解决细粒度分类中层级标签的梯度冲突问题
FlexiGrad: Adaptive Gradient Modulation for Hierarchical Fine-Grained Classification
- 通过层次感知加权函数调节反向传播梯度
- 在多个数据集上提升多粒度分类准确率
- 无需修改模型结构，即插即用
许多细粒度识别任务包含层级标签，如目、科、种。尽管这种监督信息本应有益，但同时优化所有层级常导致训练不稳定，因粗粒度和细粒度分类器对共享主干网络施加不一致的梯度。这种层级梯度冲突阻碍模型学习一致的由粗到精表示。本文提出 FlexiGrad，一种简单且无参数的方法，在反向传播中调节梯度交互。FlexiGrad 仅移除任务分歧时的有害冲突分量，并在任务部分一致时增强共享方向，通过平滑的层次感知加权函数实现。该方法带来稳定优化，同时保留全局结构与细粒度判别特征。FlexiGrad 可无缝集成至现有架构，无需修改，在 CUB-200-2011、FGVC-Aircraft 与 Stanford Cars 数据集上均提升了多粒度分类性能。代码将发布于 PRIS-CV/FlexiGrad。
原文摘要 · Abstract (English)
Many fine-grained recognition tasks contain hierarchical labels such as order, family and species. Although this supervision should be beneficial, jointly optimising all levels often leads to unstable training because coarse and fine classifiers impose inconsistent gradients on the shared backbone. This hierarchical gradient conflict prevents the model from learning a coherent coarse-to-fine representation. In this paper, we propose FlexiGrad, a simple and parameter-free method that regulates gradient interactions during backpropagation. FlexiGrad removes only the harmful conflicting component when tasks disagree and reinforces the shared direction when they partially agree through a smooth hierarchy-aware weighting function. This produces stable optimisation and preserves both global structure and fine-grained discriminative cues. FlexiGrad integrates into existing architectures without modification while improves multi-granularity accuracy on CUB-200-2011, FGVC-Aircraft and Stanford Cars. The code will be available at PRIS-CV/FlexiGrad.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。