归因引导的不可靠神经网络行为模型矫正
Attribution-Guided Model Rectification of Unreliable Neural Network Behaviors
用秩一模型编辑加逐层归因定位，在几乎不重训、仅需一张干净样本的条件下修正视觉模型对非鲁棒特征的依赖（如神经木马、虚假相关），为视觉模型的可信部署提供了一种低成本的后处理修复手段。
原文依据 we leverage rank-one model editing to establish an attribution-guided model rectification framework
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论神经网络模型由于在受损样本的非鲁棒特征上表现出不可靠行为，其性能会随之下降。由于模型本身不透明，为矫正模型以解决该问题，往往需要繁琐的数据清洗与模型重训练，带来巨大的计算与人工开销。在本工作中，我们利用秩一模型编辑，构建了一个归因引导的模型矫正框架，能够有效定位并修正模型的不不可靠行为。我们首先将本文的矫正设定与已有的模型编辑区分开来，给出一种在修正不可靠行为的同时保持模型性能、并降低对大规模清洗样本依赖的形式化表述。我们进一步揭示了模型矫正的一个瓶颈，即不同层之间可编辑性存在异质性。为对准错误行为的主要来源，我们提出一种归因引导的层定位方法，量化逐层可编辑性并识别出对不可靠性负最主要责任的层。大量实验表明，我们的方法在修正神经木马、虚假相关与特征泄漏所表现出的不可靠性方面是有效的。我们的方法仅用一张清洗样本即可达成其编辑目标，展现出显著性能，因而具有实际应用吸引力。
英文原文摘要 1363 字符
The performance of neural network models deteriorates due to their unreliable behavior on non-robust features of corrupted samples. Owing to their opaque nature, rectifying models to address this problem often necessitates arduous data cleaning and model retraining, resulting in huge computational and manual overhead. In this work, we leverage rank-one model editing to establish an attribution-guided model rectification framework that effectively locates and corrects model unreliable behaviors. We first distinguish our rectification setting from existing model editing, yielding a formulation that corrects unreliable behavior while preserving model performance and reducing reliance on large budgets of cleansed samples. We further reveal a bottleneck of model rectifying arising from heterogeneous editability across layers. To target the primary source of misbehavior, we introduce an attribution-guided layer localization method that quantifies layer-wise editability and identifies the layer most responsible for unreliabilities. Extensive experiments demonstrate the effectiveness of our method in correcting unreliabilities observed for neural Trojans, spurious correlations and feature leakage. Our method shows remarkable performance by achieving its editing objective with as few as a single cleansed sample, which makes it appealing for practice.