Appraisal of Citation Reliability Using a Gan-Based Approach

Dvora Toledano Kitai, Renata Avros, Ilya Lev, Biran Fridman, Zeev Volkovich

2025

Abstract

This paper addresses the pressing issue of citation manipulation in academic publications. Traditional detection methods, which rely on expert manual review, struggle to keep pace with the ever-growing volume of research output. To overcome these limitations, this study introduces an automated, network-based approach for identifying unreliable citations using an Encoder-Decoder model. By learning regular citation patterns, the model detects anomalies through reconstruction errors. Citation reliability is assessed by systematically removing edges from a citation network and predicting their reinstatement using a modified GAN-based framework. Successful predictions indicate legitimate citations, while failures suggest potential manipulation. The proposed methodology is validated on the CORA dataset, demonstrating its effectiveness in distinguishing genuine references from manipulated ones. This approach provides a scalable and data-driven solution for enhancing research integrity and mitigating citation distortions in scholarly literature.

Paper Citation

in Harvard Style

Kitai D., Avros R., Lev I., Fridman B. and Volkovich Z. (2025). Appraisal of Citation Reliability Using a Gan-Based Approach. In Proceedings of the 14th International Conference on Data Science, Technology and Applications - Volume 1: DMBDA; ISBN 978-989-758-758-0, SciTePress, pages 778-785. DOI: 10.5220/0013586500003967

in Bibtex Style

@conference{dmbda25,
author={Dvora Kitai and Renata Avros and Ilya Lev and Biran Fridman and Zeev Volkovich},
title={Appraisal of Citation Reliability Using a Gan-Based Approach},
booktitle={Proceedings of the 14th International Conference on Data Science, Technology and Applications - Volume 1: DMBDA},
year={2025},
pages={778-785},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013586500003967},
isbn={978-989-758-758-0},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Conference on Data Science, Technology and Applications - Volume 1: DMBDA
TI - Appraisal of Citation Reliability Using a Gan-Based Approach
SN - 978-989-758-758-0
AU - Kitai D.
AU - Avros R.
AU - Lev I.
AU - Fridman B.
AU - Volkovich Z.
PY - 2025
SP - 778
EP - 785
DO - 10.5220/0013586500003967
PB - SciTePress