@inproceedings{matos-etal-2026-benchmarking,
title = "Benchmarking Retrieval-Augmented Generation for Scientific Knowledge {QA} in {E}uropean {P}ortuguese",
author = "Matos, Jose and
Silva, Catarina and
Goncalo Oliveira, Hugo",
editor = "Rehm, Georg and
Dietze, Stefan and
Dessi, Danilo and
Maynard, Diana and
Schimmler, Sonja",
booktitle = "Proceedings of Natural Scientific Language Processing ({NSLP}) @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.nslp-1.3/",
doi = "10.63317/3muergicuxwk",
pages = "25--31",
abstract = "Retrieval-Augmented Generation (RAG) enables grounding of model outputs in external evidence, but its impact on European Portuguese (pt-PT) scientific question answering (QA) remains unclear. We present a controlled evaluation of RAG on pt-PT knowledge QA across different scientific domains using the Portuguese test split of the Global MMLU Lite dataset. As external evidence, we use a Portuguese scientific literature knowledge base containing over 32,000 documents converted to Markdown. We benchmark five instruction-tuned small language models (4-12B) and compare closed-book baselines against 16 RAG configurations that vary by: (i) dense retriever specialization (multilingual vs. Portuguese-specific), (ii) reranking (on/off), and (iii) number of retrieved chunks (k {\ensuremath{\in}} {1, 3, 5, 10}). Results suggest that RAG gains are model-dependent. Some models improve consistently, others are highly sensitive to retrieval choices, and some degrade under retrieval noise, especially at larger values of k. Findings highlight the importance of model-specific retrieval tuning and ensuring that the retriever and reranker languages and domains align when deploying RAG systems for Portuguese natural scientific language processing."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="matos-etal-2026-benchmarking">
<titleInfo>
<title>Benchmarking Retrieval-Augmented Generation for Scientific Knowledge QA in European Portuguese</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jose</namePart>
<namePart type="family">Matos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Catarina</namePart>
<namePart type="family">Silva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hugo</namePart>
<namePart type="family">Goncalo Oliveira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of Natural Scientific Language Processing (NSLP) @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Georg</namePart>
<namePart type="family">Rehm</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stefan</namePart>
<namePart type="family">Dietze</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Danilo</namePart>
<namePart type="family">Dessi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Diana</namePart>
<namePart type="family">Maynard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sonja</namePart>
<namePart type="family">Schimmler</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Retrieval-Augmented Generation (RAG) enables grounding of model outputs in external evidence, but its impact on European Portuguese (pt-PT) scientific question answering (QA) remains unclear. We present a controlled evaluation of RAG on pt-PT knowledge QA across different scientific domains using the Portuguese test split of the Global MMLU Lite dataset. As external evidence, we use a Portuguese scientific literature knowledge base containing over 32,000 documents converted to Markdown. We benchmark five instruction-tuned small language models (4-12B) and compare closed-book baselines against 16 RAG configurations that vary by: (i) dense retriever specialization (multilingual vs. Portuguese-specific), (ii) reranking (on/off), and (iii) number of retrieved chunks (k \ensuremathın 1, 3, 5, 10). Results suggest that RAG gains are model-dependent. Some models improve consistently, others are highly sensitive to retrieval choices, and some degrade under retrieval noise, especially at larger values of k. Findings highlight the importance of model-specific retrieval tuning and ensuring that the retriever and reranker languages and domains align when deploying RAG systems for Portuguese natural scientific language processing.</abstract>
<identifier type="citekey">matos-etal-2026-benchmarking</identifier>
<identifier type="doi">10.63317/3muergicuxwk</identifier>
<location>
<url>https://aclanthology.org/2026.nslp-1.3/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>25</start>
<end>31</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Benchmarking Retrieval-Augmented Generation for Scientific Knowledge QA in European Portuguese
%A Matos, Jose
%A Silva, Catarina
%A Goncalo Oliveira, Hugo
%Y Rehm, Georg
%Y Dietze, Stefan
%Y Dessi, Danilo
%Y Maynard, Diana
%Y Schimmler, Sonja
%S Proceedings of Natural Scientific Language Processing (NSLP) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F matos-etal-2026-benchmarking
%X Retrieval-Augmented Generation (RAG) enables grounding of model outputs in external evidence, but its impact on European Portuguese (pt-PT) scientific question answering (QA) remains unclear. We present a controlled evaluation of RAG on pt-PT knowledge QA across different scientific domains using the Portuguese test split of the Global MMLU Lite dataset. As external evidence, we use a Portuguese scientific literature knowledge base containing over 32,000 documents converted to Markdown. We benchmark five instruction-tuned small language models (4-12B) and compare closed-book baselines against 16 RAG configurations that vary by: (i) dense retriever specialization (multilingual vs. Portuguese-specific), (ii) reranking (on/off), and (iii) number of retrieved chunks (k \ensuremathın 1, 3, 5, 10). Results suggest that RAG gains are model-dependent. Some models improve consistently, others are highly sensitive to retrieval choices, and some degrade under retrieval noise, especially at larger values of k. Findings highlight the importance of model-specific retrieval tuning and ensuring that the retriever and reranker languages and domains align when deploying RAG systems for Portuguese natural scientific language processing.
%R 10.63317/3muergicuxwk
%U https://aclanthology.org/2026.nslp-1.3/
%U https://doi.org/10.63317/3muergicuxwk
%P 25-31
Markdown (Informal)
[Benchmarking Retrieval-Augmented Generation for Scientific Knowledge QA in European Portuguese](https://aclanthology.org/2026.nslp-1.3/) (Matos et al., NSLP 2026)
ACL