Logo do repositório
 
Publicação

A Retrieval-Augmented Natural Language Interface for Data Description and Meta-Analysis in the Pathogens-in-Foods (PIF) Database

datacite.subject.fosCiências Agrárias::Biotecnologia Agrária e Alimentar
datacite.subject.fosEngenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática
datacite.subject.sdg04:Educação de Qualidade
datacite.subject.sdg02:Erradicar a Fome
datacite.subject.sdg12:Produção e Consumo Sustentáveis
dc.contributor.authorSilva, Lucas Ribeiro
dc.contributor.authorGonzales-Barron, Ursula
dc.contributor.authorCadavez, Vasco
dc.date.accessioned2026-07-08T15:45:40Z
dc.date.available2026-07-08T15:45:40Z
dc.date.issued2026
dc.description.abstractFood-safety occurrence databases are increasingly important for surveillance, evidence appraisal, and quantitative risk assessment, yet their routine analytical use remains constrained by the need for database literacy and statistical programming. Building on the curated and harmonized Pathogens-in-Foods (PIF) database, we developed and evaluated a retrieval-augmented natural-language interface designed to support grounded querying and reproducible evidence synthesis. The system includes two complementary modes: an Open Chat Mode for exploratory, tool-mediated interrogation of the database and a Guided Meta-Analysis Mode that couples structured user input to a deterministic R-based analytical pipeline. Evaluation included four compact language models: Phi-4 Mini (3.8B), DeepSeek-R1 Tool-Calling (14B), Cogito (14B), and Qwen 3 (8B), together with Gemini 2.5 Pro as a larger proprietary baseline model. Within a 10-query benchmark, all models achieved 100% tool selection accuracy and retrieval correctness; for the five argument bearing queries, all models also achieved 100% argument extraction F1-score, indicating reliable grounding of database operations for the evaluated query set. In a guided case study on Toxoplasma in meat and meat products (153 records from 65 studies), the system achieved 100% numerical concordance and high visual informativeness; the highest report quality index was 93% with Qwen 3 (8B). Performance differences across models arose primarily from the factual precision and economy of their written interpretations rather than from failures in tool execution. These findings support hybrid, evidence-grounded analytical interfaces built on curated data resources and deterministic statistical backends as practical tools for accelerating surveillance-oriented evidence synthesis in food protection.eng
dc.description.sponsorshipThis work was funded by the European Food Safety Authority (EFSA) through GP/EFSA/BIOHAW/2022/01 and GP/EFSA/BIOHAW/2023/05. The positions and opinions presented in this study are those of the authors and are not intended to represent the views or scientific works of EFSA.
dc.identifier.citationSilva, Lucas Ribeiro; Gonzales-Barron, Ursula; Cadavez, Vasco (2026). A Retrieval-Augmented Natural Language Interface for Data Description and Meta-Analysis in the Pathogens-in-Foods (PIF) Database ✩. Journal of Food Protection. ISSN 0362-028X. 89:8, p. 1-11
dc.identifier.doi10.1016/j.jfp.2026.100829
dc.identifier.issn0362-028X
dc.identifier.urihttp://hdl.handle.net/10198/36992
dc.language.isoeng
dc.peerreviewedyes
dc.publisherElsevier
dc.relationPathogens in foods database - IPB - GP/EFSA/BIOHAW/2022/01
dc.relationNewPIF - Pathogens in foods database: extension with Vibrio and parasites - GP/EFSA/BIOHAW/2023/05
dc.relation.ispartofJournal of Food Protection
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subjectData visualization
dc.subjectFood safety
dc.subjectMeta-analysis
dc.subjectNatural language interface
dc.subjectOpen-weight language models
dc.subjectTool-using agents
dc.titleA Retrieval-Augmented Natural Language Interface for Data Description and Meta-Analysis in the Pathogens-in-Foods (PIF) Databaseeng
dc.typejournal article
dspace.entity.typePublication
oaire.awardNumberGP/EFSA/BIOHAW/2022/01
oaire.awardNumberGP/EFSA/BIOHAW/2023/05
oaire.awardTitlePathogens in foods database - IPB - GP/EFSA/BIOHAW/2022/01
oaire.awardTitleNewPIF - Pathogens in foods database: extension with Vibrio and parasites - GP/EFSA/BIOHAW/2023/05
oaire.awardURIhttp://hdl.handle.net/10198/36991
oaire.awardURIhttps://pif.esa.ipb.pt/blog/2023/09/01/extension-with-vibrio-and-parasites/
oaire.citation.endPage11
oaire.citation.issue8
oaire.citation.startPage1
oaire.citation.titleJournal of Food Protection
oaire.citation.volume89
oaire.fundingStreamEFSA grants
oaire.versionhttp://purl.org/coar/version/c_970fb48d4fbd8a85
person.familyNameGonzales-Barron
person.familyNameCadavez
person.givenNameUrsula
person.givenNameVasco
person.identifierR-000-HDG
person.identifier.ciencia-id0813-C319-B62A
person.identifier.ciencia-id441B-01AB-A12E
person.identifier.orcid0000-0002-8462-9775
person.identifier.orcid0000-0002-3077-7414
person.identifier.ridA-3958-2010
person.identifier.scopus-author-id9435483700
person.identifier.scopus-author-id9039121900
relation.isAuthorOfPublication17c6b98f-4fb5-41d3-839a-6f77ec70021a
relation.isAuthorOfPublication57b410e9-f6b7-42ff-ab3d-b526278715eb
relation.isAuthorOfPublication.latestForDiscovery17c6b98f-4fb5-41d3-839a-6f77ec70021a
relation.isProjectOfPublication54e63134-dc30-4865-8a60-4ef3598f9fee
relation.isProjectOfPublication52ec4558-4739-4120-b887-4183f5da2c12
relation.isProjectOfPublication.latestForDiscovery54e63134-dc30-4865-8a60-4ef3598f9fee

Ficheiros

Principais
A mostrar 1 - 1 de 1
A carregar...
Miniatura
Nome:
1-s2.0-S0362028X26001341-main.pdf
Tamanho:
1.66 MB
Formato:
Adobe Portable Document Format
Licença
A mostrar 1 - 1 de 1
Miniatura indisponível
Nome:
license.txt
Tamanho:
1.75 KB
Formato:
Item-specific license agreed upon to submission
Descrição: