Preprint
Aug 2026
Geometry-Constrained Kolmogorov-Arnold Networks: Learning Edge Geometry via Banach Duality
Geometry-constrained KANs are introduced, a family of edge activations derived from Banach duality maps in which the geometry itself is learned through a scalar exponent $p>1$ per edge, which controls the qualitative response.
K. S. Sesh Kumar
· 0 citations