Data Augmented Flatness-aware Gradient Projection for Continual Learning

Enneng Yang, Li Shen, Zhenyi Wang, Shiwei Liu, Guibing Guo, Xingwei Wang; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023, pp. 5630-5639

The goal of continual learning (CL) is to continuously learn new tasks without forgetting previously learned old tasks. To alleviate catastrophic forgetting, gradient projection based CL methods require that the gradient updates of new tasks are orthogonal to the subspace spanned by old tasks. This limits the learning process and leads to poor performance on the new task due to the projection constraint being too strong. In this paper, we first revisit the gradient projection method from the perspective of flatness of loss surface, and find that unflatness of the loss surface leads to catastrophic forgetting of the old tasks when the projection constraint is reduced to improve the performance of new tasks. Based on our findings, we propose a Data Augmented Flatness-aware Gradient Projection (DFGP) method to solve the problem, which consists of three modules: data and weight perturbation, flatness-aware optimization, and gradient projection. Specifically, we first perform a flatness-aware perturbation on the task data and current weights to find the case that makes the task loss worst. Next, flatness-aware optimization optimizes both the loss and the flatness of the loss surface on raw and worst-case perturbed data to obtain a flatness-aware gradient. Finally, gradient projection updates the network with the flatness-aware gradient along directions orthogonal to the subspace of the old tasks. Extensive experiments on four datasets show that our method improves the flatness of loss surface and the performance of new tasks, and achieves state-of-the-art (SOTA) performance in the average accuracy of all tasks.