dann

The Domain Adversarial Neural Network (DANN) approach uses a domain discriminator trained on distinguishing the source and target features produced by a shared feature extractor. A Gradient Reversal Layer (GRL) is used to train the feature extractor on making its source and target outputs indistinguishable.

Source --> FeatEx --> Source Feats -----------> Regressor  --> RUL Prediction
        ^         |                 |
        |         |                 v
Target --         --> Target Feats -->  GRL --> DomainDisc --> Domain Prediction

It was originally introduced by Ganin et al. for image classification.

Used In

• da Costa et al. (2020). Remaining useful lifetime prediction via deep domain adaptation. Reliability Engineering & System Safety, 195, 106682. 10.1016/J.RESS.2019.106682
• Krokotsch et al. (2020). A Novel Evaluation Framework for Unsupervised Domain Adaption on Remaining Useful Lifetime Estimation. 2020 IEEE International Conference on Prognostics and Health Management (ICPHM). 10.1109/ICPHM49022.2020.9187058

DannApproach

Bases: AdaptionApproach

The DANN approach introduces a domain discriminator that is trained on distinguishing source and target features as a binary classification problem. The features are produced by a shared feature extractor. The loss in the domain discriminator is binary cross-entropy.

The regressor and domain discriminator need the same number of input units as the feature extractor has output units. The discriminator is not allowed to have an activation function on its last layer and needs to use only a single output neuron because BCEWithLogitsLoss is used.

Examples:

>>> from rul_adapt import model
>>> from rul_adapt import approach
>>> feat_ex = model.CnnExtractor(1, [16, 16, 1], 10, fc_units=16)
>>> reg = model.FullyConnectedHead(16, [1])
>>> disc = model.FullyConnectedHead(16, [8, 1], act_func_on_last_layer=False)
>>> dann = approach.DannApproach(1.0)
>>> dann.set_model(feat_ex, reg, disc)

domain_disc property

The domain discriminator network.

__init__(dann_factor, loss_type='mae', rul_score_mode='phm08', evaluate_degraded_only=False, **optim_kwargs)

Create a new DANN approach.

The strength of the domain discriminator's influence on the feature extractor is controlled by the dann_factor. The higher it is, the stronger the influence.

Possible options for the regression loss are mae, mse and rmse.

The domain discriminator is set by the set_model function together with the feature extractor and regressor. For more information, see the approach module page.

For more information about the possible optimizer keyword arguments, see here.

Parameters:

Name	Type	Description	Default
dann_factor	float	Strength of the domain DANN loss.	required
loss_type	Literal['mae', 'mse', 'rmse']	Type of regression loss.	'mae'
rul_score_mode	Literal['phm08', 'phm12']	The mode for the val and test RUL score, either 'phm08' or 'phm12'.	'phm08'
evaluate_degraded_only	bool	Whether to only evaluate the RUL score on degraded samples.	False
**optim_kwargs	Any	Keyword arguments for the optimizer, e.g. learning rate.	{}

configure_optimizers()

Configure an optimizer for the whole model.

forward(inputs)

Predict the RUL values for a batch of input features.

set_model(feature_extractor, regressor, domain_disc=None, *args, **kwargs)

Set the feature extractor, regressor, and domain discriminator for this approach.

The discriminator is not allowed to have an activation function on its last layer and needs to use only a single output neuron. It is wrapped by a DomainAdversarialLoss.

Parameters:

Name	Type	Description	Default
feature_extractor	Module	The feature extraction network.	required
regressor	Module	The RUL regression network.	required
domain_disc	Optional[Module]	The domain discriminator network.	None

test_step(batch, batch_idx, dataloader_idx)

Execute one test step.

The batch argument is a list of two tensors representing features and labels. A RUL prediction is made from the features and the validation RMSE and RUL score are calculated. The metrics recorded for dataloader idx zero are assumed to be from the source domain and for dataloader idx one from the target domain. The metrics are written to the configured logger under the prefix test.

Parameters:

Name	Type	Description	Default
batch	List[Tensor]	A list containing a feature and a label tensor.	required
batch_idx	int	The index of the current batch.	required
dataloader_idx	int	The index of the current dataloader (0: source, 1: target).	required

training_step(batch, batch_idx)

Execute one training step.

The batch argument is a list of three tensors representing the source features, source labels and target features. Both types of features are fed to the feature extractor. Then the regression loss for the source domain and the DANN loss between domains is computed. The regression, DANN and combined loss are logged.

Parameters:

Name	Type	Description	Default
batch	List[Tensor]	A list of a source feature, source label and target feature tensors.	required
batch_idx	int	The index of the current batch.	required

Returns: The combined loss.

validation_step(batch, batch_idx, dataloader_idx)

Execute one validation step.

The batch argument is a list of two tensors representing features and labels. A RUL prediction is made from the features and the validation RMSE and RUL score are calculated. The metrics recorded for dataloader idx zero are assumed to be from the source domain and for dataloader idx one from the target domain. The metrics are written to the configured logger under the prefix val.

Parameters:

Name	Type	Description	Default
batch	List[Tensor]	A list containing a feature and a label tensor.	required
batch_idx	int	The index of the current batch.	required
dataloader_idx	int	The index of the current dataloader (0: source, 1: target).	required