arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2511.04401cs.LGcs.AI
收藏

针对最差群体鲁棒性的虚假相关性感知嵌入正则化

Spurious Correlation-Aware Embedding Regularization for Worst-Group Robustness

Yonsei University(延世大学);LG CNS;Korea Development Bank(韩国发展银行)

Subeen Park, Joowang Kim, Hakyung Lee, Sunjae Yoo, Kyungwoo Song

AI总结:

SCER通过正则化特征表示,减少模型对虚假线索的依赖,提升最差群体的鲁棒性。

摘要:

深度学习模型在各种领域中实现了强大的性能,但往往依赖于虚假相关性,使其容易受到分布偏移的影响。这个问题在亚群体偏移场景中尤为严重,其中模型在代表性不足的群体中表现不佳。尽管现有方法在缓解这一问题上取得了进展,但其性能提升仍然受限。它们缺乏将嵌入空间表示与最差群体误差联系起来的严格理论框架。为了解决这一限制,我们提出了针对最差群体鲁棒性的虚假相关性感知嵌入正则化(SCER),一种新的方法,直接正则化特征表示以抑制虚假线索。我们理论证明,最差群体误差受分类器对虚假方向和核心方向依赖程度的影响,这些差异体现在跨域和类别的群体平均嵌入差异中。通过在嵌入层面施加理论约束,SCER鼓励模型专注于核心特征,同时减少对虚假模式的敏感性。通过系统评估多个视觉和语言数据集,我们证明SCER在最差群体准确性上优于先前的最先进研究。我们的代码可在https://github.com/MLAI-Yonsei/SCER获得。

英文摘要:

Deep learning models achieve strong performance across various domains but often rely on spurious correlations, making them vulnerable to distribution shifts. This issue is particularly severe in subpopulation shift scenarios, where models struggle in underrepresented groups. While existing methods have made progress in mitigating this issue, their performance gains are still constrained. They lack a rigorous theoretical framework connecting the embedding space representations with worst-group error. To address this limitation, we propose Spurious Correlation-Aware Embedding Regularization for Worst-Group Robustness (SCER), a novel approach that directly regularizes feature representations to suppress spurious cues. We show theoretically that worst-group error is influenced by how strongly the classifier relies on spurious versus core directions, identified from differences in group-wise mean embeddings across domains and classes. By imposing theoretical constraints at the embedding level, SCER encourages models to focus on core features while reducing sensitivity to spurious patterns. Through systematic evaluation on multiple vision and language, we show that SCER outperforms prior state-of-the-art studies in worst-group accuracy. Our code is available at \href{https://github.com/MLAI-Yonsei/SCER}{https://github.com/MLAI-Yonsei/SCER}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