https://hdl.handle.net/10216/1731 2026-04-17T09:20:53Z 2026-04-17T09:20:53Z https://hdl.handle.net/10216/166437 2026-04-14T06:46:36Z 2024-01-01T00:00:00Z

Title: Expanding FLORES+ Benchmark for More Low-Resource Settings: Portuguese-Emakhuwa Machine Translation Evaluation Abstract: As part of the Open Language Data Initiative shared tasks, we have expanded the FLORES+ evaluation set to include Emakhuwa, a low-resource language widely spoken in Mozambique. We translated the dev and devtest sets from Portuguese into Emakhuwa, and we detail the translation process and quality assurance measures used. Our methodology involved various quality checks, including post-editing and adequacy assessments. The resulting datasets consist of multiple reference sentences for each source. We present baseline results from training a Neural Machine Translation system and fine-tuning existing multilingual translation models. Our findings suggest that spelling inconsistencies remain a challenge in Emakhuwa. Additionally, the baseline models underperformed on this evaluation set, underscoring the necessity for further research to enhance machine translation quality for Emakhuwa. The data is publicly available at https://huggingface.co/datasets/LIACC/Emakhuwa-FLORES (c)2024 Association for Computational Linguistics.

2024-01-01T00:00:00Z https://hdl.handle.net/10216/108884 2026-03-21T07:47:19Z 2017-01-01T00:00:00Z

Title: Exportação lexical do português para o cingalês do século XVI: processos de nativização fonológica

2017-01-01T00:00:00Z https://hdl.handle.net/10216/122300 2026-03-21T07:47:10Z 2017-01-01T00:00:00Z

Title: Língua e literatura, ou a Mátria revisitada

2017-01-01T00:00:00Z https://hdl.handle.net/10216/172822 2026-03-17T07:45:10Z 2025-01-01T00:00:00Z

Title: A Memória e a Cidade (Reflexões sobre a prosa de David Mourão-Ferreira)

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171771 2026-03-16T07:41:43Z 2025-01-01T00:00:00Z

Title: The incremental process of building an annotation scheme for clinical narratives in portuguese: the contribution of human variation analysis Abstract: The development of a robust annotation scheme and corresponding guidelines is crucial for pro- ducing annotated datasets that advance both lin- guistic and computational research. This paper presents a case study that outlines a method- ology for designing an annotation scheme and its guidelines, specifically aimed at represent- ing morphosyntactic and semantic information regarding temporal features, as well as medi- cal information in medical reports written in Portuguese. We detail a multi-step process that includes reviewing existing frameworks, con- ducting an annotation experiment to determine the optimal approach, and designing a model based on these findings. We validated the ap- proach through a pilot experiment where we assessed the reliability and applicability of the annotation scheme and guidelines. In this ex- periment, two annotators independently anno- tated a patient's medical report consisting of six documents using the proposed model, while a curator established the ground truth. The analy- sis of inter-annotator agreement and the annota- tion results enabled the identification of sources of human variation and provided insights for further refinement of the annotation scheme and guidelines.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171504 2026-03-16T07:41:38Z 2025-01-01T00:00:00Z

Title: Can ISO 24617-1 go clinical? Extending a General-Domain Scheme to Medical Narratives Abstract: The definition of rigorous and well-structured annotation schemes is a key element in the advancement of Natural Language Processing (NLP). This paper aims to compare the perfor- mance of a general-purpose annotation scheme - Text2Story, based on the ISO 24617-1 stan- dard - with that of a domain-specific scheme - i2b2 - in the context of clinical narrative annotation; and to assess the feasibility of har- monizing ISO 24617-1, originally designed for general-domain applications, with a special- ized extension tailored to the medical domain. Based on the results of this comparative analy- sis, we present Med2Story, a medical-specific extension of ISO 24617-1 developed to address the particularities of clinical text annotation.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171561 2026-03-16T07:41:36Z 2025-01-01T00:00:00Z

