SEP 4, 2026 · PREPRINT
Communication-Efficient Personalized Federated Learning via Layer-Wise Multi-Threshold Random Sketching
arXiv
This is a methodological preprint proposing a novel compression algorithm for federated learning; it presents no empirical validation, clinical outcome, or peer-reviewed publication, and thus does not constitute evidence for clinical or practice-relevant claims.