The XAE learning architecture is designed to identify domain-sharable information learned through gated cycle consistent translation between modalities, particularly with respect to segmented 3D multichannel image tensors and single-cell genomic profiles.
