@inproceedings{campese-etal-2026-proprietary,
title = "From Proprietary to Public: Dataset Declassification for Question Answering",
author = "Campese, Stefano and
Moschitti, Alessandro and
Lauriola, Ivano",
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.1112/",
pages = "23526--23541",
ISBN = "979-8-89176-462-0",
abstract = "Training Question Answering (QA) systems on real customer interaction data is often infeasible to reproduce or share due to privacy and intellectual property constraints. We introduce dataset declassification, a framework that replaces proprietary QA datasets with public, semantically equivalent counterparts that preserve training and evaluation utility while satisfying typical legal and privacy requirements. The approach maps proprietary questions to public equivalents via large-scale retrieval and reconstructs answers from domain-matched corpora using LLM-based annotation. Experiments on four QA benchmarks show that models trained and evaluated on declassified datasets achieve performance within $\Delta \approx 0$ of the original data. The method targets datasets whose underlying information needs already exist in public knowledge sources (e.g., knowledge available on the web and used by public-facing services). We further provide explicit applicability conditions showing that declassification succeeds when sufficient semantic overlap exists between proprietary and public corpora, and degrades predictably when it does not."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="campese-etal-2026-proprietary">
<titleInfo>
<title>From Proprietary to Public: Dataset Declassification for Question Answering</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stefano</namePart>
<namePart type="family">Campese</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alessandro</namePart>
<namePart type="family">Moschitti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ivano</namePart>
<namePart type="family">Lauriola</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>Training Question Answering (QA) systems on real customer interaction data is often infeasible to reproduce or share due to privacy and intellectual property constraints. We introduce dataset declassification, a framework that replaces proprietary QA datasets with public, semantically equivalent counterparts that preserve training and evaluation utility while satisfying typical legal and privacy requirements. The approach maps proprietary questions to public equivalents via large-scale retrieval and reconstructs answers from domain-matched corpora using LLM-based annotation. Experiments on four QA benchmarks show that models trained and evaluated on declassified datasets achieve performance within Δ \approx 0 of the original data. The method targets datasets whose underlying information needs already exist in public knowledge sources (e.g., knowledge available on the web and used by public-facing services). We further provide explicit applicability conditions showing that declassification succeeds when sufficient semantic overlap exists between proprietary and public corpora, and degrades predictably when it does not.</abstract>
<identifier type="citekey">campese-etal-2026-proprietary</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1112/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>23526</start>
<end>23541</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T From Proprietary to Public: Dataset Declassification for Question Answering
%A Campese, Stefano
%A Moschitti, Alessandro
%A Lauriola, Ivano
%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 campese-etal-2026-proprietary
%X Training Question Answering (QA) systems on real customer interaction data is often infeasible to reproduce or share due to privacy and intellectual property constraints. We introduce dataset declassification, a framework that replaces proprietary QA datasets with public, semantically equivalent counterparts that preserve training and evaluation utility while satisfying typical legal and privacy requirements. The approach maps proprietary questions to public equivalents via large-scale retrieval and reconstructs answers from domain-matched corpora using LLM-based annotation. Experiments on four QA benchmarks show that models trained and evaluated on declassified datasets achieve performance within Δ \approx 0 of the original data. The method targets datasets whose underlying information needs already exist in public knowledge sources (e.g., knowledge available on the web and used by public-facing services). We further provide explicit applicability conditions showing that declassification succeeds when sufficient semantic overlap exists between proprietary and public corpora, and degrades predictably when it does not.
%U https://aclanthology.org/2026.findings-emnlp.1112/
%P 23526-23541
Markdown (Informal)
[From Proprietary to Public: Dataset Declassification for Question Answering](https://aclanthology.org/2026.findings-emnlp.1112/) (Campese et al., Findings 2026)
ACL