LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization

<p>Sequence-to-sequence neural networks have recently achieved significant success in abstractive summarization, especially through fine-tuning large pre-trained language models on downstream datasets. However, these models frequently suffer from exposure bias, which can impair their performan...

Full description

Saved in:
Bibliographic Details
Main Author: Eman Aloraini (21797867) (author)
Other Authors: Hozaifa Kassab (21797870) (author), Ali Hamdi (13432680) (author), Khaled Shaban (20074425) (author)
Published: 2025
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513543151288320
author Eman Aloraini (21797867)
author2 Hozaifa Kassab (21797870)
Ali Hamdi (13432680)
Khaled Shaban (20074425)
author2_role author
author
author
author_facet Eman Aloraini (21797867)
Hozaifa Kassab (21797870)
Ali Hamdi (13432680)
Khaled Shaban (20074425)
author_role author
dc.creator.none.fl_str_mv Eman Aloraini (21797867)
Hozaifa Kassab (21797870)
Ali Hamdi (13432680)
Khaled Shaban (20074425)
dc.date.none.fl_str_mv 2025-07-05T09:00:00Z
dc.identifier.none.fl_str_mv 10.1016/j.neucom.2025.130816
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/LexiSem_A_re-ranker_balancing_lexical_and_semantic_quality_for_enhanced_abstractive_summarization/29655764
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Information and computing sciences
Artificial intelligence
Distributed computing and systems software
Machine learning
Abstractive summarization
Re-ranking
Lexical quality
Semantic quality
Deep learning
dc.title.none.fl_str_mv LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p>Sequence-to-sequence neural networks have recently achieved significant success in abstractive summarization, especially through fine-tuning large pre-trained language models on downstream datasets. However, these models frequently suffer from exposure bias, which can impair their performance. To address this, re-ranking systems have been introduced, but their potential remains underexplored despite some demonstrated performance gains. Most prior work relies on ROUGE scores and aligned candidate summaries for ranking, exposing a substantial gap between semantic similarity and lexical overlap metrics. In this study, we demonstrate that a second-stage model can be trained to re-rank a set of summary candidates, significantly enhancing performance. Our novel approach leverages a re-ranker that balance lexical and semantic quality. Additionally, we introduce a new strategy for defining negative samples in ranking models. Through experiments on the CNN/DailyMail, XSum and Reddit TIFU datasets, we show that our method effectively estimates the semantic content of summaries without compromising lexical quality. In particular, our method sets a new performance benchmark on the CNN/DailyMail dataset (48.18 R1, 24.46 R2, 45.05 RL) and on Reddit TIFU (30.37 R1,RL 23.87).</p><h2>Other Information</h2> <p> Published in: Neurocomputing<br> License: <a href="http://creativecommons.org/licenses/by/4.0/" target="_blank">http://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1016/j.neucom.2025.130816" target="_blank">https://dx.doi.org/10.1016/j.neucom.2025.130816</a></p>
eu_rights_str_mv openAccess
id Manara2_0333aa3c465b3c9dad28b58b9cd0b59e
identifier_str_mv 10.1016/j.neucom.2025.130816
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/29655764
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarizationEman Aloraini (21797867)Hozaifa Kassab (21797870)Ali Hamdi (13432680)Khaled Shaban (20074425)Information and computing sciencesArtificial intelligenceDistributed computing and systems softwareMachine learningAbstractive summarizationRe-rankingLexical qualitySemantic qualityDeep learning<p>Sequence-to-sequence neural networks have recently achieved significant success in abstractive summarization, especially through fine-tuning large pre-trained language models on downstream datasets. However, these models frequently suffer from exposure bias, which can impair their performance. To address this, re-ranking systems have been introduced, but their potential remains underexplored despite some demonstrated performance gains. Most prior work relies on ROUGE scores and aligned candidate summaries for ranking, exposing a substantial gap between semantic similarity and lexical overlap metrics. In this study, we demonstrate that a second-stage model can be trained to re-rank a set of summary candidates, significantly enhancing performance. Our novel approach leverages a re-ranker that balance lexical and semantic quality. Additionally, we introduce a new strategy for defining negative samples in ranking models. Through experiments on the CNN/DailyMail, XSum and Reddit TIFU datasets, we show that our method effectively estimates the semantic content of summaries without compromising lexical quality. In particular, our method sets a new performance benchmark on the CNN/DailyMail dataset (48.18 R1, 24.46 R2, 45.05 RL) and on Reddit TIFU (30.37 R1,RL 23.87).</p><h2>Other Information</h2> <p> Published in: Neurocomputing<br> License: <a href="http://creativecommons.org/licenses/by/4.0/" target="_blank">http://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1016/j.neucom.2025.130816" target="_blank">https://dx.doi.org/10.1016/j.neucom.2025.130816</a></p>2025-07-05T09:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1016/j.neucom.2025.130816https://figshare.com/articles/journal_contribution/LexiSem_A_re-ranker_balancing_lexical_and_semantic_quality_for_enhanced_abstractive_summarization/29655764CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/296557642025-07-05T09:00:00Z
spellingShingle LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
Eman Aloraini (21797867)
Information and computing sciences
Artificial intelligence
Distributed computing and systems software
Machine learning
Abstractive summarization
Re-ranking
Lexical quality
Semantic quality
Deep learning
status_str publishedVersion
title LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
title_full LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
title_fullStr LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
title_full_unstemmed LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
title_short LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
title_sort LexiSem: A re-ranker balancing lexical and semantic quality for enhanced abstractive summarization
topic Information and computing sciences
Artificial intelligence
Distributed computing and systems software
Machine learning
Abstractive summarization
Re-ranking
Lexical quality
Semantic quality
Deep learning