Beyond Sharpness: A Flatness Decomposition Framework for Efficient Continual Learning
This work addresses computational bottlenecks in continual learning for AI practitioners, offering a more efficient optimization method, though it is incremental in improving existing sharpness-aware approaches.
The paper tackled the problem of computational inefficiency in sharpness-aware methods for continual learning by proposing FLAD, a framework that decomposes sharpness perturbations and retains only the noise component, achieving significant performance gains with reduced overhead.
Continual Learning (CL) aims to enable models to sequentially learn multiple tasks without forgetting previous knowledge. Recent studies have shown that optimizing towards flatter loss minima can improve model generalization. However, existing sharpness-aware methods for CL suffer from two key limitations: (1) they treat sharpness regularization as a unified signal without distinguishing the contributions of its components. and (2) they introduce substantial computational overhead that impedes practical deployment. To address these challenges, we propose FLAD, a novel optimization framework that decomposes sharpness-aware perturbations into gradient-aligned and stochastic-noise components, and show that retaining only the noise component promotes generalization. We further introduce a lightweight scheduling scheme that enables FLAD to maintain significant performance gains even under constrained training time. FLAD can be seamlessly integrated into various CL paradigms and consistently outperforms standard and sharpness-aware optimizers in diverse experimental settings, demonstrating its effectiveness and practicality in CL.