fusilli.fusionmodels.tabularfusion

Collection of fusion models for tabular-tabular fusion.

Modules

fusilli.fusionmodels.tabularfusion.activation	Activation-function fusion model for tabular data.
fusilli.fusionmodels.tabularfusion.attention_and_activation	Using activation functions to fuse tabular data, with self-attention on the second tabular modality.
fusilli.fusionmodels.tabularfusion.attention_weighted_GNN	Attention weighted GNN model: the edge weights are the attention weights from a pre-trained MLP and the node features are the second modality.
fusilli.fusionmodels.tabularfusion.channelwise_att	Channel-wise multiplication fusion model for tabular data.
fusilli.fusionmodels.tabularfusion.concat_data	Concatenating the two tabular modalities at the data-level (early fusion)
fusilli.fusionmodels.tabularfusion.concat_feature_maps	Concatenating the feature maps of the two tabular modalities.
fusilli.fusionmodels.tabularfusion.crossmodal_att	Crossmodal multi-head attention for tabular data.
fusilli.fusionmodels.tabularfusion.decision	Decision fusion of two types of tabular data.
fusilli.fusionmodels.tabularfusion.edge_corr_gnn	Edge correlation GNN model: edges are weighted by the correlation between the nodes' first tabular modality features.
fusilli.fusionmodels.tabularfusion.mcvae_model	This module implements the MCVAE (multi-channel variational autoencoder) model for fusing two types of tabular data.