<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2026-retain">
    <titleInfo>
        <title>When Retain Constraints Conflict: Mitigating Forget–Retain Interference in Tabular Data</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Zijie</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jinhao</namePart>
        <namePart type="family">Duan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Bingqi</namePart>
        <namePart type="family">Shang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xinming</namePart>
        <namePart type="family">An</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sijia</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Tianlong</namePart>
        <namePart type="family">Chen</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>Machine unlearning aims to remove the influence of designated training data while preserving model utility, but its behavior on tabular data remains underexplored. This gap is important because tabular prediction is widely used in high-stakes domains and is increasingly adapted to language models through record serialization and schema-aware prompting. We identify a key challenge that distinguishes tabular unlearning from unlearning in free-form text or other modalities: schema-induced forget–retain overlap. In serialized tabular data, records share fixed column-name/value slots, similar attribute ranges, and common output spaces. Consequently, a forget row may have nearby retain rows that rely on the same high-signal attributes, causing retain preservation to oppose the update required for forgetting. Motivated by this failure mode, we propose Conflict-Aware Unlearning (CAU), a schema-aware approach that reduces forget–retain interference by relaxing preservation constraints on retained rows that most conflict with the forget set. Across sample-level and feature-level unlearning on clinical and non-medical tabular tasks, more closely matches a retraining oracle while maintaining predictive utility and retain-region behavior. Our results show that reliable tabular LLM unlearning depends not only on the forgetting objective, but also on how retain constraints are constructed.</abstract>
    <identifier type="citekey">liu-etal-2026-retain</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.527/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>11052</start>
            <end>11065</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T When Retain Constraints Conflict: Mitigating Forget–Retain Interference in Tabular Data
%A Liu, Zijie
%A Duan, Jinhao
%A Shang, Bingqi
%A An, Xinming
%A Liu, Sijia
%A Chen, Tianlong
%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 liu-etal-2026-retain
%X Machine unlearning aims to remove the influence of designated training data while preserving model utility, but its behavior on tabular data remains underexplored. This gap is important because tabular prediction is widely used in high-stakes domains and is increasingly adapted to language models through record serialization and schema-aware prompting. We identify a key challenge that distinguishes tabular unlearning from unlearning in free-form text or other modalities: schema-induced forget–retain overlap. In serialized tabular data, records share fixed column-name/value slots, similar attribute ranges, and common output spaces. Consequently, a forget row may have nearby retain rows that rely on the same high-signal attributes, causing retain preservation to oppose the update required for forgetting. Motivated by this failure mode, we propose Conflict-Aware Unlearning (CAU), a schema-aware approach that reduces forget–retain interference by relaxing preservation constraints on retained rows that most conflict with the forget set. Across sample-level and feature-level unlearning on clinical and non-medical tabular tasks, more closely matches a retraining oracle while maintaining predictive utility and retain-region behavior. Our results show that reliable tabular LLM unlearning depends not only on the forgetting objective, but also on how retain constraints are constructed.
%U https://aclanthology.org/2026.findings-emnlp.527/
%P 11052-11065

Markdown (Informal)

[When Retain Constraints Conflict: Mitigating Forget–Retain Interference in Tabular Data](https://aclanthology.org/2026.findings-emnlp.527/) (Liu et al., Findings 2026)

• When Retain Constraints Conflict: Mitigating Forget–Retain Interference in Tabular Data (Liu et al., Findings 2026)

ACL

• Zijie Liu, Jinhao Duan, Bingqi Shang, Xinming An, Sijia Liu, and Tianlong Chen. 2026. When Retain Constraints Conflict: Mitigating Forget–Retain Interference in Tabular Data. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 11052–11065, Budapest, Hungary. Association for Computational Linguistics.