Abstract:Most Continual Learning (CL) methods maintain performance on earlier tasks by storing exemplars in a replay buffer, introducing memory overhead that scales with the number of tasks and raising privacy concerns in regulated domains. We propose No Forgetting Learning (NFL), a buffer-free framework for class- and task-incremental learning that instead exploits the inherent redundancy of overparameterized networks.
NFL decomposes the network into a shared backbone and task-specific heads, then applies a stepwise freezing protocol: new capabilities are first isolated, shared representations are adapted under knowledge distillation, and all components are jointly refined with dual soft-target anchoring. NFL+ augments this pipeline with an under-complete auto-encoder that preserves informative features from previous tasks and corrects the prediction bias caused by class imbalance. NFL+LoRA further extends the framework to pre-trained Vision Transformers by confining updates to a low-rank subspace with Fisher-weighted regularization, maintaining constant backbone memory cost regardless of the number of tasks.
On CIFAR-100, Tiny-ImageNet, and ImageNet-1000 across up to 50 incremental tasks, NFL+ outperforms all buffer-free baselines and matches memory-based methods while requiring only 2.53\% of their model size. We also propose a Plasticity--Stability score for more balanced trade-off evaluation.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2503.04638 [cs.LG] |
| (or arXiv:2503.04638v3 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2503.04638 arXiv-issued DOI via DataCite |
From: Mohammad Ali Vahedifar [view email]
[v1]
Thu, 6 Mar 2025 17:25:46 UTC (153 KB)
[v2]
Fri, 7 Mar 2025 09:18:06 UTC (153 KB)
[v3]
Tue, 5 May 2026 21:18:33 UTC (345 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。