提出新方法提升模型对图像退化的鲁棒性，无需修改网络结构
Suppress and Diversify: Refining Robust Pathways for Corruption Robustness
- 通过分析网络各层特征退化规律，识别出影响鲁棒性的关键路径
- 在8个基准上验证，显著提升多种视觉任务的抗退化能力
- 无需额外参数，可适配不同模型，适合实际部署场景
模型对自然图像退化的鲁棒性对安全关键应用至关重要。现有方法多依赖隐式表征学习，本文首次系统探索显式计算路径，揭示了特征鲁棒性随网络层数递减的现象，并建立了其与模型性能的功能依赖关系。基于此，提出无侵入式精炼方法Suppress and Diversify（S&D），通过动态选择鲁棒路径并利用保持对称性的变换进行多样化，提升模型鲁棒性。S&D具有架构无关性、零参数、测试时无开销等优势。在八个基准上的广泛评估表明，该方法在多种视觉任务、多样主干网络和复杂真实场景中均能持续提升性能，展现出广泛的有效性与可扩展性。
原文摘要 · Abstract (English)
Model robustness against natural image corruptions is essential for safety-critical applications. While existing methods primarily focus on implicit representation learning, we provide the first systematic exploration of computational pathways to explicitly characterize internal robustness. We identify a progressive decay of robust features across network layers and establish a functional dependency between the prevalence of these features and model performance. To exploit these insights, we propose Suppress and Diversify (S\&D), a non-intrusive refinement approach that enhances robustness by dynamically selecting robust pathways and diversifying them through symmetry-preserving transformations. S\&D is architecture-agnostic, parameter-free, and incurs zero test-time overhead. Extensive evaluations across eight benchmarks demonstrate that S\&D consistently improves performance across multiple vision tasks, diverse backbones, and complex real-world scenarios, highlighting its broad efficacy and scalability.
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。