提出双侧优化框架，提升异构环境下的个性化联邦学习稳定性和效果。
HEART-PFL: Stable Personalized Federated Learning under Heterogeneity with Hierarchical Directional Alignment and Adversarial Knowledge Transfer
- 分阶段对齐客户端特征：浅层用余弦相似度，深层用MSE匹配，保留个体差异。
- 在非独立同分布下，三个数据集准确率分别达63.42%、84.23%、95.67%。
- 轻量适配器仅146万参数，适合资源受限场景，且对异常代理数据鲁棒。
个性化联邦学习（PFL）旨在应对异构数据分布，实现客户端专属模型，但现有方法存在原型对齐浅层化与服务器端蒸馏脆弱的问题。本文提出HEART-PFL，一种双侧优化框架：(i) 通过早期阶段的余弦相似度和深层阶段的MSE匹配实现深度感知的分层方向对齐（HDA），以保留客户端特异性；(ii) 采用对称KL蒸馏，在干净与对抗性代理数据上进行对抗知识迁移（AKT），稳定全局更新。使用仅含146万可训练参数的轻量适配器，HEART-PFL在CIFAR-100、Flowers-102、Caltech-101数据集上，于狄利克雷非独立同分布划分下，分别取得63.42%、84.23%、95.67%的个性化准确率，且对域外代理数据保持鲁棒。消融实验表明，HDA与AKT在对齐、鲁棒性与优化稳定性方面提供互补增益，共同促进有效个性化。整体结果证明，该方法同时提升个性化性能与全局稳定性，具备强扩展性，代码已开源。
原文摘要 · Abstract (English)
Personalized Federated Learning (PFL) aims to deliver effective client-specific models under heterogeneous distributions, yet existing methods suffer from shallow prototype alignment and brittle server-side distillation. We propose HEART-PFL, a dual-sided framework that (i) performs depth-aware Hierarchical Directional Alignment (HDA) using cosine similarity in the early stage and MSE matching in the deep stage to preserve client specificity, and (ii) stabilizes global updates through Adversarial Knowledge Transfer (AKT) with symmetric KL distillation on clean and adversarial proxy data. Using lightweight adapters with only 1.46M trainable parameters, HEART-PFL achieves state-of-the-art personalized accuracy on CIFAR-100, Flowers-102, and Caltech-101 (63.42%, 84.23%, and 95.67%, respectively) under Dirichlet non-IID partitions, and remains robust to out-of-domain proxy data. Ablation studies further confirm that HDA and AKT provide complementary gains in alignment, robustness, and optimization stability, offering insights into how the two components mutually reinforce effective personalization. Overall, these results demonstrate that HEART-PFL simultaneously enhances personalization and global stability, highlighting its potential as a strong and scalable solution for PFL(code available at https://github.com/danny0628/HEART-PFL).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。