@inproceedings{li-etal-2026-beyond-single,
title = "Beyond Single Vector: Query-Aware Multi-Perspective Document Embeddings for Retrieval",
author = "Li, Zhen and
Tao, Chongyang and
Chen, Yihan and
Zhang, Huishuai and
Zhao, Dongyan",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.613/",
doi = "10.18653/2026.findings-emnlp.613",
pages = "12898--12914",
ISBN = "979-8-89176-462-0",
abstract = "Recently, large language models (LLMs) have demonstrated excellent performance on general information retrieval tasks, which encode query and document into an embedding vector representing semantic information respectively. However, single-vector representations often fail to capture the multifaceted semantics of informative documents, limiting the adaptability across diverse scenarios. To address this, we propose PRISM (multi-PeRspectIve Semantic inforMation retrieval), a novel framework that leverages LLMs' implicit reasoning capabilities to extract and structure diverse semantic perspectives. PRISM introduces query-aware semantic fusion to dynamically integrate relevant perspectives during retrieval through weighted query embedding interactions. Furthermore, we propose the contrastive semantic diversification, a training mechanism that disperses semantic embeddings to maximize perspective diversity. Evaluated on the English retrieval subsets of massive multilingual text embedding benchmark (MMTEB), PRISM achieves state-of-the-art general results compared with models trained exclusively on public data."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-beyond-single">
<titleInfo>
<title>Beyond Single Vector: Query-Aware Multi-Perspective Document Embeddings for Retrieval</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhen</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chongyang</namePart>
<namePart type="family">Tao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yihan</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huishuai</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dongyan</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Recently, large language models (LLMs) have demonstrated excellent performance on general information retrieval tasks, which encode query and document into an embedding vector representing semantic information respectively. However, single-vector representations often fail to capture the multifaceted semantics of informative documents, limiting the adaptability across diverse scenarios. To address this, we propose PRISM (multi-PeRspectIve Semantic inforMation retrieval), a novel framework that leverages LLMs’ implicit reasoning capabilities to extract and structure diverse semantic perspectives. PRISM introduces query-aware semantic fusion to dynamically integrate relevant perspectives during retrieval through weighted query embedding interactions. Furthermore, we propose the contrastive semantic diversification, a training mechanism that disperses semantic embeddings to maximize perspective diversity. Evaluated on the English retrieval subsets of massive multilingual text embedding benchmark (MMTEB), PRISM achieves state-of-the-art general results compared with models trained exclusively on public data.</abstract>
<identifier type="citekey">li-etal-2026-beyond-single</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.613</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.613/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>12898</start>
<end>12914</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Beyond Single Vector: Query-Aware Multi-Perspective Document Embeddings for Retrieval
%A Li, Zhen
%A Tao, Chongyang
%A Chen, Yihan
%A Zhang, Huishuai
%A Zhao, Dongyan
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F li-etal-2026-beyond-single
%X Recently, large language models (LLMs) have demonstrated excellent performance on general information retrieval tasks, which encode query and document into an embedding vector representing semantic information respectively. However, single-vector representations often fail to capture the multifaceted semantics of informative documents, limiting the adaptability across diverse scenarios. To address this, we propose PRISM (multi-PeRspectIve Semantic inforMation retrieval), a novel framework that leverages LLMs’ implicit reasoning capabilities to extract and structure diverse semantic perspectives. PRISM introduces query-aware semantic fusion to dynamically integrate relevant perspectives during retrieval through weighted query embedding interactions. Furthermore, we propose the contrastive semantic diversification, a training mechanism that disperses semantic embeddings to maximize perspective diversity. Evaluated on the English retrieval subsets of massive multilingual text embedding benchmark (MMTEB), PRISM achieves state-of-the-art general results compared with models trained exclusively on public data.
%R 10.18653/2026.findings-emnlp.613
%U https://aclanthology.org/2026.findings-emnlp.613/
%U https://doi.org/10.18653/2026.findings-emnlp.613
%P 12898-12914
Markdown (Informal)
[Beyond Single Vector: Query-Aware Multi-Perspective Document Embeddings for Retrieval](https://aclanthology.org/2026.findings-emnlp.613/) (Li et al., Findings 2026)
ACL