On the expressiveness and spectral bias of KANs

Kolmogorov-Arnold Networks (KANs) [2] Theorem Formula (Shallow) Model (Shallow) Model (Deep) Multi-Layer Perceptron (MLP) Kolmogorov-Arnold Network (KAN) Universal Approximation Theorem Kolmogorov-Arnold Representation Theorem f(x) ≈ N(ϵ)