art.estimators.classification.scikitlearn

This module implements the classifiers for scikit-learn models.

Base Class Scikit-learn

class art.estimators.classification.scikitlearn.ScikitlearnClassifier(model: sklearn.base.BaseEstimator, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0), use_logits: bool = False)

Class for scikit-learn classifier models.

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn DecisionTreeClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnDecisionTreeClassifier(model: sklearn.tree.DecisionTreeClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Decision Tree Classifier models.

__init__(model: sklearn.tree.DecisionTreeClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn Decision Tree Classifier model.

Parameters:
  • model – scikit-learn Decision Tree Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_classes_at_node(node_id: int) ndarray

Returns the classification for a given node.

Returns:

Major class in node.

get_decision_path(x: ndarray) ndarray

Returns the path through nodes in the tree when classifying x. Last one is leaf, first one root node.

Returns:

The indices of the nodes in the array structure of the tree.

get_feature_at_node(node_id: int) int

Returns the feature of given id for a node.

Returns:

Feature index of feature split in this node.

get_left_child(node_id: int) int

Returns the id of the left child node of node_id.

Returns:

The indices of the left child in the tree.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

get_right_child(node_id: int) int

Returns the id of the right child node of node_id.

Returns:

The indices of the right child in the tree.

get_samples_at_node(node_id: int) int

Returns the number of training samples mapped to a node.

Returns:

Number of samples mapped this node.

get_threshold_at_node(node_id: int) float

Returns the threshold of given id for a node.

Returns:

Threshold value of feature split in this node.

get_values_at_node(node_id: int) ndarray

Returns the feature of given id for a node.

Returns:

Normalized values at node node_id.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn ExtraTreeClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnExtraTreeClassifier(model: sklearn.tree.ExtraTreeClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Extra TreeClassifier Classifier models.

__init__(model: sklearn.tree.ExtraTreeClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn Extra TreeClassifier Classifier model.

Parameters:
  • model – scikit-learn Extra TreeClassifier Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_classes_at_node(node_id: int) ndarray

Returns the classification for a given node.

Returns:

Major class in node.

get_decision_path(x: ndarray) ndarray

Returns the path through nodes in the tree when classifying x. Last one is leaf, first one root node.

Returns:

The indices of the nodes in the array structure of the tree.

get_feature_at_node(node_id: int) int

Returns the feature of given id for a node.

Returns:

Feature index of feature split in this node.

get_left_child(node_id: int) int

Returns the id of the left child node of node_id.

Returns:

The indices of the left child in the tree.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

get_right_child(node_id: int) int

Returns the id of the right child node of node_id.

Returns:

The indices of the right child in the tree.

get_samples_at_node(node_id: int) int

Returns the number of training samples mapped to a node.

Returns:

Number of samples mapped this node.

get_threshold_at_node(node_id: int) float

Returns the threshold of given id for a node.

Returns:

Threshold value of feature split in this node.

get_values_at_node(node_id: int) ndarray

Returns the feature of given id for a node.

Returns:

Normalized values at node node_id.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn AdaBoostClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnAdaBoostClassifier(model: sklearn.ensemble.AdaBoostClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn AdaBoost Classifier models.

__init__(model: sklearn.ensemble.AdaBoostClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn AdaBoost Classifier model.

Parameters:
  • model – scikit-learn AdaBoost Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn BaggingClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnBaggingClassifier(model: sklearn.ensemble.BaggingClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Bagging Classifier models.

__init__(model: sklearn.ensemble.BaggingClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn Bagging Classifier model.

Parameters:
  • model – scikit-learn Bagging Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn ExtraTreesClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnExtraTreesClassifier(model: sklearn.ensemble.ExtraTreesClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Extra Trees Classifier models.

__init__(model: sklearn.ensemble.ExtraTreesClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Create a Classifier instance from a scikit-learn Extra Trees Classifier model.

