27 Oct 2026 - 30 Oct 2026: LSDS Lab Presents new research on Clustered Distributed Learning at FLTA 2026 

Our researcher Mohsen Salimi Khanghah presented our latest paper at the IEEE International Conference on Federated Learning Technologies and Applications (FLTA 2026), held in Paris, France.

The paper, titled:”Dynamic Proxy-Based Clustering for Reliable Federated and Split Learning under Heterogeneity”   Authors: Mohsen Salimi Khanghah, Khalid Ali, Phil Aupke, Andreas Kassler, Nabil Abdennadher, and Giovanna Di Marzo Serugendo

The study re-examines clustered Federated Learning and Split Learning as iterative, topology-adaptive processes rather than static pre-processing steps. By introducing privacy-preserving proxies and evaluating them under information-geometric distances, the proposed framework dynamically adapts client grouping and aggregation topologies across training rounds.