伪特征排序：通过数据排序来度量与缓解偏差
Spuriosity Rankings: Sorting Data to Measure and Mitigate Biases
AI总结：
本文提出通过按伪特征程度排序图像来度量并缓解模型偏差，无需改动训练数据或流程，并在ImageNet上验证了该方法可高效识别偏差、微调去偏且跨模型偏差高度相关。
摘要：
我们提出了一种简单但有效的方法，用于度量和缓解因依赖伪相关线索而导致的模型偏差。我们的方法无需对数据或模型训练过程进行代价高昂的改动，而是通过排序更好地利用已有的数据。具体而言，我们根据伪特征程度（即常见伪相关线索出现的程度）对每个类别内的图像进行排序，并以一个可解释网络的深度神经特征作为代理。借助伪特征排序，可以很容易地识别少数子群体（即低伪特征图像），并将模型偏差评估为高伪特征图像与低伪特征图像之间的准确率差距。甚至可以通过在低伪特征图像上微调分类头，以很小的准确率代价高效地消除模型偏差，从而使样本无论伪特征程度如何都能得到更公平的处理。我们在ImageNet上演示了该方法，标注了5000个类别-特征依赖关系（其中630个被发现是伪相关的），并在此过程中为这些特征生成了包含325k个软分割的数据集。通过识别出的伪相关神经特征计算伪特征排序后，我们评估了89个不同模型的偏差，发现各类别的偏差在不同模型之间高度相关。我们的结果表明，由伪特征依赖引起的模型偏差受训练数据内容的影响远大于受训练方式的影响。
英文摘要：
We present a simple but effective method to measure and mitigate model biases caused by reliance on spurious cues. Instead of requiring costly changes to one's data or model training, our method better utilizes the data one already has by sorting them. Specifically, we rank images within their classes based on spuriosity (the degree to which common spurious cues are present), proxied via deep neural features of an interpretable network. With spuriosity rankings, it is easy to identify minority subpopulations (i.e. low spuriosity images) and assess model bias as the gap in accuracy between high and low spuriosity images. One can even efficiently remove a model's bias at little cost to accuracy by finetuning its classification head on low spuriosity images, resulting in fairer treatment of samples regardless of spuriosity. We demonstrate our method on ImageNet, annotating $5000$ class-feature dependencies ($630$ of which we find to be spurious) and generating a dataset of $325k$ soft segmentations for these features along the way. Having computed spuriosity rankings via the identified spurious neural features, we assess biases for $89$ diverse models and find that class-wise biases are highly correlated across models. Our results suggest that model bias due to spurious feature reliance is influenced far more by what the model is trained on than how it is trained.
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。
