<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mansoori-etal-2026-defer">
    <titleInfo>
        <title>When to Defer: Scope-Aware Cascades for Structured Attribute Generation in E-commerce Product Catalogs</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Fatemeh</namePart>
        <namePart type="family">Mansoori</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Andrea</namePart>
        <namePart type="family">Scarinci</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Aditya</namePart>
        <namePart type="family">Aggarwal</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ashwin</namePart>
        <namePart type="family">Chandramouli</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Suleiman</namePart>
        <namePart type="given">A</namePart>
        <namePart type="family">Khan</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: Industry Track</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Pepa</namePart>
            <namePart type="family">Atanasova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Gholamreza</namePart>
            <namePart type="family">Haffari</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Minghan</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">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-460-6</identifier>
    </relatedItem>
    <abstract>Fine-tuned small language models (SLMs) are attractive defaults for large-scale structured at- tribute generation—producing values such as materials, colors, or dimensions from product descriptions—as they are inexpensive to serve. However, a fine-tuned SLM is not reliable ev- erywhere as its correctness varies sharply, so deploying it still requires deciding, per output, whether to keep the SLM’s value or escalate it to a more capable fallback. The standard pol- icy applies a single confidence threshold, but one threshold cannot separate scores on reli- able tasks from those on unreliable ones, so it reduces coverage at no gain in quality. We demonstrate that in a fixed deployment scope, where the target tasks recur and validation data is available, a lightweight scope-aware router using only SLM-side features (token confi- dence plus task-scope identity) selects risky predictions far better. Our experiments show that at a fixed precision target it accepts about 11 points more values than a global confidence threshold (family-level calibration helps only marginally). Ablation experiments confirm the gain comes from task-scope reliability priors rather than a better confidence transformation. The router is an 8-feature gradient-boosted tree that adds negligible inference cost, making a strong fine-tuned SLM the low-cost tier of a cascade that optimizes the quality–cost tradeoff in real production settings.</abstract>
    <identifier type="citekey">mansoori-etal-2026-defer</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-industry.187/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>3149</start>
            <end>3160</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T When to Defer: Scope-Aware Cascades for Structured Attribute Generation in E-commerce Product Catalogs
%A Mansoori, Fatemeh
%A Scarinci, Andrea
%A Aggarwal, Aditya
%A Chandramouli, Ashwin
%A Khan, Suleiman A.
%Y Atanasova, Pepa
%Y Haffari, Gholamreza
%Y Wang, Minghan
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-460-6
%F mansoori-etal-2026-defer
%X Fine-tuned small language models (SLMs) are attractive defaults for large-scale structured at- tribute generation—producing values such as materials, colors, or dimensions from product descriptions—as they are inexpensive to serve. However, a fine-tuned SLM is not reliable ev- erywhere as its correctness varies sharply, so deploying it still requires deciding, per output, whether to keep the SLM’s value or escalate it to a more capable fallback. The standard pol- icy applies a single confidence threshold, but one threshold cannot separate scores on reli- able tasks from those on unreliable ones, so it reduces coverage at no gain in quality. We demonstrate that in a fixed deployment scope, where the target tasks recur and validation data is available, a lightweight scope-aware router using only SLM-side features (token confi- dence plus task-scope identity) selects risky predictions far better. Our experiments show that at a fixed precision target it accepts about 11 points more values than a global confidence threshold (family-level calibration helps only marginally). Ablation experiments confirm the gain comes from task-scope reliability priors rather than a better confidence transformation. The router is an 8-feature gradient-boosted tree that adds negligible inference cost, making a strong fine-tuned SLM the low-cost tier of a cascade that optimizes the quality–cost tradeoff in real production settings.
%U https://aclanthology.org/2026.emnlp-industry.187/
%P 3149-3160

Markdown (Informal)

[When to Defer: Scope-Aware Cascades for Structured Attribute Generation in E-commerce Product Catalogs](https://aclanthology.org/2026.emnlp-industry.187/) (Mansoori et al., EMNLP 2026)

• When to Defer: Scope-Aware Cascades for Structured Attribute Generation in E-commerce Product Catalogs (Mansoori et al., EMNLP 2026)

ACL

• Fatemeh Mansoori, Andrea Scarinci, Aditya Aggarwal, Ashwin Chandramouli, and Suleiman A. Khan. 2026. When to Defer: Scope-Aware Cascades for Structured Attribute Generation in E-commerce Product Catalogs. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 3149–3160, Budapest, Hungary. Association for Computational Linguistics.