EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Kolmogorov–Arnold networks for genomic tasks

Oleksandr Cherednichenko, Maria Poptsova iD

Briefings in Bioinformatics · 04.03.2025 · 26 (2)

DOI: 10.1093/bib/bbaf129 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 31.03.2025. Verfasst von Oleksandr Cherednichenko und 1 weiteren Autor:innen. Beteiligte Einrichtung: International Laboratory of Bioinformatics, HSE University, 11 Pokrovksy Bulvar, Moscow, 109028, Russia. Hauptthemen laut MEDLINE: Genomics, Deep Learning, Neural Networks, Computer. Die Arbeit wurde bisher 17-mal zitiert (Crossref) und nennt 66 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Basic Research Program at HSE University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Kolmogorov–Arnold networks (KANs) emerged as a promising alternative for multilayer perceptrons (MLPs) in dense fully connected networks. Multiple attempts have been made to integrate KANs into various deep learning architectures in the domains of computer vision and natural language processing. Integrating KANs into deep learning models for genomic tasks has not been explored. Here, we tested linear KANs (LKANs) and convolutional KANs (CKANs) as a replacement for MLP in baseline deep learning architectures for classification and generation of genomic sequences. We used three genomic benchmark datasets: Genomic Benchmarks, Genome Understanding Evaluation, and Flipon Benchmark. We demonstrated that LKANs outperformed both baseline and CKANs on almost all datasets. CKANs can achieve comparable results but struggle with scaling over large number of parameters. Ablation analysis demonstrated that the number of KAN layers correlates with the model performance. Overall, linear KANs show promising results in improving the performance of deep learning models with relatively small number of parameters. Unleashing KAN potential in different state-of-the-art deep learning architectures currently used in genomics requires further research.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

GenomicsDeep LearningNeural Networks, ComputerHumansAlgorithmsComputational Biology

Stichwörter der Autor:innen

Kolmogorov–Arnold networksdeep learningdiffusion modelsgenomic benchmarkslarge language modelsregulatory genomics

Autor:innen und Institutionen

  • International Laboratory of Bioinformatics, HSE University, 11 Pokrovksy Bulvar, Moscow, 109028, Russia.

ORCID-Profile: Maria Poptsova

Förderung

  • Basic Research Program at HSE University

Zitieren

APA

Cherednichenko, O., & Poptsova, M. (2025). Kolmogorov–Arnold networks for genomic tasks. Briefings in Bioinformatics, 26(2). https://doi.org/10.1093/bib/bbaf129

Vancouver

Cherednichenko O, Poptsova M. Kolmogorov–Arnold networks for genomic tasks. Briefings in Bioinformatics. 2025;26(2). doi:10.1093/bib/bbaf129

BibTeX
@article{Cherednichenko2025,
  author = {Cherednichenko, Oleksandr and Poptsova, Maria},
  title = {Kolmogorov–Arnold networks for genomic tasks},
  journal = {Briefings in Bioinformatics},
  year = {2025},
  volume = {26},
  number = {2},
  doi = {10.1093/bib/bbaf129}
}
Kostenlos starten →