Pairwise learning to rank by neural networks revisited: reconstruction, theoretical analysis and practical performance
Machine Learning · 2025-04 · 114 (4)
DOI: 10.1007/s10994-024-06644-6 ↗Kurz gefasst
Fachartikel in der Zeitschrift Machine Learning, online erschienen am 04.03.2025. Verfasst von Marius Köppel und 5 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Carl-Zeiss-Stiftung, Johannes Gutenberg-Universität Mainz.
Abstract (Originalsprache)
We reevaluate the pairwise learning to rank approach based on neural nets, called RankNet, and present a theoretical analysis of its architecture. We show mathematically that the model can, under certain conditions, learn reflexive, antisymmetric, and transitive relations, enabling simplified training and improved performance. Experimental results on the LETOR MSLR-WEB10K, MQ2007 and MQ2008 datasets show that the model outperforms numerous state-of-the-art methods (including a listwise approach), while being inherently simpler in structure and using a pairwise approach only.
Förderung
- Carl-Zeiss-Stiftung
- Johannes Gutenberg-Universität Mainz
Zitieren
APA
Köppel, M., Segner, A., Wagener, M., Pensel, L., Karwath, A., & Kramer, S. (2025). Pairwise learning to rank by neural networks revisited: reconstruction, theoretical analysis and practical performance. Machine Learning, 114(4). https://doi.org/10.1007/s10994-024-06644-6
Vancouver
Köppel M, Segner A, Wagener M, Pensel L, Karwath A, Kramer S. Pairwise learning to rank by neural networks revisited: reconstruction, theoretical analysis and practical performance. Machine Learning. 2025;114(4). doi:10.1007/s10994-024-06644-6
BibTeX
@article{Kppel2025,
author = {Köppel, Marius and Segner, Alexander and Wagener, Martin and Pensel, Lukas and Karwath, Andreas and Kramer, Stefan},
title = {Pairwise learning to rank by neural networks revisited: reconstruction, theoretical analysis and practical performance},
journal = {Machine Learning},
year = {2025},
volume = {114},
number = {4},
doi = {10.1007/s10994-024-06644-6}
}
EUVIMED