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

Indonesian

Dalam pembelajaran federasi data terdesentralisasi, klien heterogen dapat mengembangkan keahlian komplementer, namun setiap kueri biasanya hanya dievaluasi oleh klien di mana ia tiba. Untuk memanfaatkan keahlian terdesentralisasi ini tanpa koordinasi global, kami merumuskan kolaborasi inferensi-waktu sebagai referral ahli lokal yang dibiayai, proses routing terbatas graf di mana kueri hanya dapat bergerak antara klien tetangga dan hanya klien yang dikunjungi yang berkontribusi pada prediksi akhir. EdgeDRM menerapkan kerangka kerja ini dengan mempelajari potensi routing berbasis kepadatan untuk setiap klien: perbedaan antara potensi tetangga menyediakan perbandingan yang digunakan untuk memilih referral berturut-turut, sementara model tugas lokal menghasilkan prediksi yang agregasi di sepanjang rute yang dihasilkan. Kedua pelatihan dan inferensi terdesentralisasi, menggunakan hanya informasi tetangga-lok

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