面向皮肤病诊断公平性的公平多出口框架研究
Toward Fairness Through Fair Multi-Exit Framework for Dermatological Disease Diagnosis
AI总结：
本文提出一种公平导向的多出口框架，通过训练更准确且更公平的内部分类器并允许高置信度实例提前退出，在两个皮肤病数据集上改善了公平性表现。
摘要：
公平性在医学图像识别中日益关键。然而，若不缓解偏见，部署不公平的医学人工智能系统可能损害弱势群体的利益。本文观察到，虽然从神经网络更深层提取的特征通常能提供更高的准确率，但随着从更深层提取特征，公平性状况会恶化。这一现象促使我们扩展多出口框架的概念。与现有主要关注准确率的工作不同，我们的多出口框架以公平为导向；内部分类器被训练得更加准确且更公平，并具有高可扩展性，可应用于大多数现有的公平感知框架。在推理过程中，任何来自内部分类器具有高置信度的实例都被允许提前退出。实验结果表明，所提出的框架在两个皮肤病数据集上能够相较现有最优方法改善公平性状况。
英文摘要：
Fairness has become increasingly pivotal in medical image recognition. However, without mitigating bias, deploying unfair medical AI systems could harm the interests of underprivileged populations. In this paper, we observe that while features extracted from the deeper layers of neural networks generally offer higher accuracy, fairness conditions deteriorate as we extract features from deeper layers. This phenomenon motivates us to extend the concept of multi-exit frameworks. Unlike existing works mainly focusing on accuracy, our multi-exit framework is fairness-oriented; the internal classifiers are trained to be more accurate and fairer, with high extensibility to apply to most existing fairness-aware frameworks. During inference, any instance with high confidence from an internal classifier is allowed to exit early. Experimental results show that the proposed framework can improve the fairness condition over the state-of-the-art in two dermatological disease datasets.
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。
