Search papers, labs, and topics across Lattice.
This paper introduces FedTopo, a novel relation-level framework for federated learning that addresses the challenges posed by heterogeneous local architectures. By encoding global knowledge as class relation topology instead of absolute feature space, FedTopo allows clients to share and aggregate knowledge more reliably, even when local representations are misaligned. Experimental results demonstrate that FedTopo significantly outperforms traditional sharing methods across multiple datasets and architectures, achieving better performance with reduced communication overhead.
FedTopo transforms federated learning by enabling reliable knowledge sharing through class relation topology, outperforming traditional methods in heterogeneous settings.
Federated learning (FL) enables collaborative learning over decentralized data silos without centralizing raw data. However, heterogeneous local architectures often induce non-aligned representation spaces, making it difficult to transfer global knowledge across silos. Existing paradigms share this knowledge as model parameters, distilled predictions, or class prototypes, yet all encode it in an absolute space that must be aligned across clients. Heterogeneous backbones break this alignment, so the shared knowledge becomes unreliable and misleads local training. We propose FedTopo, a relation-level framework that encodes global knowledge as class relation topology, capturing how classes relate within each client rather than where they lie in feature space. Each client builds its relation topology from local prototypes and uploads it with class statistics. The server then aggregates these relations in a reliability-aware manner that down-weights weakly supported ones, and broadcasts the global topology to clients. The global topology guides local training by emphasizing topology-similar negative classes. Experiments on three datasets under eight heterogeneous backbones show that FedTopo consistently outperforms parameter-, distillation-, and prototype-sharing baselines, with low communication and no inference overhead. Our code is available at https://github.com/Zhaoyang-Ma/FedTopo.
