@inproceedings{wang-etal-2025-unsupervised,
title = "Unsupervised Sentence Representation Learning with Syntactically Aligned Negative Samples",
author = "Wang, Zhilan and
Zhi, Zekai and
Jin, Rize and
Song, Kehui and
Wang, He and
Cho, Da-Jung",
editor = "Chiruzzo, Luis and
Ritter, Alan and
Wang, Lu",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2025",
month = apr,
year = "2025",
address = "Albuquerque, New Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-naacl.461/",
pages = "8247--8259",
ISBN = "979-8-89176-195-7",
abstract = "Sentence representation learning benefits from data augmentation strategies to improve model performance and generalization, yet existing approaches often encounter issues such as semantic inconsistencies and feature suppression. To address these limitations, we propose a method for generating Syntactically Aligned Negative (SAN) samples through a semantic importance-aware Masked Language Model (MLM) approach. Our method quantifies semantic contributions of individual words to produce negative samples that have substantial textual overlap with the original sentences while conveying different meanings. We further introduce Hierarchical-InfoNCE (HiNCE), a novel contrastive learning objective employing differential temperature weighting to optimize the utilization of both in-batch and syntactically aligned negative samples. Extensive evaluations across seven semantic textual similarity benchmarks demonstrate consistent improvements over state-of-the-art models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2025-unsupervised">
<titleInfo>
<title>Unsupervised Sentence Representation Learning with Syntactically Aligned Negative Samples</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhilan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zekai</namePart>
<namePart type="family">Zhi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rize</namePart>
<namePart type="family">Jin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kehui</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">He</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Da-Jung</namePart>
<namePart type="family">Cho</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: NAACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Luis</namePart>
<namePart type="family">Chiruzzo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alan</namePart>
<namePart type="family">Ritter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Albuquerque, New Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-195-7</identifier>
</relatedItem>
<abstract>Sentence representation learning benefits from data augmentation strategies to improve model performance and generalization, yet existing approaches often encounter issues such as semantic inconsistencies and feature suppression. To address these limitations, we propose a method for generating Syntactically Aligned Negative (SAN) samples through a semantic importance-aware Masked Language Model (MLM) approach. Our method quantifies semantic contributions of individual words to produce negative samples that have substantial textual overlap with the original sentences while conveying different meanings. We further introduce Hierarchical-InfoNCE (HiNCE), a novel contrastive learning objective employing differential temperature weighting to optimize the utilization of both in-batch and syntactically aligned negative samples. Extensive evaluations across seven semantic textual similarity benchmarks demonstrate consistent improvements over state-of-the-art models.</abstract>
<identifier type="citekey">wang-etal-2025-unsupervised</identifier>
<location>
<url>https://aclanthology.org/2025.findings-naacl.461/</url>
</location>
<part>
<date>2025-04</date>
<extent unit="page">
<start>8247</start>
<end>8259</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Unsupervised Sentence Representation Learning with Syntactically Aligned Negative Samples
%A Wang, Zhilan
%A Zhi, Zekai
%A Jin, Rize
%A Song, Kehui
%A Wang, He
%A Cho, Da-Jung
%Y Chiruzzo, Luis
%Y Ritter, Alan
%Y Wang, Lu
%S Findings of the Association for Computational Linguistics: NAACL 2025
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-195-7
%F wang-etal-2025-unsupervised
%X Sentence representation learning benefits from data augmentation strategies to improve model performance and generalization, yet existing approaches often encounter issues such as semantic inconsistencies and feature suppression. To address these limitations, we propose a method for generating Syntactically Aligned Negative (SAN) samples through a semantic importance-aware Masked Language Model (MLM) approach. Our method quantifies semantic contributions of individual words to produce negative samples that have substantial textual overlap with the original sentences while conveying different meanings. We further introduce Hierarchical-InfoNCE (HiNCE), a novel contrastive learning objective employing differential temperature weighting to optimize the utilization of both in-batch and syntactically aligned negative samples. Extensive evaluations across seven semantic textual similarity benchmarks demonstrate consistent improvements over state-of-the-art models.
%U https://aclanthology.org/2025.findings-naacl.461/
%P 8247-8259
Markdown (Informal)
[Unsupervised Sentence Representation Learning with Syntactically Aligned Negative Samples](https://aclanthology.org/2025.findings-naacl.461/) (Wang et al., Findings 2025)
ACL