Temporal Functional Circuits: From Spline Plots to Faithful Explanations in KAN Forecasting

arXiv:2605.05685v1 Announce Type: cross Abstract: Unlike MLPs, Kolmogorov-Arnold Networks (KANs) expose explicit learnable edge functions on every connection, enabling mechanistic explanation in time-series forecasting. This paper introduces Temporal Functional Circuits, a framework that transforms...