KIT | KIT-Bibliothek | Impressum | Datenschutz

Probabilistic Predictions with Federated L3earning

Thorgeirsson, Adam Thor; Gauterin, Frank

Probabilistic predictions with machine learning are important in many applications. These are commonly done with Bayesian learning algorithms. However, Bayesian learning methods are computationally expensive in comparison with non-Bayesian methods. Furthermore, the data used to train these algorithms are often distributed over a large group of end devices. Federated learning can be applied in this setting in a communication-efficient and privacy-preserving manner but does not include predictive uncertainty. To represent predictive uncertainty in federated learning, our suggestion is to introduce uncertainty in the aggregation step of the algorithm by treating the set of local weights as a posterior distribution for the weights of the global model. We compare our approach to state-of-the-art Bayesian and non-Bayesian probabilistic learning algorithms. By applying proper scoring rules to evaluate the predictive distributions, we show that our approach can achieve similar performance as the benchmark would achieve in a non-distributed setting.

Open Access Logo

Verlagsausgabe §
DOI: 10.5445/IR/1000129115
Veröffentlicht am 05.02.2021
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Fahrzeugsystemtechnik (FAST)
Publikationstyp Zeitschriftenaufsatz
Publikationsdatum 30.12.2020
Sprache Englisch
Identifikator ISSN: 1099-4300
KITopen-ID: 1000129115
Erschienen in Entropy
Verlag MDPI
Band 23
Heft 1
Seiten Art.-Nr. 41
Schlagwörter Bayesian deep learning; Federated learning; Predictive uncertainty; Probabilistic machine learning
Nachgewiesen in Dimensions
Web of Science
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page