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 2601.19341cs.LGcs.AI
收藏

在分布偏移下通过差分重建实现鲁棒不确定性估计

Robust Uncertainty Estimation under Distribution Shift via Difference Reconstruction

1 Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore 2 College of Computing;Data Science, Nanyang Technological University, Singapore 3 Centre for Medical Technologies \& Innovations, National Health Group, Singapore

Xinran Xu, Li Rong Wang, Xiuyi Fan

AI总结:

本文提出DRUE方法,通过差分重建提升模型在分布偏移下的不确定性估计鲁棒性。

摘要:

深度学习模型中的不确定性估计对于医疗影像等高风险应用中的可靠决策至关重要。先前研究已建立输入样本与其由辅助模型生成的重建版本之间的差异可以作为不确定性有用的代理。然而,直接将重建与原始输入进行比较会因信息损失和对表面细节的敏感性而受到限制,这限制了其有效性。在本文中,我们提出了差分重建不确定性估计(DRUE),一种通过从两个中间层重建输入并测量其输出之间的差异作为不确定性分数来减轻这一限制的方法。为了评估实际中的不确定性估计,我们遵循广泛使用的分布外(OOD)检测范式,其中分布内(ID)训练数据与具有增加领域偏移的数据集进行比较。使用青光眼检测作为ID任务,我们证明DRUE在多个OOD数据集上 consistently 实现了优越的AUC和AUPR,突显了其在分布偏移下的鲁棒性和可靠性。这项工作提供了一个原则性和有效的框架,用于增强不确定环境中的模型可靠性。

英文摘要:

Estimating uncertainty in deep learning models is critical for reliable decision-making in high-stakes applications such as medical imaging. Prior research has established that the difference between an input sample and its reconstructed version produced by an auxiliary model can serve as a useful proxy for uncertainty. However, directly comparing reconstructions with the original input is degraded by information loss and sensitivity to superficial details, which limits its effectiveness. In this work, we propose Difference Reconstruction Uncertainty Estimation (DRUE), a method that mitigates this limitation by reconstructing inputs from two intermediate layers and measuring the discrepancy between their outputs as the uncertainty score. To evaluate uncertainty estimation in practice, we follow the widely used out-of-distribution (OOD) detection paradigm, where in-distribution (ID) training data are compared against datasets with increasing domain shift. Using glaucoma detection as the ID task, we demonstrate that DRUE consistently achieves superior AUC and AUPR across multiple OOD datasets, highlighting its robustness and reliability under distribution shift. This work provides a principled and effective framework for enhancing model reliability in uncertain environments.

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。

↑