Interpretable knowledge distillation via symbolic regression for feedforward neural networks
Neural Computing and Applications · 2026-04 · 38 (7)
DOI: 10.1007/s00521-026-12067-2 ↗Kurz gefasst
Fachartikel in der Zeitschrift Neural Computing and Applications, online erschienen am 02.04.2026. Verfasst von Assaf Shmuel und 3 weiteren Autor:innen. Die Arbeit nennt 54 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Bar-Ilan University.
Abstract (Originalsprache)
Neural networks (NNs) are widely used for modeling complex, high-dimensional relationships but often lack interpretability, limiting their adoption in critical domains such as healthcare, finance, and engineering. Symbolic regression (SR), in contrast, generates explicit mathematical expressions that enhance transparency but typically underperform in predictive accuracy. To bridge this gap, we propose a knowledge distillation framework that approximates the activations of a trained feedforward NN’s final hidden layer using SR models. This approach enhances interpretability while retaining a substantial fraction of the neural network’s predictive performance in structured data settings. Our method is evaluated across 20 diverse datasets, demonstrating a 7-21% improvement in RMSE over baseline SR models. These results highlight the potential of symbolic knowledge distillation as a practical tool for enhancing model transparency in structured data applications.
Förderung
- Bar-Ilan University
Zitieren
APA
Shmuel, A., Koren, N., Glickman, O., & Lazebnik, T. (2026). Interpretable knowledge distillation via symbolic regression for feedforward neural networks. Neural Computing and Applications, 38(7). https://doi.org/10.1007/s00521-026-12067-2
Vancouver
Shmuel A, Koren N, Glickman O, Lazebnik T. Interpretable knowledge distillation via symbolic regression for feedforward neural networks. Neural Computing and Applications. 2026;38(7). doi:10.1007/s00521-026-12067-2
BibTeX
@article{Shmuel2026,
author = {Shmuel, Assaf and Koren, Nir and Glickman, Oren and Lazebnik, Teddy},
title = {Interpretable knowledge distillation via symbolic regression for feedforward neural networks},
journal = {Neural Computing and Applications},
year = {2026},
volume = {38},
number = {7},
doi = {10.1007/s00521-026-12067-2}
}
EUVIMED