This work investigates catastrophic forgetting in continual SSL from an explicitly geometric perspective, moving beyond downstream linear probe accuracy to characterize the evolution of the representation space directly and inform the development of novel methods that specifically preserve the embedding space geometry.
This work proposes FlowLess-R, a representation-space regularization method that constrains replay representations relative to stored references while allowing continued learning and introduces representation flux, a geometric measure of sample-level representation displacement across training.
TASSO, a new paradigm that efficiently preserves the latent space geometry while ensuring network plasticity, is introduced with two complementary techniques: subspace learning and geometry-aware knowledge distillation.
Chang-Ming Sun, Francesco Barbato, Matteo Caligiuri et al.· 0 citations
This paper studies what happens to the rest of the model when a class is forgotten, using a label-conditioned energy-based model (EBM) that assigns per-class energies, making the effect directly observable.
Syed Ali Ahmed, Syed Bilal Ahsan, Muhammad Zaigham Zaheer National University of Computer et al.· 0 citations
This work presents a comprehensive survey of CSSL for vision, with connections to emerging vision-language settings and argues that advancing CSSL requires moving beyond small-scale benchmarks towards continual pre-training paradigms for large-scale systems.
Sergi Masip Cabeza, Alicja Dobrzeniecka, Jonathan Swinnen et al.· 0 citations
In non-stationary streaming environments, simultaneously adapting to complex, non-linear domain shifts via continual learning while mitigating the catastrophic effects of severe, uncalibrated label noise poses a fundamental mathematical challenge. In this paper, we propose \FlatManifold{}, a novel, streamlined robust c...
Experiments show that FiUni can effectively infer latent batch-level task affiliations and achieve competitive performance against advanced task-aware CL methods with fewer trainable parameters.
Dezheng Han, Anlan Zhang, Zhiwu Zhu et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
