无处不在的虚假特征——ImageNet中有害虚假特征的大规模检测
Spurious Features Everywhere -- Large-Scale Detection of Harmful Spurious Features in ImageNet
AI总结：
针对深度学习分类器因习得虚假特征导致部署后预测失效的问题，本文提出基于神经PCA成分的框架，系统检测ImageNet中的有害虚假特征，构建了Spurious ImageNet数据集，并提出无需重训的SpuFix缓解方法。
摘要：
仅靠深度学习分类器的基准性能并不能可靠预测部署后模型的表现。尤其是当图像分类器在训练数据中习得虚假特征时，其预测可能会以意想不到的方式失效。本文提出了一个框架，可系统识别ImageNet这类大型数据集中的虚假特征，该框架基于我们提出的神经PCA成分及其可视化结果。以往的虚假特征研究通常在小规模实验场景中开展，或需要成本高昂的逐像素标注。与之不同，我们在ImageNet上开展研究，并通过验证某一类别的有害虚假特征本身的存在就足以触发该类别的预测，证实了我们的结果。我们推出了全新的“Spurious ImageNet”数据集，可用于衡量任意ImageNet分类器对有害虚假特征的依赖程度。此外，我们提出了SpuFix这一简单的缓解方法，无需额外标注或重新训练模型，就能降低任意ImageNet分类器对已识别有害虚假特征的依赖。相关代码和数据可在https://github.com/YanNeu/spurious_imagenet 获取。
英文摘要：
Benchmark performance of deep learning classifiers alone is not a reliable predictor for the performance of a deployed model. In particular, if the image classifier has picked up spurious features in the training data, its predictions can fail in unexpected ways. In this paper, we develop a framework that allows us to systematically identify spurious features in large datasets like ImageNet. It is based on our neural PCA components and their visualization. Previous work on spurious features often operates in toy settings or requires costly pixel-wise annotations. In contrast, we work with ImageNet and validate our results by showing that presence of the harmful spurious feature of a class alone is sufficient to trigger the prediction of that class. We introduce the novel dataset "Spurious ImageNet" which allows to measure the reliance of any ImageNet classifier on harmful spurious features. Moreover, we introduce SpuFix as a simple mitigation method to reduce the dependence of any ImageNet classifier on previously identified harmful spurious features without requiring additional labels or retraining of the model. We provide code and data at https://github.com/YanNeu/spurious_imagenet .
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。
