Avoiding spurious correlations via logit correction
This addresses the issue of poor generalization in models due to spurious correlations, which is a common problem in machine learning, though it is an incremental improvement over existing methods.
The paper tackles the problem of machine learning models relying on spurious correlations in training data, which harms inference performance, by proposing a logit correction loss that improves group-balanced accuracy and achieves an average 5.5% absolute improvement over state-of-the-art methods on benchmarks without needing spurious attribute labels.
Empirical studies suggest that machine learning models trained with empirical risk minimization (ERM) often rely on attributes that may be spuriously correlated with the class labels. Such models typically lead to poor performance during inference for data lacking such correlations. In this work, we explicitly consider a situation where potential spurious correlations are present in the majority of training data. In contrast with existing approaches, which use the ERM model outputs to detect the samples without spurious correlations and either heuristically upweight or upsample those samples, we propose the logit correction (LC) loss, a simple yet effective improvement on the softmax cross-entropy loss, to correct the sample logit. We demonstrate that minimizing the LC loss is equivalent to maximizing the group-balanced accuracy, so the proposed LC could mitigate the negative impacts of spurious correlations. Our extensive experimental results further reveal that the proposed LC loss outperforms state-of-the-art solutions on multiple popular benchmarks by a large margin, an average 5.5\% absolute improvement, without access to spurious attribute labels. LC is also competitive with oracle methods that make use of the attribute labels. Code is available at https://github.com/shengliu66/LC.