1 code implementation • 19 Jul 2022 • Rakib Hyder, Ken Shao, Boyu Hou, Panos Markopoulos, Ashley Prater-Bennette, M. Salman Asif
Our method also offers better memory efficiency compared to episodic memory- and mask-based approaches.
no code implementations • 29 Sep 2021 • Rakib Hyder, Ken Shao, Boyu Hou, Panos Markopoulos, Ashley Prater-Bennette, Salman Asif
To update the network for a new task, we learn a low-rank (or rank-1) matrix and add that to the weights of every layer.