Cross-Domain Attention

This paper introduces Transferable-guided Attention (TransferAttn) model, a novel ViT for unsupervised domain adaptation (UDA). To improve the transferability of ViT, we propose Domain Transferable-guided Attention Block (DTAB), a new attention mechanism that improves adaptation and domain transferability.

Do you want to know more about improving the transferability of a ViT for unsupervised domain adaptation? Check our paper “Transferable-guided Attention Is All You Need for Video Domain Adaptation”, just accepted for publication in the IEEE/CVF Winter Conference on Applications of Computer Vision.