<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lin-etal-2026-asking">
    <titleInfo>
        <title>Who Is Asking? Personalized and Tiered Memory for Long-Lived AI Assistants</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yueqian</namePart>
        <namePart type="family">Lin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sathiya</namePart>
        <namePart type="given">Kumaran</namePart>
        <namePart type="family">Mani</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yiran</namePart>
        <namePart type="family">Chen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sujata</namePart>
        <namePart type="family">Banerjee</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Vaishnavi</namePart>
        <namePart type="family">Ranganathan</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: System Demonstrations</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Sweta</namePart>
            <namePart type="family">Agrawal</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Alberto</namePart>
            <namePart type="family">Abad</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">979-8-89176-461-3</identifier>
    </relatedItem>
    <abstract>Personalized AI assistants in collaborative settings must balance private user context with shared team knowledge; i.e. the same question can have different answers, depending on who is asking. However, retrieving and reasoning over a growing shared archive, as agents do today, increases latency and cost. We present Goldilocks, a multimodal assistive-agent demo that explores a tiered personal memory architecture which combines a lightweight private working resident memory alongside a shared team memory, both grounded in a common eventstore. A sufficiency gate first tries to answer from the low-cost resident tier and only escalates other tiers when needed. To evaluate long-lived personalized collaborative assistants, we also introduce WhoQA, a 500-question evaluation set built from public GitHub history, where work-related questions and answers trace to real pull requests and commits. On WhoQA, Goldilocks’s deployed small-model configuration beats a strong hybrid RAG baseline by 7.0 points (59.0 vs. 52.0) at 13% lower estimated energy (a tier-weighted token cost), and reaches 62.4 (+10.4) with the answer model held fixed. The system serves 42.6% of queries from the resident tier. A compact social-memory scoping further lifts accuracy on teammate-specific personal details from 53 to 80. While we demonstrate with a set tier and scoping configuration, the system is generalizable to multiple tiers depending on the use case. The live demo lets users explore memory tiers, gating decisions, and interactively compare personalized responses across team personas.</abstract>
    <identifier type="citekey">lin-etal-2026-asking</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-demos.16/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>167</start>
            <end>177</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Who Is Asking? Personalized and Tiered Memory for Long-Lived AI Assistants
%A Lin, Yueqian
%A Mani, Sathiya Kumaran
%A Chen, Yiran
%A Banerjee, Sujata
%A Ranganathan, Vaishnavi
%Y Agrawal, Sweta
%Y Abad, Alberto
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-461-3
%F lin-etal-2026-asking
%X Personalized AI assistants in collaborative settings must balance private user context with shared team knowledge; i.e. the same question can have different answers, depending on who is asking. However, retrieving and reasoning over a growing shared archive, as agents do today, increases latency and cost. We present Goldilocks, a multimodal assistive-agent demo that explores a tiered personal memory architecture which combines a lightweight private working resident memory alongside a shared team memory, both grounded in a common eventstore. A sufficiency gate first tries to answer from the low-cost resident tier and only escalates other tiers when needed. To evaluate long-lived personalized collaborative assistants, we also introduce WhoQA, a 500-question evaluation set built from public GitHub history, where work-related questions and answers trace to real pull requests and commits. On WhoQA, Goldilocks’s deployed small-model configuration beats a strong hybrid RAG baseline by 7.0 points (59.0 vs. 52.0) at 13% lower estimated energy (a tier-weighted token cost), and reaches 62.4 (+10.4) with the answer model held fixed. The system serves 42.6% of queries from the resident tier. A compact social-memory scoping further lifts accuracy on teammate-specific personal details from 53 to 80. While we demonstrate with a set tier and scoping configuration, the system is generalizable to multiple tiers depending on the use case. The live demo lets users explore memory tiers, gating decisions, and interactively compare personalized responses across team personas.
%U https://aclanthology.org/2026.emnlp-demos.16/
%P 167-177

Markdown (Informal)

[Who Is Asking? Personalized and Tiered Memory for Long-Lived AI Assistants](https://aclanthology.org/2026.emnlp-demos.16/) (Lin et al., EMNLP 2026)

• Who Is Asking? Personalized and Tiered Memory for Long-Lived AI Assistants (Lin et al., EMNLP 2026)

ACL

• Yueqian Lin, Sathiya Kumaran Mani, Yiran Chen, Sujata Banerjee, and Vaishnavi Ranganathan. 2026. Who Is Asking? Personalized and Tiered Memory for Long-Lived AI Assistants. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 167–177, Budapest, Hungary. Association for Computational Linguistics.