Newest papers

Learning to Refer: Inference-Time Collaboration in Decentralized Federated Learning

Yuheng Ma, Chendi Wang, Qiong Zhang · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26800 · cited by 0 works (OpenAlex) · open-access PDF

German

In dezentralisiertem federiertem Lernen können heterogene Kunden komplementäre Kompetenzen entwickeln, aber jede Abfrage wird normalerweise nur von dem Kunden ausgewertet, an dem sie ankommt. Um dieses verteilte Know-how ohne globale Koordination zu nutzen, formulieren wir Inferenzzeit-Kollaboration als budgetierte lokale Experten-Referenzierung, ein grafbegrenztes Routing-Prozess, bei dem eine Abfrage nur zwischen benachbarten Kunden bewegen kann und nur besuchte Kunden zu ihrer endgültigen Prognose beitragen. EdgeDRM implementiert diesen Rahmen, indem er ein dichtheitsbasiertes Routing-Potenzial für jeden Kunden lernt: Unterschiede zwischen benachbarten Potenzialen bieten die Vergleichungen, die verwendet werden, um aufeinanderfolgende Referenzen aus

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

Abstract

In decentralized federated learning, heterogeneous clients may develop complementary expertise, yet each query is typically evaluated only by the client where it arrives. To exploit this distributed expertise without global coordination, we formulate inference-time collaboration as budgeted local expert referral, a graph-constrained routing process in which a query can move only between neighboring clients and only visited clients contribute to its final prediction. EdgeDRM implements this framework by learning a density-based routing potential for each client: differences between neighboring potentials provide the comparisons used to select successive referrals, while local task models generate predictions that are aggregated along the resulting route. Both training and inference are decentralized, using only neighbor-local information without a central router or graph-wide client scoring. We establish estimation guarantees for the learned comparisons and characterize how local improvement and graph structure govern prior-weighted density gains from referral. Experiments on benchmark and real-world medical data show that EdgeDRM outperforms graph-matched decentralized baselines in most settings, with the clearest benefits under sparse connectivity.

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

Topics: Computer science, Exploit, Inference, Task (project management), Process (computing), Benchmark (surveying)

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