Federated learning (FL) enables collaborative model training without sharing raw patient data, but standard approaches such as FedAvg treat each client as a black box and provide no mechanism for isolating an adversarial contributor, auditing per-client influence, or honoring a departed participant's right to be forgotten. We present Fed-FBD (Federated Functional Block Diversification), a modular federated architecture that decomposes a ResNet backbone into six functional blocks (the stem, four residual groups, and the classification head) and maintains a warehouse of N color variants, each assembled from independently tracked and contributor-stamped blocks. Fed-FBD provides three capabilities absent in FedAvg: (i) architecturally guaranteed block-level isolation, so that an adversarial or mislabelled client cannot contaminate the clean colous; (ii) privacy-by-design, where membership inference advantage is already indistinguishable from chance before any privacy mechanism is applied; and (iii) surgical machine unlearning of a departed participant's contribution at sub-second cost and without retraining. Experiments on six MedMNIST-2D datasets, PathMNIST at 224x224, and CIFAR-10 show that Fed-FBD trades a modest 0.3%-3.1% IID accuracy gap on the adequately sized datasets for these guarantees, remains within 0.8%-4.0% of FedAvg at Dirichlet alpha=1.0 on three of four datasets, and confines all six adversarial attacks we study to the poisoned client's own blocks with at most +/-0.01 AUC drift on the clean colors.

翻译：联邦学习（FL）无需共享原始患者数据即可实现协同模型训练，但FedAvg等标准方法将每个客户端视为黑盒，无法隔离对抗性贡献者、审计单个客户端影响或满足已退出参与者的被遗忘权。我们提出Fed-FBD（联邦功能块多样化），一种模块化联邦架构，将ResNet骨干网络分解为六个功能块（茎模块、四个残差组和分类头），并维护一个包含N种颜色变体的仓库，每种变体由独立追踪且带有贡献者标识的功能块组装而成。Fed-FBD提供了FedAvg所不具备的三种能力：（i）架构级块隔离保障，使对抗性或错误标注的客户端无法污染洁净颜色变体；（ii）隐私内生设计，使得在应用任何隐私机制前，成员推断攻击的优势已与随机猜测无异；（iii）对已退出参与者的贡献实现亚秒级外科式机器遗忘，无需重新训练。在六个MedMNIST-2D数据集、224x224的PathMNIST以及CIFAR-10上的实验表明，Fed-FBD在规模适当的数据集上以0.3%-3.1%的适度IID精度差距换取这些保障，在四个数据集的三个中，Dirichlet alpha=1.0时与FedAvg的差距保持在0.8%-4.0%以内，并且将所有六种对抗攻击限制在被投毒客户端自身的功能块内，洁净颜色变体的AUC波动不超过±0.01。