论文标题
最后一层重新训练足以使稳健性具有虚假相关性
Last Layer Re-Training is Sufficient for Robustness to Spurious Correlations
论文作者
论文摘要
神经网络分类器在很大程度上可以依靠简单的虚假特征（例如背景）来做出预测。但是，即使在这些情况下，我们也表明他们仍然经常学习与最新发现相反的数据属性相关的核心功能。受到这种见解的启发，我们证明了简单的最后一层再培训可以匹配或超过伪造相关基准的最先进方法，但复杂性和计算费用较低。此外，我们表明，在大型成像网训练的大型模型上进行最后一层重新培训也可以显着降低对背景和纹理信息的依赖，从而在单个GPU进行了几分钟的培训后，改善了协变量转移的鲁棒性。
Neural network classifiers can largely rely on simple spurious features, such as backgrounds, to make predictions. However, even in these cases, we show that they still often learn core features associated with the desired attributes of the data, contrary to recent findings. Inspired by this insight, we demonstrate that simple last layer retraining can match or outperform state-of-the-art approaches on spurious correlation benchmarks, but with profoundly lower complexity and computational expenses. Moreover, we show that last layer retraining on large ImageNet-trained models can also significantly reduce reliance on background and texture information, improving robustness to covariate shift, after only minutes of training on a single GPU.