Parameters:
  • model – scikit-learn Extra Trees Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

get_trees() List[Tree]

Get the decision trees.

Returns:

A list of decision trees.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn GradientBoostingClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnGradientBoostingClassifier(model: sklearn.ensemble.GradientBoostingClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Gradient Boosting Classifier models.

__init__(model: sklearn.ensemble.GradientBoostingClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn Gradient Boosting Classifier model.

Parameters:
  • model – scikit-learn Gradient Boosting Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

get_trees() List[Tree]

Get the decision trees.

Returns:

A list of decision trees.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn RandomForestClassifier Classifier

class art.estimators.classification.scikitlearn.ScikitlearnRandomForestClassifier(model: sklearn.ensemble.RandomForestClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Random Forest Classifier models.

__init__(model: sklearn.ensemble.RandomForestClassifier, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn Random Forest Classifier model.

Parameters:
  • model – scikit-learn Random Forest Classifier model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

get_trees() List[Tree]

Get the decision trees.

Returns:

A list of decision trees.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn LogisticRegression Classifier

class art.estimators.classification.scikitlearn.ScikitlearnLogisticRegression(model: sklearn.linear_model.LogisticRegression, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn Logistic Regression models.

__init__(model: sklearn.linear_model.LogisticRegression, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn Logistic Regression model.

Parameters:
  • model – scikit-learn LogisticRegression model

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

class_gradient(x: ndarray, label: int | List[int] | ndarray | None = None, **kwargs) ndarray

Compute per-class derivatives w.r.t. x.

Return type:

ndarray

Parameters:
  • x (ndarray) – Sample input with shape as expected by the model.

  • label – Index of a specific per-class derivative. If an integer is provided, the gradient of that class output is computed for all samples. If multiple values as provided, the first dimension should match the batch size of x, and each value will be used as target for its corresponding sample in x. If None, then gradients for all classes will be computed for each sample.

Returns:

Array of gradients of input features w.r.t. each class in the form (batch_size, nb_classes, input_shape) when computing for all classes, otherwise shape becomes (batch_size, 1, input_shape) when label parameter is specified.

Raises:
  • ValueError – If the model has not been fitted prior to calling this method or if the number of classes in the classifier is not known.

  • TypeError – If the requested label cannot be processed.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

static get_trainable_attribute_names() Tuple[str, str]

Get the names of trainable attributes.

Returns:

A tuple of trainable attributes.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

loss_gradient(x: ndarray, y: ndarray, **kwargs) ndarray

Compute the gradient of the loss function w.r.t. x.

Return type:

ndarray

Parameters:
  • x (ndarray) – Sample input with shape as expected by the model.

  • y (ndarray) – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes) or indices of shape (nb_samples,).

Returns:

Array of gradients of the same shape as x.

Raises:

ValueError – If the model has not been fitted prior to calling this method.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

Raises:

ValueError – If the classifier does not have methods predict or predict_proba.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.

Scikit-learn SVC Classifier

class art.estimators.classification.scikitlearn.ScikitlearnSVC(model: sklearn.svm.SVC | sklearn.svm.LinearSVC, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0))

Class for scikit-learn C-Support Vector Classification models.

__init__(model: sklearn.svm.SVC | sklearn.svm.LinearSVC, clip_values: CLIP_VALUES_TYPE | None = None, preprocessing_defences: Preprocessor | List[Preprocessor] | None = None, postprocessing_defences: Postprocessor | List[Postprocessor] | None = None, preprocessing: PREPROCESSING_TYPE = (0.0, 1.0)) None

Create a Classifier instance from a scikit-learn C-Support Vector Classification model.

Parameters:
  • model – scikit-learn C-Support Vector Classification model.

  • clip_values – Tuple of the form (min, max) representing the minimum and maximum values allowed for features.

  • preprocessing_defences – Preprocessing defence(s) to be applied by the classifier.

