Pairwise Learning to Rank by Neural Networks Revisited: Reconstruction, Theoretical Analysis and Practical Performance

Marius Köppel*, Alexander Segner, Martin Wagener, Lukas Pensel, Andreas Karwath, Stefan Kramer

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

We present a pairwise learning to rank approach based on a neural net, called DirectRanker, that generalizes the RankNet architecture. We show mathematically that our model is reflexive, antisymmetric, and transitive allowing for simplified training and improved performance. Experimental results on the LETOR MSLR-WEB10K, MQ2007 and MQ2008 datasets show that our model outperforms numerous state-of-the-art methods, while being inherently simpler in structure and using a pairwise approach only.

Original languageEnglish
Title of host publicationMachine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2019, Proceedings
EditorsUlf Brefeld, Elisa Fromont, Andreas Hotho, Arno Knobbe, Marloes Maathuis, Céline Robardet
PublisherSpringer Vieweg
Pages237-252
Number of pages16
ISBN (Print)9783030461324
DOIs
Publication statusPublished - 2020
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019 - Wurzburg, Germany
Duration: 16 Sep 201920 Sep 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11908 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019
Country/TerritoryGermany
CityWurzburg
Period16/09/1920/09/19

Bibliographical note

Funding Information:
We would like to thank Dr. Christian Schmitt for his contributions to the work presented in this paper. We also thank Luiz Frederic Wagner for proof(read)ing the mathematical aspects of our model. Parts of this research were conducted using the supercomputer Mogon and/or advisory services offered by Johannes Gutenberg University Mainz (hpc.uni-mainz.de), which is a member of the AHRP (Alliance for High Performance Computing in Rhineland Palatinate, www.ahrp.info) and the Gauss Alliance e.V. The authors gratefully acknowledge the computing time granted on the supercomputer Mogon at Johannes Gutenberg University Mainz (hpc.uni-mainz.de). This research was partially funded by the Carl Zeiss Foundation Project: ?Compe-tence Centre for High-Performance-Computing in the Natural Sciences? at the University of Mainz. Furthermore, Andreas Karwath has been co-funded by the MRC grant MR/S003991/1.

Funding Information:
This research was partially funded by the Carl Zeiss Foundation Project: ‘Competence Centre for High-Performance-Computing in the Natural Sciences’ at the University of Mainz. Furthermore, Andreas Karwath has been co-funded by the MRC grant MR/S003991/1.

Publisher Copyright:
© Springer Nature Switzerland AG 2020.

Keywords

  • Information retrieval
  • Learning to rank
  • Machine learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Pairwise Learning to Rank by Neural Networks Revisited: Reconstruction, Theoretical Analysis and Practical Performance'. Together they form a unique fingerprint.

Cite this