quantus.metrics.robustness.consistency module

This module contains the implementation of the Consistency metric.

final class quantus.metrics.robustness.consistency.Consistency(discretise_func: Callable | None = None, abs: bool = True, normalise: bool = True, normalise_func: Callable[[ndarray], ndarray] | None = None, normalise_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 the Consistency metric which measures the expected local consistency, i.e., the probability of the prediction label for a given datapoint coinciding with the prediction labels of other data points that the same explanation is being attributed to. For example, if the explanation of a given image is “contains zebra”, the local consistency metric measures the probability a different image that the explanation “contains zebra” is being attributed to having the same prediction label.

Assumptions:

A used-defined discreization function is used to discretize continuous explanation spaces.

References:

Sanjoy Dasgupta et al.: “Framework for Evaluating Faithfulness of Local Explanations.” ICML (2022): 4794-4815.

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.
evaluation_category: What property/ explanation quality that this metric measures.

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, ...)	Compute additional arguments required for Consistency on batch-level.
`custom_postprocess`(*, model, x_batch, ...)	Implement this method if you need custom postprocessing of results or additional attributes.
`custom_preprocess`(*, model, x_batch, ...)	Implement this method if you need custom preprocessing of data, model alteration or simply for creating/initialising additional attributes or assertions.
`evaluate_batch`(a_batch, i_batch, ...)	This method performs XAI evaluation on a single batch of explanations.
`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.

__call__(model, x_batch: ndarray, y_batch: ndarray, a_batch: ndarray | None = None, s_batch: ndarray | None = None, channel_first: bool | None = None, explain_func: Callable | None = None, explain_func_kwargs: Dict | None = None, model_predict_kwargs: Dict | None = None, softmax: bool | None = True, device: str | None = None, batch_size: int = 64, **kwargs) → List[float]

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__(discretise_func: Callable | None = None, abs: bool = True, normalise: bool = True, normalise_func: Callable[[ndarray], ndarray] | None = None, normalise_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:

discretise_func: callable: Discretisation function applied to explantions. If None, the default value is used, default=top_n_sign.
normalise: boolean: Indicates whether normalise operation is applied on the attribution, default=True.
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={}.
return_aggregate: boolean: Indicates if an aggregated score should be computed over all instances.
aggregate_func: callable: Callable that aggregates the scores given an evaluation call.
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_batch_preprocess(model: ModelInterface, x_batch: ndarray, a_batch: ndarray, **kwargs) → Dict[str, ndarray]: Compute additional arguments required for Consistency on batch-level.

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

evaluate_batch(a_batch: ndarray, i_batch: ndarray, a_label_batch: ndarray, y_pred_classes: ndarray, **kwargs) → List[float]

This method performs XAI evaluation on a single batch of explanations. For more information on the specific logic, we refer the metric’s initialisation docstring.

Parameters:

a_batch:: Batch of explanation to be evaluated.
i_batch:: Batch of segmentations to be evaluated.
a_label_batch:: Batch of discretised attribution labels.
y_pred_classes:: The class predictions of the complete input dataset.
kwargs:: Unused.

Returns:

scores_batch:: Evaluation results.

evaluation_category: ClassVar[EvaluationCategory] = 'Robustness'

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

name: ClassVar[str] = 'Consistency'

score_direction: ClassVar[ScoreDirection] = 'lower'