Title: Visual representations of temporal relations between events and time expressions in news stories Abstract: High-quality annotation is essential for the ef- fective predictions of machine learning mod- els. When annotations are dense, achieving accurate human labeling can be challenging since the most used annotation tools present an overloaded visualization of labels. Thus, we present Vitra (Visualizer of temporal relation annotations), a tool designed for viewing anno- tations made in corpora, specifically focusing on the temporal relations between events and temporal expressions. This tool aims to fill a gap in the available resources for this purpose. Our focus is on narrative text, which is a rich source for these types of elements. Vitra was developed to increase the human capacity for detecting annotation errors and uncover rela- tions between narrative components or issues about the annotation scheme. To show how this can be done, we present an analysis of a subset of the Text2Story Lusa corpus, a dataset of Portuguese news stories. Such analysis fo- cuses on the linguistic properties of the events and temporal expressions that occur in the an- notated texts, in particular, of short news. We highlight that annotation is an iterative process that involves multiple rounds of revision, and our tool facilitates this process by helping users detect inconsistencies and improve the annota- tion scheme, thus offering added value to the community

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172579 2026-03-16T07:43:11Z 2025-01-01T00:00:00Z

Title: From the image of the molecule to the molecule of the Image: exploring Dianne Iverglynne's methods with a Online platform for the production of ecological images

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172558 2026-03-16T07:41:33Z 2025-01-01T00:00:00Z

Title: Decyclopeding: provoking the error in arts education through a wikimedia-based platform Abstract: he Wikimedia ecosystem has long been integrated into education, both indirectly through platforms like Wikipedia and more directly in classroom settings. However, its evolution-encompassing Wikibase extensions and semantically structured data-introduces new possibilities for educational and research methodologies. This research, conducted as part of the DARIAH project Breaking the Code: Algorithmic Non-Normativity in Creative Digital Humanities, critically examines how Wikimedia-based tools can embody situated knowledge (Haraway, 1988) by integrating intersectional and environmental perspectives into arts education. Our approach embraces error, unpredictability, and misuse as epistemological and pedagogical strategies. We present a prototype platform developed within the Arts Education, Technology, and Society (AETS) course, designed to support pedagogical resources in four key areas: conceptual frameworks for intersectional perspectives, instructional decentralization, sustainability in media production, and speculative approaches to taxonomies. Methodologically, the project employs an alternative epistemic framework where error is embraced as a condition of knowledge production. While error is often assimilated into creative processes in digital and visual arts, its application to wiki platforms-structured by taxonomic logic and encyclopedic norms-reveals flaws in knowledge organization. Our platform actively integrates error as both a methodological and pedagogical tool, challenging epistemic normativity and inviting students to explore non-standard classifications. From a technological standpoint, the platform employs a semantic data structure that deliberately introduces ambiguity. While aligned with Wikidata standards, it simultaneously cultivates an alternative knowledge vocabulary. This is achieved through unconventional taxonomic applications, such as incorporating kinship structures into coin classifications or designing fictional taxonomies inspired by Jorge L. Borges. Additionally, the platform critiques imaging technologies' environmental and social impact, advocating for sustainable media production and shifting from data to capta (Drucker, 2011), emphasizing narrative construction over presumed objectivity. The pedagogical approach mirrors the platform's epistemic critique by incorporating errance, humor, and play as strategies to destabilize rigid knowledge structures. Through interactive storytelling and visualization tools-including cartographic, chronological, and graphical representations-students engage in a participatory process that situates their lived experiences within alternative knowledge frameworks. This methodology transforms students from passive recipients into active agents of knowledge production. Furthermore, the platform fosters a critical engagement with digital infrastructures, emphasizing their role in shaping knowledge validation and classification. The project challenges epistemic normativity in digital environments by introducing speculative taxonomies and deliberate misclassifications. Expected student learning outcomes include developing critical reflexivity, problematizing taxonomic structures, acquiring practical digital methodologies, and exploring experimental pedagogical approaches. Ultimately, this project reimagines digital platforms as spaces for artistic intervention and epistemic critique, expanding their function beyond encyclopedic knowledge production.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172479 2026-03-15T07:47:33Z 2025-01-01T00:00:00Z

