AI RESEARCH

Kolmogorov-Arnold Fourier Networks

arXiv CS.AI

ArXi:2502.06018v3 Announce Type: replace-cross Although Kolmogoro-Arnold-based interpretable networks (KANs) possess strong theoretical expressiveness, they suffer from severe parameter explosion and limited ability to capture high-frequency features in high-dimensional tasks. To address these issues, we propose the Kolmogoro-Arnold Fourier Network (KAF), which fundamentally redefines the KAN paradigm through spectral reparameterization.