Multi-Task Transfer

Hard Parameter Sharing

Multi-Task Learning approach where the lower layers of the network are shared between all tasks, while only the upper layers are specific to each task.

← Terug