[论文解读] Achieving Fairness in Dermatological Disease Diagnosis through Automatic Weight Adjusting Federated Learning and Personalization

Gelei Xu, Yawen Wu+2 more|arXiv (Cornell University)|Aug 23, 2022

Cutaneous Melanoma Detection and Management被引用 5

一句话总结

本文提出了一种两阶段公平性感知联邦学习框架，用于皮肤病诊断，通过在全局模型聚合过程中自动基于损失调整权重，并在联邦学习后进行个性化微调，以减少肤色偏差。该方法通过优先考虑表现较差的客户端（例如深色肤色）并为每种肤色选择最优模型，实现了相较于基线模型9.6%的准确率提升和68%的准确率方差降低。

ABSTRACT

Dermatological diseases pose a major threat to the global health, affecting almost one-third of the world's population. Various studies have demonstrated that early diagnosis and intervention are often critical to prognosis and outcome. To this end, the past decade has witnessed the rapid evolvement of deep learning based smartphone apps, which allow users to conveniently and timely identify issues that have emerged around their skins. In order to collect sufficient data needed by deep learning and at the same time protect patient privacy, federated learning is often used, where individual clients aggregate a global model while keeping datasets local. However, existing federated learning frameworks are mostly designed to optimize the overall performance, while common dermatological datasets are heavily imbalanced. When applying federated learning to such datasets, significant disparities in diagnosis accuracy may occur. To address such a fairness issue, this paper proposes a fairness-aware federated learning framework for dermatological disease diagnosis. The framework is divided into two stages: In the first in-FL stage, clients with different skin types are trained in a federated learning process to construct a global model for all skin types. An automatic weight aggregator is used in this process to assign higher weights to the client with higher loss, and the intensity of the aggregator is determined by the level of difference between losses. In the latter post-FL stage, each client fine-tune its personalized model based on the global model in the in-FL stage. To achieve better fairness, models from different epochs are selected for each client to keep the accuracy difference of different skin types within 0.05. Experiments indicate that our proposed framework effectively improves both fairness and accuracy compared with the state-of-the-art.

研究动机与目标

• 解决皮肤病人工智能模型中显著的肤色偏差问题，即深色肤色在训练数据中代表性不足。
• 通过缓解不同肤色间的表现差异，提升联邦学习在皮肤病诊断中的公平性。
• 开发一种动态、自动的权重调整机制，在全局模型聚合过程中优先考虑损失较高的客户端。
• 在联邦学习后实现个性化模型适配，以在多样化肤色间平衡公平性与准确率。
• 在保证不同肤色间准确率差距≤5%的前提下，最大化整体性能。

提出的方法

• 在联邦学习阶段，自动权重聚合器为局部损失较高的客户端分配更高的全局模型更新权重，其强度根据损失差异动态调整。
• 权重调整函数采用非线性指数映射，以放大高损失客户端的影响，其中上界M和缩放因子m通过调优以获得最佳性能。
• 在联邦学习后阶段，每个客户端使用不同训练阶段的检查点对全局模型进行个性化微调。
• 模型选择受公平性约束：通过选择每种肤色准确率在0.7至0.75之间的模型，确保不同肤色间的准确率差距控制在5%以内。
• 该框架使用Fitzpatrick 17k数据集，其中皮肤类型1–6代表不同色素水平，通过准确率差距和方差指标评估公平性。
• 超参数如M和m通过消融实验调优，确定M=3为在准确率与稳定性之间取得最佳平衡的最优值。

实验结果

研究问题

• RQ1 在联邦学习中，基于损失的自动权重调整能否减少皮肤病诊断中不同肤色间的性能差异？
• RQ2 与固定或静态权重方案相比，动态权重调整在数据不平衡条件下能否更有效地提升公平性？
• RQ3 在全局模型训练后，个性化微调在多大程度上能进一步提升代表性不足肤色类型的公平性与准确率？
• RQ4 在从不同训练阶段选择模型时，公平性（准确率差距≤5%）与整体准确率之间的最优权衡是什么？
• RQ5 所提出的框架在真实世界皮肤病数据集上是否优于当前最先进的联邦学习方法，在公平性与准确率方面表现更优？

主要发现

• 所提出的FedAuto方法相较于基线FedAvg模型，整体准确率绝对提升了9.6%。
• 不同肤色间的准确率方差降低至原始模型的32%，表明公平性显著改善。
• 肤色类型6（最深肤色）的准确率提升最为显著，经联邦学习后微调后达到0.725，缩小了与高性能肤色类型之间的差距。
• 权重调整函数的最优上界M=3，使F1得分达到最高（0.7003），准确率方差最低（0.00015）。
• 动态权重调整优于固定超参数设置，证明了基于损失差异自动适应强度的优越性。
• 从不同训练阶段选择模型实现了公平性与准确率的权衡，所选模型在所有肤色类型中准确率差异不超过0.05。