Kolmogorov Arnold Networks and Multi-Layer Perceptrons: A Paradigm Shift in Neural Modelling

Aradhya Gaonkar 1, Nihal Jain 2 Vignesh Chougule 3, Nikhil Deshpande 4, Sneha Varur 5, Channabasappa Muttal 6 · 2026

Read Kolmogorov Arnold Networks and Multi-Layer Perceptrons: A Paradigm Shift in Neural Modelling by Aradhya Gaonkar 1, Nihal Jain 2 Vignesh Chougule 3, Nikhil Deshpande 4, Sneha Varur 5, Channabasappa Muttal 6 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-10563.txt.

Description

Aradhya Gaonkar and co authors set two net kinds side by side: KANs and MLPs. MLPs are the base net form, with fixed curve steps at each node. KANs swap those steps for spline curves that learn. They rest on a known theorem of Kolmogorov. The team test both on base tasks.

Tasks span curve fits, time based calls, and class calls with many inputs. The paper weighs which form wins on each. KANs bring a new shape to the task. Their grid based curves can fit hard bends with fewer nodes. The work is a guide for when to pick each form.

Catalog record

CREATORAradhya Gaonkar 1, Nihal Jain 2 Vignesh Chougule 3, Nikhil Deshpande 4, Sneha Varur 5, Channabasappa Muttal 6
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-10563

Browse the full library · About the Tower of Babel