  • postprocessing_defences – Postprocessing defence(s) to be applied by the classifier.

  • preprocessing – Tuple of the form (subtrahend, divisor) of floats or np.ndarray of values to be used for data preprocessing. The first value will be subtracted from the input. The input will then be divided by the second one.

class_gradient(x: ndarray, label: int | List[int] | ndarray | None = None, **kwargs) ndarray

Compute per-class derivatives w.r.t. x.

Return type:

ndarray

Parameters:
  • x (ndarray) – Sample input with shape as expected by the model.

  • label – Index of a specific per-class derivative. If an integer is provided, the gradient of that class output is computed for all samples. If multiple values as provided, the first dimension should match the batch size of x, and each value will be used as target for its corresponding sample in x. If None, then gradients for all classes will be computed for each sample.

Returns:

Array of gradients of input features w.r.t. each class in the form (batch_size, nb_classes, input_shape) when computing for all classes, otherwise shape becomes (batch_size, 1, input_shape) when label parameter is specified.

property clip_values: CLIP_VALUES_TYPE | None

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() ScikitlearnClassifier

Create a copy of the classifier that can be refit from scratch.

Returns:

new estimator

compute_loss(x: ndarray, y: Any, **kwargs) ndarray

Compute the loss of the estimator for samples x.

Parameters:
  • x (ndarray) – Input samples.

  • y – Target values.

Returns:

Loss values.

Return type:

Format as expected by the model

compute_loss_from_predictions(pred: ndarray, y: ndarray, **kwargs) ndarray

Compute the loss of the estimator for predictions pred.

Return type:

ndarray

Parameters:
  • pred (ndarray) – Model predictions.

  • y (ndarray) – Target values.

Returns:

Loss values.

fit(*args, **kwargs)

Fit the classifier on the training set (x, y).

Parameters:
  • x – Training data.

  • y – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes).

  • kwargs – Dictionary of framework-specific arguments. These should be parameters supported by the fit function in sklearn classifier and will be passed to this function as such.

get_params() Dict[str, Any]

Get all parameters and their values of this estimator.

Returns:

A dictionary of string parameter names to their value.

property input_shape: Tuple[int, ...]

Return the shape of one input sample.

Returns:

Shape of one input sample.

loss_gradient(x: ndarray, y: ndarray, **kwargs) ndarray

Compute the gradient of the loss function w.r.t. x. Following equation (1) with lambda=0.

Return type:

ndarray

Parameters:
  • x (ndarray) – Sample input with shape as expected by the model.

  • y (ndarray) – Target values (class labels) one-hot-encoded of shape (nb_samples, nb_classes) or indices of shape (nb_samples,).

Returns:

Array of gradients of the same shape as x.

property model

Return the model.

Returns:

The model.

property nb_classes: int

Return the number of output classes.

Returns:

Number of classes in the data.

predict(*args, **kwargs)

Perform prediction for a batch of inputs.

Parameters:

x – Input samples.

Returns:

Array of predictions of shape (nb_inputs, nb_classes).

q_submatrix(rows: ndarray, cols: ndarray) ndarray

Returns the q submatrix of this SVM indexed by the arrays at rows and columns.

Return type:

ndarray

Parameters:
  • rows (ndarray) – The row vectors.

  • cols (ndarray) – The column vectors.

Returns:

A submatrix of Q.

reset() None

Resets the weights of the classifier so that it can be refit from scratch.

save(filename: str, path: str | None = None) None

Save a model to file in the format specific to the backend framework.

Parameters:
  • filename (str) – Name of the file where to store the model.

  • path – Path of the folder where to store the model. If no path is specified, the model will be stored in the default data location of the library ART_DATA_PATH.

set_params(**kwargs) None

Take a dictionary of parameters and apply checks before setting them as attributes.

Parameters:

kwargs – A dictionary of attributes.

property use_logits: bool

Return the Boolean for using logits.

Returns:

Boolean for using logits.