Answers are grounded in this paper and cite their sources.
This record contains a research preprint and its supplementary reproducibility materials. The preprint has not undergone peer review. Abstract Public-service queries can mention several locations while only one controls the requested service. We study this distinction in UAE GovTrust, a controlled English-Arabic retrieval diagnostic comprising 22 authored service summaries and 50 scenarios with 110 query variants. Twelve methods separate lexical and dense ranking, rank fusion, scope filtering, a multilingual cross-encoder, and direct service lookup. On 72 matched answerable queries with bilingual evidence, role-filtered reciprocal rank fusion retrieves a relevant first result for 69 queries (95.8%), versus 47 (65.3%) for unfiltered fusion; filtered dense retrieval and direct lookup each achieve 68 (94.4%). With English-only evidence, excluding zero-score lexical candidates raises filtered fusion from 52 to 64 correct first hits. However, the unchanged router selects the correct action and scope for only 2 of 168 post-audit perturbations covering paraphrases, irrelevant facts, negation, and corrections. These perturbations are deliberately authored diagnostics, not held-out evaluation. Fresh-environment execution reproduces all 1,320 original ranking orders and actions. The results show that correct scope metadata can improve ranking while brittle scope inference can remove the evidence needed to answer. The study evaluates retrieval and routing, not generated answers or legal correctness; independent Arabic and domain validation remains necessary. Files UAE_GovTrust_Manuscript_Preflight_Checked.pdf — Research manuscript. UAE_GovTrust_Supplementary_S1_Reproducibility.zip — Reproducibility materials, including source code, authored diagnostic data, stored experimental outputs, dependency information, model manifests and checksum verification instructions. Extract the archive and consult reproducibility/README.txt. Authorship and funding Aleksandar Mihajlovic is the sole author. The study was self-funded by the author. No external funding or sponsorship supported the research. Academic affiliation: Technical Faculty, Singidunum University, Belgrade, Serbia. The author is a PhD student based in Abu Dhabi, United Arab Emirates. The manuscript includes employment and AI-assistance disclosures. The author retains responsibility for the work. Software licensing The MIT License included in the supplementary archive applies to the original software code only. It does not relicense the manuscript, candidate dataset, government websites, external model weights or third-party publications.
No takes yet. Share an insight, caveat, or question.
Aleksandar Mihajlović (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: