art.estimators.gan

GAN Estimator API.

TensorFlowV2 GAN

class art.estimators.gan.TensorFlowV2GAN(generator: GENERATOR_TYPE, discriminator: CLASSIFIER_TYPE, generator_loss=None, discriminator_loss=None, generator_optimizer_fct=None, discriminator_optimizer_fct=None)

This class implements a GAN with the TensorFlow v2 framework.

__init__(generator: GENERATOR_TYPE, discriminator: CLASSIFIER_TYPE, generator_loss=None, discriminator_loss=None, generator_optimizer_fct=None, discriminator_optimizer_fct=None)

Initialization of a test TensorFlow v2 GAN

Parameters:

  • generator – a TensorFlow2 generator

  • discriminator – a TensorFlow v2 discriminator

  • generator_loss – the loss function to use for the generator

  • discriminator_loss – the loss function to use for the discriminator

  • generator_optimizer_fct – the optimizer function to use for the generator

  • discriminator_optimizer_fct – the optimizer function to use for the discriminator

property channels_first: bool
Returns:

Boolean to indicate index of the color channels in the sample x.

property clip_values: Optional[CLIP_VALUES_TYPE]

Return the clip values of the input samples.

Returns:

Clip values (min, max).

clone_for_refitting() → ESTIMATOR_TYPE

Clone estimator for refitting.

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.

property discriminator: CLASSIFIER_TYPE
Returns:

the discriminator

property discriminator_loss: tf.Tensor
Returns:

the loss fct used for the discriminator

property discriminator_optimizer_fct: tf.Tensor
Returns:

the optimizer function for the discriminator

fit(x: ndarray, y: ndarray, batch_size: int = 128, nb_epochs: int = 20, **kwargs) → None

Creates a generative model

Parameters:

  • x (ndarray) – the secret backdoor trigger that will produce the target

  • y (ndarray) – the target to produce when using the trigger

  • batch_size (int) – batch_size of images used to train generator

  • nb_epochs (int) – total number of iterations for performing the attack

fit_generator(generator: DataGenerator, nb_epochs: int = 20, **kwargs) → None

Fit the estimator using a generator yielding training batches. Implementations can provide framework-specific versions of this function to speed-up computation.

Parameters:

  • generator – Batch generator providing (x, y) for each epoch.

  • nb_epochs (int) – Number of training epochs.

property generator: GENERATOR_TYPE
Returns:

the generator

property generator_loss: tf.Tensor
Returns:

the loss fct used for the generator

property generator_optimizer_fct: tf.Tensor
Returns:

the optimizer function for the generator

get_activations(x: ndarray, layer: Union[int, str], batch_size: int, framework: bool = False) → ndarray

Return the output of a specific layer for samples x where layer is the index of the layer between 0 and nb_layers - 1 or the name of the layer. The number of layers can be determined by counting the results returned by calling `layer_names.

Return type:

ndarray

Parameters:

  • x (ndarray) – Samples

  • layer – Index or name of the layer.

  • batch_size (int) – Batch size.

  • framework (bool) – If true, return the intermediate tensor representation of the activation.

Returns:

The output of layer, where the first dimension is the batch size corresponding to x.

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, int]

Return the shape of one input sample.

Returns:

Shape of one input sample.

property layer_names: Optional[List[str]]

Return the names of the hidden layers in the model, if applicable.

Returns:

The names of the hidden layers in the model, input and output layers are ignored.

Warning

layer_names tries to infer the internal structure of the model. This feature comes with no guarantees on the correctness of the result. The intended order of the layers tries to match their order in the model, but this is not guaranteed either.

loss_gradient(x, y, **kwargs)

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

Parameters:

  • x (Format as expected by the model) – Samples.

  • y (Format as expected by the model) – Target values.

Returns:

Loss gradients w.r.t. x in the same format as x.

Return type:

Format as expected by the model

property model

Return the model.

Returns:

The model.

predict(x: ndarray, batch_size: int = 128, **kwargs) → ndarray

Generates a sample.

Return type:

ndarray

Parameters:

  • x (ndarray) – A input seed.

  • batch_size (int) – The batch size for predictions.

Returns:

The generated sample.

set_params(**kwargs) → None

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

Parameters:

kwargs – A dictionary of attributes.