TransformerABMIL

torchmil.models.TransformerABMIL

Bases: MILModel

Transformer Attention-based Multiple Instance Learning model. Proposed in Sm: enhanced localization in Multiple Instance Learning for medical imaging classification.

Given an input bag \(\mathbf{X} = \left[ \mathbf{x}_1, \ldots, \mathbf{x}_N \right]^\top \in \mathbb{R}^{N \times P}\), the model optionally applies a feature extractor, \(\text{FeatExt}(\cdot)\), to transform the instance features: \(\mathbf{X} = \text{FeatExt}(\mathbf{X}) \in \mathbb{R}^{N \times D}\). Then, it transforms the instance features using a transformer encoder,

\[ \mathbf{X} = \text{TransformerEncoder}(\mathbf{X}) \in \mathbb{R}^{N \times D}, \]

and finally it aggregates the instance features into a bag representation \(\mathbf{z} \in \mathbb{R}^{D}\) using the attention-based pooling,

\[ \mathbf{z}, \mathbf{f} = \operatorname{AttentionPool}(\mathbf{X}). \]

where \(\mathbf{f} = \operatorname{MLP}(\mathbf{X}) \in \mathbb{R}^{N}\) are the attention values. The bag representation \(\mathbf{z}\) is then fed into a classifier (one linear layer) to predict the bag label.

See AttentionPool for more details on the attention-based pooling, and TransformerEncoder for more details on the transformer encoder.

__init__(in_shape, pool_att_dim=128, pool_act='tanh', pool_gated=False, feat_ext=torch.nn.Identity(), transf_att_dim=512, transf_n_layers=1, transf_n_heads=8, transf_use_mlp=True, transf_add_self=True, transf_dropout=0.0, criterion=torch.nn.BCEWithLogitsLoss())

Class constructor.

Parameters:

• in_shape (tuple) –

  Shape of input data expected by the feature extractor (excluding batch dimension). If not provided, it will be lazily initialized.

• pool_att_dim (int, default: 128 ) –

  Attention dimension for pooling.

• pool_act (str, default: 'tanh' ) –

  Activation function for pooling. Possible values: 'tanh', 'relu', 'gelu'.

• pool_gated (bool, default: False ) –

  If True, use gated attention in the attention pooling.

• feat_ext (Module, default: Identity() ) –

  Feature extractor.

• transf_att_dim (int, default: 512 ) –

  Attention dimension for transformer encoder.

• transf_n_layers (int, default: 1 ) –

  Number of layers in transformer encoder.

• transf_n_heads (int, default: 8 ) –

  Number of heads in transformer encoder.

• transf_use_mlp (bool, default: True ) –

  Whether to use MLP in transformer encoder.

• transf_add_self (bool, default: True ) –

  Whether to add input to output in transformer encoder.

• transf_dropout (float, default: 0.0 ) –

  Dropout rate in transformer encoder.

• criterion (Module, default: BCEWithLogitsLoss() ) –

  Loss function. By default, Binary Cross-Entropy loss from logits for binary classification.

forward(X, mask=None, return_att=False)

Forward pass.

Parameters:

• X (Tensor) –

  Bag features of shape (batch_size, bag_size, ...).

• mask (Tensor, default: None ) –

  Mask of shape (batch_size, bag_size).

• return_att (bool, default: False ) –

  If True, returns attention values (before normalization) in addition to Y_logits_pred.

Returns:

• Y_pred ( Tensor ) –

  Bag label logits of shape (batch_size,).

• att ( Tensor ) –

  Only returned when return_att=True. Attention values (before normalization) of shape (batch_size, bag_size).

compute_loss(Y, X, mask=None)

Compute loss given true bag labels.

Parameters:

• Y (Tensor) –

  Bag labels of shape (batch_size,).

• X (Tensor) –

  Bag features of shape (batch_size, bag_size, ...).

• mask (Tensor, default: None ) –

  Mask of shape (batch_size, bag_size).

Returns:

• Y_pred ( Tensor ) –

  Bag label logits of shape (batch_size,).

• loss_dict ( dict ) –

  Dictionary containing the loss value.

predict(X, mask=None, return_inst_pred=True)

Predict bag and (optionally) instance labels.

Parameters:

• X (Tensor) –

  Bag features of shape (batch_size, bag_size, ...).

• mask (Tensor, default: None ) –

  Mask of shape (batch_size, bag_size).

• return_inst_pred (bool, default: True ) –

  If True, returns instance labels predictions, in addition to bag label predictions.

Returns:

• Y_pred ( Tensor ) –

  Bag label logits of shape (batch_size,).

• y_inst_pred ( Tensor ) –

  If return_inst_pred=True, returns instance labels predictions of shape (batch_size, bag_size).