Rethinking the Representational Continuity: Towards Unsupervised Continual Learning

18 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

SLCA: Slow Learner with Classifier Alignment for Continual Learning on a Pre-trained Model (2023, background): “…typically unlabeled and may also arrive incrementally, it seems more reasonable to use self-supervised pre-training than supervised pre-training [5, 42], also regarding that (upstream) continual learning in a self-supervised manner is generally more robust to catastrophic forgetting [17, 26, 9].”
CoSCL: Cooperation of Small Continual Learners is Stronger than a Big One (2022, positive): “An empirical explanation is that UCL achieves a flatter loss landscape and more meaningful feature representations [25], which is consistent with our analysis.”
Self-Supervised Models are Continual Learners (2021, methodology): “We also compare with two concurrent works that propose approaches for CSSL (LUMP [22], Lin et al.”
Continually Learning Self-Supervised Representations with Projected Functional Regularization (2021, background): “In a very recent paper, the authors used contrastive self-supervised learning with a memory buffer for storing exemplars [40].”