Fabíola Andrade Souza

Mrs. Fabíola Andrade Souza

Polytechnic School, Federal University of Bahia, Salvador, Bahia, Brazil
orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
Fabíola Andrade SouzaFabíola Andrade Souza
Fabíola Andrade Souza
SCSilvana Philippi Camboim
Silvana Philippi Camboim
Semantic interoperability remains a critical challenge in Spatial Data Infrastructures (SDIs), particularly when aligning authoritative taxonomies with collaborative folksonomies. Traditional alignment tools often fail to bridge the semantic and structural asymmetry between these schemas. This paper evaluates the capability of Large Language Models (LLMs), specifically distinguishing between traditional architectures and emerging Large Reasoning Models (LRMs), to perform semantic alignment between the Brazilian national topographic data model standard (EDGV) and OpenStreetMap (OSM). Using a formal ontology as a prompting scaffold, we tested seven model versions (including ChatGPT 5, DeepSeek R1, and Gemini 2.5) on their ability to bridge the gap between rigid hierarchical classes and the dynamic, ‘long-tail’ vocabulary of the folksonomy. Results reveal a distinct trade-off: while traditional LLMs exhibited ‘lexical rigidity’ and popularity bias—failing to map low-frequency tags—Reasoning Models demonstrated significantly improved capacity for semantic expansion, correctly identifying complex many-to-one (n:1) relationships across linguistic barriers. However, this reasoning depth often came at the cost of ‘hallucination by over-specification’ and syntactic instability in generating OWL code. We conclude that a neuro-symbolic approach, positioning LRMs as ‘Semantic Catalysts’ within a Human-in-the-Loop (HITL) workflow, provides a viable pathway for interoperability, balancing generative power with the need for logical rigor and spatial validation.

13 March 2026 in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Verified
Journal Article
Open access
PDF file
JPJaqueline Pisetta
Jaqueline Pisetta
Fabíola Andrade SouzaFabíola Andrade Souza
Fabíola Andrade Souza
JAJaqueline Amorim
Jaqueline Amorim
NANathan D. Antonio
Nathan D. Antonio
DNDarlan M. Nunes
Darlan M. Nunes
SCSilvana P. Camboim
Silvana P. Camboim
For decades, the focus of geospatial artificial intelligence (AI) has been on imagery and data, with linguistic interfaces remaining unexplored. This study presents a systematic review of research applying Large Language Models (LLMs) to Cartography and GIScience. We analysed 54 peer-reviewed articles published between 2023 and 2025, mapping the use of LLMs in data acquisition, semantic enrichment, spatial analysis, cartographic design, and user interaction, among other topics. The corpus reveals four dominant application clusters: (1) semantic creation/alignment of Geo-knowledge graphs; (2) text- and vision-based data acquisition; (3) language-driven spatial analysis and Geometric Question Answering (GeoQA); and (4) automated map styling and symbolisation. GPT-3.5/4 underpins two-thirds of the studies, while open-weight models, such as LLaMA-2, FLAN-T5, Gemini, and DeepSeek, are gaining traction. LLM work aligns most strongly with the challenges of openness and reproducibility, as well as cartographic design automation, but is noticeably weaker in areas such as provenance ethics, causal inference, participatory mapping, and mobile multimodal interaction. We outline three priorities for future research: (i) open benchmark datasets for spatial reasoning and map quality; (ii) ethics checklists that surface bias, privacy and hallucination risks; and (iii) investment in multilingual, low-resource Geo- LLMs to broaden global participation. By mapping current advances against long-standing research gaps, the review provides an actionable agenda for guiding large language models (LLMs) toward equitable and trustworthy cartographic practice.

