alibi_detect.od.pytorch.ensemble

AverageAggregator

Inherits from: BaseTransformTorch, Module

Constructor

AverageAggregator(self, weights: Optional[torch.Tensor] = None)

Name	Type	Default	Description
weights	Optional[torch.Tensor]	None	Optional parameter to weight the scores. If weights is left None then will be set to a vector of ones.

Methods

transform

transform(scores: torch.Tensor) -> torch.Tensor

Averages the scores of the detectors in an ensemble. If weights were passed in the __init__

then these are used to weight the scores.

Name	Type	Default	Description
scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor

BaseTransformTorch

Inherits from: Module

Constructor

BaseTransformTorch(self)

Methods

forward

forward(x: torch.Tensor) -> torch.Tensor

Name	Type	Default	Description
x	torch.Tensor

Returns

• Type: torch.Tensor

transform

transform(x: torch.Tensor)

Public transform method.

Name	Type	Default	Description
x	torch.Tensor		torch.Tensor array to be transformed

Ensembler

Inherits from: BaseTransformTorch, Module, FitMixinTorch, ABC

Constructor

Ensembler(self, normalizer: Optional[alibi_detect.od.pytorch.ensemble.BaseTransformTorch] = None, aggregator: alibi_detect.od.pytorch.ensemble.BaseTransformTorch = None)

Name	Type	Default	Description
normalizer	Optional[alibi_detect.od.pytorch.ensemble.BaseTransformTorch]	None	BaseFittedTransformTorch object to normalize the scores. If None then no normalization is applied.
aggregator	Optional[alibi_detect.od.pytorch.ensemble.BaseTransformTorch]	None	BaseTransformTorch object to aggregate the scores. If None defaults to AverageAggregator.

Methods

fit

fit(x: torch.Tensor) -> typing_extensions.Self

Fit the normalizer to the scores.

Name	Type	Default	Description
x	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: typing_extensions.Self

transform

transform(x: torch.Tensor) -> torch.Tensor

Apply the normalizer and aggregator to the scores.

Name	Type	Default	Description
x	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor

FitMixinTorch

Inherits from: ABC

Methods

check_fitted

check_fitted()

Checks to make sure object has been fitted.

fit

fit(x_ref: torch.Tensor) -> typing_extensions.Self

Abstract fit method.

Name	Type	Default	Description
x_ref	torch.Tensor
x			torch.Tensor to fit object on.

Returns

• Type: typing_extensions.Self

MaxAggregator

Inherits from: BaseTransformTorch, Module

Constructor

MaxAggregator(self)

Methods

transform

transform(scores: torch.Tensor) -> torch.Tensor

Takes the maximum score of a set of detectors in an ensemble.

Name	Type	Default	Description
scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor

MinAggregator

Inherits from: BaseTransformTorch, Module

Constructor

MinAggregator(self)

Methods

transform

transform(scores: torch.Tensor) -> torch.Tensor

Takes the minimum score of a set of detectors in an ensemble.

Name	Type	Default	Description
scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor

PValNormalizer

Inherits from: BaseTransformTorch, Module, FitMixinTorch, ABC

Constructor

PValNormalizer(self)

Methods

fit

fit(val_scores: torch.Tensor) -> typing_extensions.Self

Fit transform on scores.

Name	Type	Default	Description
val_scores	torch.Tensor		score outputs of ensemble of detectors applied to reference data.

Returns

• Type: typing_extensions.Self

transform

transform(scores: torch.Tensor) -> torch.Tensor

Transform scores to 1 - p-values.

Name	Type	Default	Description
scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor

ShiftAndScaleNormalizer

Inherits from: BaseTransformTorch, Module, FitMixinTorch, ABC

Constructor

ShiftAndScaleNormalizer(self)

Methods

fit

fit(val_scores: torch.Tensor) -> typing_extensions.Self

Computes the mean and standard deviation of the scores and stores them.

Name	Type	Default	Description
val_scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: typing_extensions.Self

transform

transform(scores: torch.Tensor) -> torch.Tensor

Transform scores to normalized values. Subtracts the mean and scales by the standard deviation.

Name	Type	Default	Description
scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor

TopKAggregator

Inherits from: BaseTransformTorch, Module

Constructor

TopKAggregator(self, k: Optional[int] = None)

Name	Type	Default	Description
k	Optional[int]	None	number of scores to take the mean of. If k is left None then will be set to half the number of scores passed in the forward call.

Methods

transform

transform(scores: torch.Tensor) -> torch.Tensor

Takes the mean of the top k scores.

Name	Type	Default	Description
scores	torch.Tensor		Torch.Tensor of scores from ensemble of detectors.

Returns

• Type: torch.Tensor