articlePhysical Review XSep 30, 2025GOLD OA

Kolmogorov-Arnold Networks Meet Science

ZLZiming LiuMTMax TegmarkPMPingchuan MaWMWojciech MatusikYWYixuan Wang

The NSF AI Institute for Artificial Intelligence and Fundamental Interactions · Massachusetts Institute of Technology · +1 more institution

Indexed incrossrefdoaj

Abstract

A major challenge of AI plus science lies in its inherent incompatibility: Today’s AI is primarily based on connectionism, while science depends on symbolism. To bridge the two worlds, we propose a framework to seamlessly synergize Kolmogorov-Arnold networks (KANs) and science. The framework highlights KANs’ usage for three aspects of scientific discovery: identifying relevant features, revealing modular structures, and discovering symbolic formulas. The synergy is bidirectional: science to KAN (incorporating scientific knowledge into KANs), and KAN to science (extracting scientific insights from KANs). We highlight major new functionalities in : (1) MultKAN, KANs with multiplication nodes, (2) kanpiler, a KAN…

Citation impact

119
total citations
FWCI
47.44
Percentile
100%
References
26
Citations per year

Authors

5
  • ZL
    Ziming LiuCorresponding

    The NSF AI Institute for Artificial Intelligence and Fundamental Interactions

  • MT
    Max Tegmark

    The NSF AI Institute for Artificial Intelligence and Fundamental Interactions

  • PM
    Pingchuan Ma

    Massachusetts Institute of Technology

  • WM
    Wojciech Matusik

    Massachusetts Institute of Technology

  • YW
    Yixuan Wang

    California Institute of Technology

Topics & keywords

Topics
Keywords
  • Bridge (graph theory)
  • Modular design
  • Compiler
  • Tree (set theory)
  • Scientific discovery
No related works found for this paper.

Funding