Can I read Factorized Tensor Networks for MTL/MDL on EtoBox?

Factorized Tensor Networks for MTL/MDL by Rakib Hyder is a document available to read on EtoBox.

What is Factorized Tensor Networks for MTL/MDL about?

This paper introduces Factorized Tensor Networks (FTN) for multi-task and multi-domain learning, which enhances efficiency by adding low-rank tensor factors to a shared backbone network. FTN achieves comparable accuracy to independent networks while significantly reducing the number of additional parameters required. The method is adaptable to various architectures, including convolutional and transformer-based models, and demonstrates flexibility in managing task complexity.

Author

Rakib Hyder

Language

EN