Long-tail learning has garnered widespread attention and achieved significant progress in recent times. However, even with pre-trained prior knowledge, models still exhibit weaker generalization performance on tail classes. The promising Sharpness-Aware Minimization (SAM) can effectively improve the generalization capability of models by seeking out flat minima in the loss landscape, which, however, comes at the cost of doubling the computational time. Since the update rule of SAM necessitates two consecutive (non-parallelizable) forward and backpropagation at each step. To address this issue, we propose a novel method called Random SAM prompt tuning (RSAM-PT) to improve the model generalization, requiring only one-step gradient computation at each step. Specifically, we search for the gradient descent direction within a random neighborhood of the parameters during each gradient update. To amplify the impact of tail-class samples and avoid overfitting, we employ the deferred re-weight scheme to increase the significance of tail-class samples. The classification accuracy of long-tailed data can be significantly improved by the proposed RSAM-PT, particularly for tail classes. RSAM-PT achieves the state-of-the-art performance of 90.3\%, 76.5\%, and 50.1\% on benchmark datasets CIFAR100-LT (IF 100), iNaturalist 2018, and Places-LT, respectively. The source code is temporarily available at https://github.com/Keke921/GNM-PT.

翻译：长尾学习近年来受到广泛关注并取得显著进展。然而，即使借助预训练先验知识，模型在尾部类别上仍表现出较弱的泛化性能。前景广阔的锐度感知最小化（SAM）方法通过寻找损失曲面中的平坦极小值，能有效提升模型泛化能力，但其代价是计算时间翻倍——因为SAM的更新规则需要在每一步执行两次连续（不可并行）的前向传播和反向传播。为解决此问题，我们提出一种名为随机SAM提示调优（RSAM-PT）的新方法以提升模型泛化性能，该方法仅需单步梯度计算。具体而言，我们在每次梯度更新时于参数随机邻域内搜索梯度下降方向。为增强尾部类别样本的影响并避免过拟合，我们采用延迟重加权策略提升尾部类别样本的重要性。所提出的RSAM-PT能显著提升长尾数据的分类准确率，尤其针对尾部类别。在基准数据集CIFAR100-LT（IF 100）、iNaturalist 2018和Places-LT上，RSAM-PT分别实现了90.3%、76.5%和50.1%的最先进性能。源代码暂存于https://github.com/Keke921/GNM-PT。