Evaluating Federated Kolmogorov-Arnold Networks on Non-IID Data

Published: 18 Oct 2024, Last Modified: 18 Nov 2024lxai-neurips-24EveryoneRevisionsBibTeXCC BY 4.0
Track: Full Paper
Abstract: Federated Kolmogorov-Arnold Networks (F-KANs) have already been proposed, but their assessment is at an initial stage. We present a comparison between KANs (using B-splines and Radial Basis Functions as activation functions) and Multi-Layer Perceptrons (MLPs) with a similar number of parameters for 100 rounds of federated learning in the MNIST classification task using non-IID partitions with 100 clients. After 15 trials for each model, we show that the best accuracies achieved by MLPs can be achieved by Spline-KANs in half of the time (in rounds), with just a moderate increase in computing time.
Submission Number: 35
Loading