate_func: Callable | None = None, default_plot_func: Callable | None = None, disable_warnings: bool = False, display_progressbar: bool = False, **kwargs)

Parameters:

abs: boolean: Indicates whether absolute operation is applied on the attribution, default=False.
normalise: boolean: Indicates whether normalise operation is applied on the attribution, default=False.
normalise_func: callable: Attribution normalisation function applied in case normalise=True. If normalise_func=None, the default value is used, default=normalise_by_max.
normalise_func_kwargs: dict: Keyword arguments to be passed to normalise_func on call, default={}.
input_shift: float, int: The value used to shift the input and the model bias as per the paper, default=-1.
perturb_func_kwargs: dict: Keyword arguments to be passed to perturb_func, default={}.
return_aggregate: boolean: Indicates if an aggregated score should be produced over all instances.
aggregate_func: callable: A Callable to aggregate the scores per instance to one float.
default_plot_func: callable: Callable that plots the metrics result.
disable_warnings: boolean: Indicates whether the warnings are printed, default=False.
display_progressbar: boolean: Indicates whether a tqdm-progress-bar is printed, default=False.
kwargs: optional: Keyword arguments.

custom_preprocess(**kwargs) → None: Additional explain_func assert, as the one in prepare() won’t be executed when a_batch != None.

data_applicability: ClassVar[Set[DataType]] = {DataType.IMAGE, DataType.TABULAR, DataType.TIMESERIES}

evaluate_batch(model: ModelInterface, x_batch: ndarray, y_batch: ndarray, a_batch: ndarray, **kwargs) → ndarray

Evaluates model and attributes on a single data batch and returns the batched evaluation result.

Parameters:

model: ModelInterface: A ModelInteface that is subject to explanation.
x_batch: np.ndarray: The input to be evaluated on a batch-basis.
y_batch: np.ndarray: The output to be evaluated on a batch-basis.
a_batch: np.ndarray: The explanation to be evaluated on a batch-basis.

Returns:

scores_batch: np.ndarray: The evaluation results.

evaluation_category: ClassVar[EvaluationCategory] = 'Axiomatic'

model_applicability: ClassVar[Set[ModelType]] = {ModelType.TF, ModelType.TORCH}

name: ClassVar[str] = 'Input Invariance'

score_direction: ClassVar[ScoreDirection] = 'higher'