Improving Group Robustness on Spurious Correlation via Evidential Alignment

TL;DR

This paper introduces Evidential Alignment, a new method that uses uncertainty quantification to reduce reliance on spurious correlations in neural networks, improving out-of-distribution robustness without needing group annotations.

Contribution

It proposes a novel evidential calibration framework that debiases models by understanding their biases through uncertainty, without external annotations or auxiliary models.

Findings

Significantly improves group robustness across architectures.

Effectively suppresses spurious correlations without group labels.

Applicable to diverse data modalities.

Abstract

Deep neural networks often learn and rely on spurious correlations, i.e., superficial associations between non-causal features and the targets. For instance, an image classifier may identify camels based on the desert backgrounds. While it can yield high overall accuracy during training, it degrades generalization on more diverse scenarios where such correlations do not hold. This problem poses significant challenges for out-of-distribution robustness and trustworthiness. Existing methods typically mitigate this issue by using external group annotations or auxiliary deterministic models to learn unbiased representations. However, such information is costly to obtain, and deterministic models may fail to capture the full spectrum of biases learned by the models. To address these limitations, we propose Evidential Alignment, a novel framework that leverages uncertainty quantification to understand the behavior of the biased models without requiring group annotations. By quantifying the evidence of model prediction with second-order risk minimization and calibrating the biased models with the proposed evidential calibration technique, Evidential Alignment identifies and suppresses spurious correlations while preserving core features. We theoretically justify the effectiveness of our method as capable of learning the patterns of biased models and debiasing the model without requiring any spurious correlation annotations. Empirical results demonstrate that our method significantly improves group robustness across diverse architectures and data modalities, providing a scalable and principled solution to spurious correlations.