White Paper Assistance: A Step Forward Beyond the Shortcut Learning

Xuan Cheng, Tianshu Xie, XiaoMin Wang, MingHui Liu, Jiali Deng, Ming Liu

Keywords: Shortcut Learning, Bias, Classification, Imbalanced Classification, Robustness

Abstract: The promising performances of CNNs often overshadow the need to examine whether they are doing in the way we are actually interested. We show through experiments that even over-parameterized models would still solve a dataset by recklessly leveraging spurious correlations, or so-called ``shortcuts’’. To combat with this unintended propensity, we borrow the idea of printer test page and propose a novel approach called White Paper Assistance. Our proposed method is two-fold; (a) we intentionally involves the white paper to detect the extent to which the model has preference for certain characterized patterns and (b) we debias the model by enforcing it to make a random guess on the white paper. We show the consistent accuracy improvements that are manifest in various architectures, datasets and combinations with other techniques. Experiments have also demonstrated the versatility of our approach on imbalanced classification and robustness to corruptions.

Community Implementations: [ 3 code implementations](https://www.catalyzex.com/paper/arxiv:2106.04178/code)