Reproduce “A Benchmarking Study of Kolmogorov-Arnold Networks on Tabular Data” 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
Eleonora Poeta, Flavio Giobergia, Eliana Pastor, Tania Cerquitelli, Elena Baralis
A benchmarking study by researchers at Politecnico di Torino empirically evaluates Kolmogorov-Arnold Networks (KANs) against Multi-Layer Perceptrons (MLPs) on ten real-world tabular datasets, finding that KANs generally achieve superior or comparable classification accuracy and F1 scores, particularly on large datasets, but incur a higher computational cost in terms of training time and FLOPS.
Key insights
- KANs offer a viable alternative to MLPs for tabular data classification, demonstrating competitive or superior performance.
- The performance gains of KANs come with a measurable increase in computational cost, impacting training time and floating-point operations.
- KANs show particular strength on datasets with a large number of instances, suggesting robustness in handling complex data distributions when ample data is available.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.