arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2511.17040cs.LG
收藏

Step-E: 一种用于具有噪声标签的鲁棒学习的可微数据清洗框架

Step-E: A Differentiable Data Cleaning Framework for Robust Learning with Noisy Labels

Wenzhang Du

AI总结:

Step-E通过整合样本选择与模型学习,提升噪声标签下的鲁棒学习性能,显著提高测试准确率并接近干净标签 oracle。

摘要:

在野外收集的训练数据往往包含噪声标签和异常值,这些会显著降低深度神经网络的性能和可靠性。虽然数据清洗通常被用作单独的预处理阶段,但这种两阶段流程既不能完全利用下游模型的反馈,也不能适应未知的噪声模式。我们提出了Step-E,一个简单的框架,将样本选择和模型学习整合到一个优化过程中。在每个epoch中,Step-E通过损失对样本进行排序,并逐渐增加高损失样本在梯度更新中被排除的比例,在短暂的预热阶段后,从而产生一个在线课程,专注于简单且一致的样本,最终忽略持续的异常值。在CIFAR-100N上,Step-E将ResNet-18模型的测试准确率从43.3%(±0.7%)提升到50.4%(±0.9%),明显优于损失截断、自 paced学习和单次过滤,接近干净标签 oracle 的60.5%(±0.2%)。在CIFAR-10N(aggre)上,Step-E也优于噪声基线(85.3% vs. 83.9%),并且几乎达到干净标签 oracle(85.9%),仅需适度的训练时间开销。

英文摘要:

Training data collected in the wild often contain noisy labels and outliers that substantially degrade the performance and reliability of deep neural networks. While data cleaning is commonly applied as a separate preprocessing stage, such two-stage pipelines neither fully exploit feedback from the downstream model nor adapt to unknown noise patterns. We propose Step-E, a simple framework that integrates sample selection and model learning into a single optimization process. At each epoch, Step-E ranks samples by loss and gradually increases the fraction of high-loss examples that are excluded from gradient updates after a brief warm-up stage, yielding an online curriculum that focuses on easy and consistent examples and eventually ignores persistent outliers. On CIFAR-100N, Step-E improves the test accuracy of a ResNet-18 model from 43.3% (+/- 0.7%) to 50.4% (+/- 0.9%), clearly outperforming loss truncation, self-paced learning, and one-shot filtering while approaching the clean-label oracle at 60.5% (+/- 0.2%). On CIFAR-10N (aggre), Step-E also improves over the noisy baseline (85.3% vs. 83.9%) and nearly matches the clean-label oracle (85.9%), with only moderate training-time overhead.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