Salvato in:
Dettagli Bibliografici
Autori principali: Kovalchuk, Viktor, Son, Denis, Bolatov, Arman, Guizani, Mohsen, Horváth, Samuel, Panov, Maxim, Takáč, Martin, Gorbunov, Eduard, Kotelevskii, Nikita
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2509.15147
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918407924875264
author Kovalchuk, Viktor
Son, Denis
Bolatov, Arman
Guizani, Mohsen
Horváth, Samuel
Panov, Maxim
Takáč, Martin
Gorbunov, Eduard
Kotelevskii, Nikita
author_facet Kovalchuk, Viktor
Son, Denis
Bolatov, Arman
Guizani, Mohsen
Horváth, Samuel
Panov, Maxim
Takáč, Martin
Gorbunov, Eduard
Kotelevskii, Nikita
contents Under data heterogeneity (e.g., $\textit{class mismatch}$), clients may produce unreliable predictions for instances belonging to unfamiliar classes. An equally weighted combination of such predictions can corrupt the teacher signal used for distillation. In this paper, we provide a theoretical analysis of Federated Distillation and show that aggregating client predictions on a shared public dataset converges to a neighborhood of the optimum, where the neighborhood size is governed by the aggregation quality. We further propose two uncertainty-aware aggregation methods, $\mathbf{UWA}$ and $\mathbf{sUWA}$, which leverage density-based uncertainty estimates to down-weight unreliable client predictions. Experiments on image and text classification benchmarks demonstrate that our methods are particularly effective under high data heterogeneity, while matching standard averaging when heterogeneity is low.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15147
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Who to Trust? Aggregating Client Predictions in Federated Distillation
Kovalchuk, Viktor
Son, Denis
Bolatov, Arman
Guizani, Mohsen
Horváth, Samuel
Panov, Maxim
Takáč, Martin
Gorbunov, Eduard
Kotelevskii, Nikita
Machine Learning
Under data heterogeneity (e.g., $\textit{class mismatch}$), clients may produce unreliable predictions for instances belonging to unfamiliar classes. An equally weighted combination of such predictions can corrupt the teacher signal used for distillation. In this paper, we provide a theoretical analysis of Federated Distillation and show that aggregating client predictions on a shared public dataset converges to a neighborhood of the optimum, where the neighborhood size is governed by the aggregation quality. We further propose two uncertainty-aware aggregation methods, $\mathbf{UWA}$ and $\mathbf{sUWA}$, which leverage density-based uncertainty estimates to down-weight unreliable client predictions. Experiments on image and text classification benchmarks demonstrate that our methods are particularly effective under high data heterogeneity, while matching standard averaging when heterogeneity is low.
title Who to Trust? Aggregating Client Predictions in Federated Distillation
topic Machine Learning
url https://arxiv.org/abs/2509.15147