With an intention to better exploit the task relationships for continual learning, we propose a transferability-based task embedding named H-embedding and present a hypernet under its guidance.

We study the properties that make machine learning models generalize their performance across dimensions.