Influence functions (IFs) are a powerful tool for detecting anomalous examples in large scale datasets. However, they are unstable when applied to deep networks. In this paper, we provide an explanation for the instability of IFs and develop a solution to this problem. We show that IFs are unreliable when the two data points belong to two different classes. Our solution leverages class information to improve the stability of IFs. Extensive experiments show that our modification significantly improves the performance and stability of IFs while incurring no additional computational cost.

翻译：影响函数（IFs）是大规模数据集中检测异常样本的有力工具。然而，当应用于深度网络时，它们表现出不稳定性。本文解释了影响函数不稳定的原因，并针对该问题提出了一种解决方案。我们证明，当两个数据点属于不同类别时，影响函数不可靠。我们的解决方案利用类别信息来提升影响函数的稳定性。大量实验表明，我们的改进显著提升了影响函数的性能与稳定性，且未引入额外计算成本。