<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zeng-etal-2026-dual">
    <titleInfo>
        <title>Dual-View Training for Instruction-Following Information Retrieval</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Qingcheng</namePart>
        <namePart type="family">Zeng</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Puxuan</namePart>
        <namePart type="family">Yu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Aman</namePart>
        <namePart type="family">Mehta</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Fuheng</namePart>
        <namePart type="family">Zhao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Rajhans</namePart>
        <namePart type="family">Samdani</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Instruction-following information retrieval (IF-IR) requires retrievers to satisfy both topical intent and instance-specific relevance constraints. We propose dual-view training, a data synthesis strategy based on polarity reversal. Given a query, an instruction-relevant document, and a topically related document excluded by that instruction, an LLM generates a complementary instruction under which the two documents exchange relevance labels. The resulting paired views directly penalize instruction-agnostic representations. At a matched 880k-example budget, replacing the non-instruction half of a Promptriever-style joint-data recipe with polarity-reversed examples improves FollowIR p-MRR from 5.27 to 8.30 (+57.5%) on a 305M-parameter encoder while increasing its instruction-conditioned Score from 20.85 to 21.38. The relative p-MRR gain repeats on a 560M-parameter backbone. The proposed recipe also improves InfoSearch nDCG@5 and shows no average degradation on four evaluated BEIR datasets with plain queries. A size-matched paraphrase control does not reproduce the instruction-sensitivity gains, supporting the conclusion that label reversal, rather than additional LLM-generated wording, provides the key supervision signal.</abstract>
    <identifier type="citekey">zeng-etal-2026-dual</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.571/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>13099</start>
            <end>13108</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Dual-View Training for Instruction-Following Information Retrieval
%A Zeng, Qingcheng
%A Yu, Puxuan
%A Mehta, Aman
%A Zhao, Fuheng
%A Samdani, Rajhans
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zeng-etal-2026-dual
%X Instruction-following information retrieval (IF-IR) requires retrievers to satisfy both topical intent and instance-specific relevance constraints. We propose dual-view training, a data synthesis strategy based on polarity reversal. Given a query, an instruction-relevant document, and a topically related document excluded by that instruction, an LLM generates a complementary instruction under which the two documents exchange relevance labels. The resulting paired views directly penalize instruction-agnostic representations. At a matched 880k-example budget, replacing the non-instruction half of a Promptriever-style joint-data recipe with polarity-reversed examples improves FollowIR p-MRR from 5.27 to 8.30 (+57.5%) on a 305M-parameter encoder while increasing its instruction-conditioned Score from 20.85 to 21.38. The relative p-MRR gain repeats on a 560M-parameter backbone. The proposed recipe also improves InfoSearch nDCG@5 and shows no average degradation on four evaluated BEIR datasets with plain queries. A size-matched paraphrase control does not reproduce the instruction-sensitivity gains, supporting the conclusion that label reversal, rather than additional LLM-generated wording, provides the key supervision signal.
%U https://aclanthology.org/2026.emnlp-main.571/
%P 13099-13108

Markdown (Informal)

[Dual-View Training for Instruction-Following Information Retrieval](https://aclanthology.org/2026.emnlp-main.571/) (Zeng et al., EMNLP 2026)

• Dual-View Training for Instruction-Following Information Retrieval (Zeng et al., EMNLP 2026)

ACL

• Qingcheng Zeng, Puxuan Yu, Aman Mehta, Fuheng Zhao, and Rajhans Samdani. 2026. Dual-View Training for Instruction-Following Information Retrieval. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 13099–13108, Budapest, Hungary. Association for Computational Linguistics.