@inproceedings{silva-etal-2026-bridging,
title = "Bridging Citizens and Public Services: Improving Service Association with Retrieval-Augmented Generation ({RAG}) Labels",
author = "Silva, Ticiana L. Coelho da and
Fran{\c{c}}a, Celso and
Gon{\c{c}}alves, Marcos Andr{\'e} and
Rocha, Leonardo and
Alamy, Leonardo and
Pereira, Fernando Sola and
Paiva, Eduardo Soares de",
editor = "Souza, Marlo and
de-Dios-Flores, Iria and
Santos, Diana and
Freitas, Larissa and
Souza, Jackson Wilke da Cruz and
Ribeiro, Eug{\'e}nio",
booktitle = "Proceedings of the 17th International Conference on Computational Processing of {P}ortuguese ({PROPOR} 2026) - Vol. 1",
month = apr,
year = "2026",
address = "Salvador, Brazil",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.propor-1.13/",
pages = "131--140",
ISBN = "979-8-89176-387-6",
abstract = "Linking citizen complaints to the public services they concern remains a major challenge in the Brazilian federal administration. In 2025, over 1.2 million manifestations were submitted across 328 agencies, yet only about 1.8{\%} are currently associated with a specific service, limiting large-scale monitoring and evidence-based management. We cast this task as an extreme multi-class text classification problem marked by severe class imbalance and strong lexical{--}semantic gaps between citizen language and official service descriptions. Building on recent work that reframes the task as information retrieval, we combine sparse retrieval with BM25 over representative complaint corpora and dense retrieval enriched with RAG-labels: semantically expanded label descriptions generated via Retrieval-Augmented Generation and Small Language Models. This approach markedly reduces vocabulary mismatch and semantic ambiguity, yielding substantial gains over direct text or embedding matching. To our knowledge, this is the first Portuguese-language application of RAG-labels for service{--}complaint association. In real operational data from the Federal Ombudsman Office, our method can automatically assign plausible services to roughly 73{\%} of previously unlabeled cases, improving coverage and supporting more effective public service evaluation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="silva-etal-2026-bridging">
<titleInfo>
<title>Bridging Citizens and Public Services: Improving Service Association with Retrieval-Augmented Generation (RAG) Labels</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ticiana</namePart>
<namePart type="given">L</namePart>
<namePart type="given">Coelho</namePart>
<namePart type="given">da</namePart>
<namePart type="family">Silva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Celso</namePart>
<namePart type="family">França</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marcos</namePart>
<namePart type="given">André</namePart>
<namePart type="family">Gonçalves</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leonardo</namePart>
<namePart type="family">Rocha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leonardo</namePart>
<namePart type="family">Alamy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fernando</namePart>
<namePart type="given">Sola</namePart>
<namePart type="family">Pereira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eduardo</namePart>
<namePart type="given">Soares</namePart>
<namePart type="given">de</namePart>
<namePart type="family">Paiva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1</title>
</titleInfo>
<name type="personal">
<namePart type="given">Marlo</namePart>
<namePart type="family">Souza</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Iria</namePart>
<namePart type="family">de-Dios-Flores</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Diana</namePart>
<namePart type="family">Santos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Larissa</namePart>
<namePart type="family">Freitas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jackson</namePart>
<namePart type="given">Wilke</namePart>
<namePart type="given">da</namePart>
<namePart type="given">Cruz</namePart>
<namePart type="family">Souza</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eugénio</namePart>
<namePart type="family">Ribeiro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Salvador, Brazil</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-387-6</identifier>
</relatedItem>
<abstract>Linking citizen complaints to the public services they concern remains a major challenge in the Brazilian federal administration. In 2025, over 1.2 million manifestations were submitted across 328 agencies, yet only about 1.8% are currently associated with a specific service, limiting large-scale monitoring and evidence-based management. We cast this task as an extreme multi-class text classification problem marked by severe class imbalance and strong lexical–semantic gaps between citizen language and official service descriptions. Building on recent work that reframes the task as information retrieval, we combine sparse retrieval with BM25 over representative complaint corpora and dense retrieval enriched with RAG-labels: semantically expanded label descriptions generated via Retrieval-Augmented Generation and Small Language Models. This approach markedly reduces vocabulary mismatch and semantic ambiguity, yielding substantial gains over direct text or embedding matching. To our knowledge, this is the first Portuguese-language application of RAG-labels for service–complaint association. In real operational data from the Federal Ombudsman Office, our method can automatically assign plausible services to roughly 73% of previously unlabeled cases, improving coverage and supporting more effective public service evaluation.</abstract>
<identifier type="citekey">silva-etal-2026-bridging</identifier>
<location>
<url>https://aclanthology.org/2026.propor-1.13/</url>
</location>
<part>
<date>2026-04</date>
<extent unit="page">
<start>131</start>
<end>140</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Bridging Citizens and Public Services: Improving Service Association with Retrieval-Augmented Generation (RAG) Labels
%A Silva, Ticiana L. Coelho da
%A França, Celso
%A Gonçalves, Marcos André
%A Rocha, Leonardo
%A Alamy, Leonardo
%A Pereira, Fernando Sola
%A Paiva, Eduardo Soares de
%Y Souza, Marlo
%Y de-Dios-Flores, Iria
%Y Santos, Diana
%Y Freitas, Larissa
%Y Souza, Jackson Wilke da Cruz
%Y Ribeiro, Eugénio
%S Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1
%D 2026
%8 April
%I Association for Computational Linguistics
%C Salvador, Brazil
%@ 979-8-89176-387-6
%F silva-etal-2026-bridging
%X Linking citizen complaints to the public services they concern remains a major challenge in the Brazilian federal administration. In 2025, over 1.2 million manifestations were submitted across 328 agencies, yet only about 1.8% are currently associated with a specific service, limiting large-scale monitoring and evidence-based management. We cast this task as an extreme multi-class text classification problem marked by severe class imbalance and strong lexical–semantic gaps between citizen language and official service descriptions. Building on recent work that reframes the task as information retrieval, we combine sparse retrieval with BM25 over representative complaint corpora and dense retrieval enriched with RAG-labels: semantically expanded label descriptions generated via Retrieval-Augmented Generation and Small Language Models. This approach markedly reduces vocabulary mismatch and semantic ambiguity, yielding substantial gains over direct text or embedding matching. To our knowledge, this is the first Portuguese-language application of RAG-labels for service–complaint association. In real operational data from the Federal Ombudsman Office, our method can automatically assign plausible services to roughly 73% of previously unlabeled cases, improving coverage and supporting more effective public service evaluation.
%U https://aclanthology.org/2026.propor-1.13/
%P 131-140
Markdown (Informal)
[Bridging Citizens and Public Services: Improving Service Association with Retrieval-Augmented Generation (RAG) Labels](https://aclanthology.org/2026.propor-1.13/) (Silva et al., PROPOR 2026)
ACL