Keywords: coreset, streaming data, non-iid, gradient matching, continual learning, GDumb
TL;DR: We devise a coreset method, based on the idea of gradient matching, that is applicable to non-iid streaming data and can be used to curate a rehearsal memory for continual learning.
Abstract: We devise a coreset selection method based on the idea of gradient matching: the gradients induced by the coreset should match, as closely as possible, those induced by the original training dataset. We evaluate the method in the context of continual learning, where it can be used to curate a rehearsal memory. Our method performs strong competitors such as reservoir sampling across a range of memory sizes.