Newest papers

The Dynamics of Collapse and Recovery in Clustered Federated Learning

Priyanka Sinha, Snigdhansu Chatterjee · 2026 · arXiv

DOI: 10.48550/arXiv.2610.28565 · open-access PDF

Chinese

集群联盟学习(CFL)可能会遭受集群崩溃,特别是在不利的初始化下,早期客户端任务将异质群集集中在一个集群模型上,然后被训练到集群妥协,而另一个模型没有收到更新。由于当前模型决定客户端路由和路由决定后续模型更新,我们将集群联盟学习作为未知动态系统的控制,而集群模型作为进化状态,客户端路由模式作为控制输入,以及路由干预作为状态底部

Machine translation (Meta M2M100); the English original follows.

Abstract

Clustered federated learning (CFL) can suffer from cluster collapse. In particular, under unfavorable initialization, early client assignments concentrate heterogeneous groups on one cluster model, which is then trained toward a pooled compromise, while another model receives no updates. Since the current models determine client routing and routing determines subsequent model updates, we formulate clustered federated learning as control of an unknown dynamical system, with cluster models as the evolving state, client routing patterns as the control input, and routing interventions as state dep

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange