Reproduce “KANLab: TensorFlow Playground for Kolmogorov-Arnold Networks” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lukas Niehaus, Robin Rawiel, Ulf Krumnack
KANLab is a browser-based tool for inspecting and editing the learnable functions on Kolmogorov–Arnold Network edges. Users can change spline control points and see effects in the network’s intermediate nodes and decision boundary. In one compact 2D example, pruning a spline is followed by increased lower-path activation variance and decreasing classification error; this does not establish a general learning benefit.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.