Learning Neural Textual Representations for Citation Recommendation

TL;DR

This paper introduces a novel citation recommendation method combining deep document representations with submodular selection, significantly improving performance on benchmark datasets.

Contribution

It is the first to integrate deep neural textual representations with submodular scoring for citation recommendation, enhancing accuracy.

Findings

Outperforms state-of-the-art approaches in MRR and F1-at-k metrics

Uses Sentence-BERT with Siamese and triplet networks for document embedding

Achieves superior results on ACL Anthology Network corpus

Abstract

With the rapid growth of the scientific literature, manually selecting appropriate citations for a paper is becoming increasingly challenging and time-consuming. While several approaches for automated citation recommendation have been proposed in the recent years, effective document representations for citation recommendation are still elusive to a large extent. For this reason, in this paper we propose a novel approach to citation recommendation which leverages a deep sequential representation of the documents (Sentence-BERT) cascaded with Siamese and triplet networks in a submodular scoring function. To the best of our knowledge, this is the first approach to combine deep representations and submodular selection for a task of citation recommendation. Experiments have been carried out using a popular benchmark dataset - the ACL Anthology Network corpus - and evaluated against baselines and a state-of-the-art approach using metrics such as the MRR and F1-at-k score. The results show that the proposed approach has been able to outperform all the compared approaches in every measured metric.