针对长尾数据中噪声标签分布不均问题，提出自适应修正框架提升模型可靠性。
CARE: Class-Adaptive Expert Consensus for Reliable Learning with Long-Tailed Noisy Labels
- 基于视觉语言模型三源监督，按类别频次动态调整修正严格度
- 在真实与合成数据集上提升3.0%准确率，尤其改善尾部类别性能
- 适合处理标注不均且含噪声的真实场景数据，如医疗图像分类
从现实数据学习常受长尾类别分布和标注噪声的复合挑战制约。现有方法虽部分缓解此问题，但通常忽略标签噪声在不同类别间的非均匀影响，导致尾部类别修正无效、头部类别过度正则化。为此，我们提出类自适应专家共识（CARE），一种参数高效框架，利用视觉语言模型（VLM）提供的三种互补监督信号：观测到的噪声标签、VLM文本嵌入和视觉特征。CARE引入类自适应专家共识机制，根据类别频率动态调节一致性要求——对尾部类别施加更严格的共识约束，对头部类别则放宽标准。通过聚合多源中的高置信度预测，CARE过滤不可靠信号并重新校准类别分布，在长尾分布下实现更可靠的标签修正。在合成与真实世界基准上的大量实验表明，CARE持续优于现有最优方法，最高实现3.0%的性能提升。代码已开源：https://github.com/qwq123-study/CARE。
原文摘要 · Abstract (English)
Learning from real-world data is frequently hindered by the compound challenge of long-tailed class distributions and noisy annotations. Existing methods partially address these issues but typically ignore the non-uniform impact of label noise across classes, resulting in ineffective correction for tail classes and over-regularization for head classes. To address this issue, we propose Class-Adaptive Rectification with Experts (CARE), a parameter-efficient framework that leverages three complementary supervision sources from vision-language models (VLM): observed noisy labels, VLM text embeddings, and visual features. CARE introduces a class-adaptive expert consensus mechanism that enforces stricter agreement for tail classes and more permissive agreement for head classes based on class frequency. By aggregating high-confidence predictions across these sources, CARE filters unreliable signals and recalibrates class distributions, yielding more reliable rectification under long-tailed distributions. Extensive experiments on both synthetic and real-world benchmarks demonstrate that CARE consistently outperforms state-of-the-art methods, achieving up to 3.0\% performance gains. The source code is available at https://github.com/qwq123-study/CARE.
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。