Three Filters: an undergraduate thesis filters scientific metadata in three layers
Leonardo Milomes Vitoriano and Mateus de Almeida Dias defended at FCTE/UnB an undergraduate thesis presenting Three Filters — a system to ensure quality and integrity of scientific metadata in JSON.
Brasília, July 28, 2025
Bad scientific metadata is expensive: once it enters a repository, it contaminates searches, harms citations, and delays bibliometric analyses. The undergraduate thesis defended in July 2025 by Leonardo Milomes Vitoriano and Mateus de Almeida Dias builds a system to address that problem, under the advising of Prof. André Luiz Peron Martins Lanna .
Three Filters — a name that describes the architecture — operates in three complementary layers: structural correctness (is the JSON well formed?), informational correctness (is the data plausible?), and semantic coherence (does the data hold up in context?). Each layer uses appropriate techniques — schema validation, plausibility rules, machine learning — and the ensemble delivers a filter capable of flagging errors before they propagate.
By tying applied AI to concrete instruments of scientific curation, the thesis speaks to an agenda gaining weight as large repositories (OpenAlex, Lattes, ORCID) become decision bases for science policy. Three Filters shows that data quality in science needs infrastructure like any other quality — and that this infrastructure can be engineered.
The full text — written in Portuguese — is available at the UnB Undergraduate Theses Digital Library.
About CEDIS: The Center for Studies, Development, and Innovation in Software (CEDIS), linked to the University of Brasília, researches and develops innovative software solutions.