Transfer learning is beneficial by allowing the expressive features of models pretrained on large-scale datasets to be finetuned for the target task of smaller, more domain-specific datasets. However, there is a concern that these pretrained models may come with their own biases which would propagate into the finetuned model. In this work, we investigate bias when conceptualized as both spurious correlations between the target task and a sensitive attribute as well as underrepresentation of a particular group in the dataset. Under both notions of bias, we find that (1) models finetuned on top of pretrained models can indeed inherit their biases, but (2) this bias can be corrected for through relatively minor interventions to the finetuning dataset, and often with a negligible impact to performance. Our findings imply that careful curation of the finetuning dataset is important for reducing biases on a downstream task, and doing so can even compensate for bias in the pretrained model.

翻译：迁移学习具有优势，它允许将在大规模数据集上预训练的模型所具有的表达能力较强的特征，微调到更小、更具领域特定性的数据集的目标任务中。然而，存在一种担忧，即这些预训练模型可能带有自身的偏见，这些偏见会传播到微调后的模型中。在本工作中，我们将偏见概念化为目标任务与敏感属性之间的虚假关联，以及特定群体在数据集中的代表性不足。在这两种偏见概念下，我们发现：(1) 在预训练模型基础上进行微调的模型确实会继承其偏见，但 (2) 这种偏见可以通过对微调数据集进行相对较小的干预来纠正，并且通常对性能的影响可以忽略不计。我们的发现表明，精心筛选微调数据集对于减少下游任务的偏见非常重要，这样做甚至能够补偿预训练模型中的偏见。