Kolmogorov-Arnold networks match multilayer perceptrons on clinical data
Kolmogorov-Arnold Networks vs. Multi-Layer Perceptrons: A Rigorous Statistical Benchmark on Clinical Tabular Data
Machine LearningArtificial Intelligence
Summary
Clinical datasets often have many more healthy cases than rare positive ones, making detection of rare cases difficult. The authors compared a newer type of neural network called Kolmogorov-Arnold Networks (KANs) with standard neural networks called multi-layer perceptrons (MLPs) on several medical datasets. They found that some KAN variants detect rare positive cases better but take longer to run and require more parameters. Overall, standard MLPs remain a solid choice for general use due to their efficiency. KANs may be better when catching rare, important cases is the main goal.
What this means in practice
- •For clinical data teams: Improve detection of rare positive cases in clinical tabular data using specific Kolmogorov-Arnold Network variants.
- •For data scientists in finance: Apply Kolmogorov-Arnold Networks to detect rare events in imbalanced structured datasets such as fraud detection or risk modeling.
Authors
Rafał Wodziński, Paweł Szczerbiak
Abstract
Kolmogorov-Arnold Networks (KANs) have recently been proposed as a promising alternative to Multi-Layer Perceptrons (MLPs). However, empirical evaluations on structured medical datasets remain scarce, and commonly reported metrics like AUROC can yield overly optimistic performance estimates on imbalanced clinical cohorts. In this study, we present a standardized benchmark comparing eleven architectures - nine KAN variants (FastKAN, WavKAN, ReLUKAN, ChebyKAN, TaylorKAN, GramKAN, JacobiKAN, LegendreKAN, and TabKAN), a tuned StandardMLP, and a Random Forest baseline - across seven heterogeneous medical tabular datasets spanning oncology, cardiology, nephrology, neurology, endocrinology, and obstetrics. Our experimental protocol employs a 3x5 Repeated Stratified K-Fold cross-validation scheme with nested Bayesian hyperparameter optimization and evaluates discrimination via Area Under the Precision-Recall Curve (AUPRC), calibration through the Brier Score, and computational efficiency. Formal equivalence testing is conducted using the Bayesian Correlated t-Test with a Region of Practical Equivalence (ROPE) at +/- 0.01. The results indicate that on datasets with severe class imbalance, dedicated KAN architectures (particularly TabKAN and ReLUKAN) can improve minority-class detection and achieve higher AUPRC than standard MLPs. Nevertheless, across pooled datasets, KAN variants demonstrate comparable overall predictive discrimination rather than a systematic advantage, while requiring 1.20x to 1.48x longer inference latency per sample and up to 12.5x more trainable parameters. Consequently, StandardMLP remains a strong baseline on the efficiency Pareto frontier, while KAN variants provide a viable alternative when precision on rare positive cases is the primary clinical objective. The code for reproducing the results is available at https://github.com/rafalwodzinski/kan_mlp_comparison.