27 May 2025 in Revista Brasileira de Cartografia
Verified
Journal Article
Open access
PDF file
Fabíola Andrade SouzaFabíola Andrade Souza
Fabíola Andrade Souza
ESEstephanie Daiane Batista da Silva
Estephanie Daiane Batista da Silva
SCSilvana Philippi Camboim
Silvana Philippi Camboim
Diante do cenário atual, em que o exponencial crescimento na produção de dados geoespaciais converge com a necessidade de sua disseminação e compartilhamento, torna-se salutar o desenvolvimento de mecanismos que facilitem a interoperabilidade dos dados, cujas fontes de produção podem ser diversas. Assim, questões voltadas a promover os processos de interoperabilidade semântica, entre distintas modelagens conceituais destes dados, tornam-se relevantes. Deste modo, este artigo investiga o potencial de utilização de uma ferramenta de processamento de linguagem natural, construída sobre um grande modelo de linguagem (Large Language Model - LLM), como um elemento facilitador para futura automatização dos mecanismos de alinhamento semântico entre esquemas conceituais diversos. Como resultado, a ferramenta utilizada – ChatGPT – apresentou 123 associações semânticas entre os esquemas utilizados: 34 classes da categoria edificações da base cartográfica de referência do Brasil e quaisquer tags aplicadas para criação de dados voluntários no OpenStreetMap (OSM). Em alguns casos, as associações foram detalhadas, em outros, mais genéricas, sendo possível sua comparação com trabalhos prévios realizados manualmente por humanos. Importante salientar o papel relevante da construção do diálogo de solicitação do alinhamento, com organização estruturada dos dados conceituais, bem como utilização de diálogo claro e sem ambiguidades. Ainda existem limitações no processo, em especial para entendimento da hierarquia dos conceitos utilizados, o que indica necessidade de novos estudos e avaliação de outros LLM disponíveis. Entretanto, o uso de inteligência artificial para interoperabilidade semântica de dados geoespaciais desponta como um caminho viável a ser aplicado.

22 December 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
Fabíola Andrade SouzaFabíola Andrade Souza
Fabíola Andrade Souza
SCSilvana Philippi Camboim
Silvana Philippi Camboim
Semantic interoperability remains a critical challenge in Spatial Data Infrastructures (SDIs), particularly when aligning authoritative taxonomies with collaborative folksonomies. Recent advances in Large Language Models (LLMs) offer new avenues for automated semantic interpretation, yet these 'sub-symbolic' approaches often lack the logical rigor required for structured geospatial data. This paper evaluates the capability of LLMs – specifically distinguishing between traditional architectures and emerging Large Reasoning Models (LRMs) – to perform semantic alignment between the Brazilian national cartographic standard (EDGV) and OpenStreetMap (OSM). Using a formal ontology as a prompting scaffold, we tested seven model versions (including ChatGPT 5.0, DeepSeek R1, and Gemini 2.5) on their ability to identify semantic equivalents and generate valid ontological mappings. Results indicate that while traditional LLMs struggle with hierarchical structures, reasoning-oriented models demonstrate significantly improved capacity for complex inference, correctly identifying many-to-one (n:1) relationships across linguistic barriers. However, all models exhibited limitations in generating syntactically valid OWL code, revealing a gap between semantic comprehension and formal structuring. We conclude that a neuro-symbolic approach, using ontologies to ground AI reasoning, provides a viable pathway for semi-automated interoperability, although future work must address the lack of explicit spatial reasoning in current architectures.

12 August 2023 in Abstracts of the ICA
Verified
Journal Article
Open access
PDF file
Fabíola Andrade SouzaFabíola Andrade Souza
Fabíola Andrade Souza
LSLeonardo Scharth Loureiro Silva
Leonardo Scharth Loureiro Silva
AMAdriana Alexandria Machado
Adriana Alexandria Machado
SSSabrina de Andrade Oliveira Santos
Sabrina de Andrade Oliveira Santos
ÉCÉrika Do Carmo Cerqueira
Érika Do Carmo Cerqueira
SCSilvana Philippi Camboim
Silvana Philippi Camboim

23 June 2025 in Revista Brasileira de Cartografia
Verified
Journal Article
Open access
PDF file
PSPablo De Almeida dos Santos
Pablo De Almeida dos Santos
Fabíola Andrade SouzaFabíola Andrade Souza
Fabíola Andrade Souza
One of the main challenges currently faced by municipal administrations concerns disorderly urban occupation, which usually leads to disasters such as landslides and floods. Unfortunately, it is observed that historically, civil defense institutions have focused more on response actions rather than disaster prevention. Aiming to change this scenario, the civil defense of the city of Salvador-Bahia has been pursuing a new management model using risk mapping of critical areas to support planning and prevention efforts. To support this initiative, a conceptual modeling of geographic data was developed to aid in the production of thematic cartography for civil defense – aligned with the Technical Specification for Structuring Vector Geospatial Data (ET-EDGV Salvador) – and enabled the organization and availability of thematic data for the stakeholders involved in the process.