Title: Some remarks on last wills at Guimarães (Portugal) in the 13th and 14th Centuries Abstract: Historians are well aware of the significant role that ecclesiastical institutions played in the Middle Ages as beneficiaries and depositories of wills. Portugal is no exception to this observation, and as a result, there are numerous documents in Portuguese archives related to the last wills of individuals who wanted to regulate their earthly possessions before their death. This type of document is created voluntarily by a testator and reflects his/her beliefs and worries about the end of his/her life. It also demonstrates the desire to entrust what mattered most during his/her life, including assets and other belongings, to the "right hands" and to assure the support of specific people connected (as family or dependent) to the testator. The Collegiate Church of Santa Maria de Guimarães is one of Portugal's most significant religious institutions. Consequently, it has been the beneficiary ofmany last wills and testamentary dispositionsmade by the town's inhabitants. The diversity of these dispositions provides an opportunity for a case study, which aims to enhance our understanding of the drafting process of wills during themedieval period. This study focuses not only on the testators but also on some strictly diplomatics aspects, particularly those related to testamentary clauses.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172478 2026-03-15T07:47:33Z 2025-01-01T00:00:00Z

Title: The formularies of medieval portuguese testaments through the study of a Monastic Archive Abstract: This paper focuses on testaments, donations, and postmortem donations preserved in the archive of the monastery of S. Pedro de Pedroso (a Benedictine coenobium of the Porto diocese) in the 13th century. To this end, we collected and analysed the documentation of the previously established typologies belonging to thismonastery, which are currently kept in the National Archive of Torre do Tombo. The first goal is to reflect on the number of documents found for each typology and relate it to their chronological distribution. Then, to analyse the formularies of all these different types of documents (namely the non-essential formulas, such as invocation, notification, sanction, corroboration, and dating) and their validation forms (concretely the notary's subscription, the testes, and the witnesses). In addition to this analysis, the vocables, verbal forms, and /or expressions that state the nature of the legal act were gathered, and a list was made of some of the conditions defined in the text, also in order to compare testaments to donations and to postmortem donations. Through this case study, we aim to identify elements that can contribute to a broader study of Portuguesemedieval testaments and other documents of last wills kept in the archives of ecclesiastical institutions, with a view to a long term study which remains to be completed.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172384 2026-01-22T07:29:16Z 2025-01-01T00:00:00Z

Title: ENEOLI Wikibase: a collaborative working platform for the European Network on Lexical Innovation Abstract: This paper presents the conceptual basis and practical implementation of a Wikibase instance as framework for collaborative workflows, developed as part of the COST action European Network on Lexical Innovation (CA22126). The platform supports the joint creation and publication of three interlinked resources, all integrated into a single knowledge graph: (1) NeoCorpus, a curated repository of bibliographic records and full text collection of the field of Neology; (2) NeoVoc, a multilingual metaterminological vocabulary designed to model, structure, and harmonise the conceptual and terminological frameworks used in research on neology and lexical innovation; and (3) a collection of neologisms, which documents and describes emergent lexical items across various European languages. Wikibase, a free and open-source software platform, provides a generic and flexible environment for collaboratively editing and publishing multilingual, cross-domain descriptions of concepts and lexical units as FAIR Linked Data. The paper explains how resources developed by different ENEOLI Working Groups are made interoperable and linked to each other on the Wikibase instance, describing workflows and preliminary results. We believe that our choice for Wikibase as infrastructure solution for a collaborative Knowledge Graph, here for the domain of Lexical Innovation, may serve as reference for similar endeavours.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172280 2026-01-20T07:39:47Z 2025-01-01T00:00:00Z

Title: MOZ-Smishing: a benchmark dataset for detecting mobile money frauds Abstract: Despite the increasing prevalence of smishing attacks targeting Mobile Money Transfer systems, there is a notable lack of publicly available SMS phishing datasets in this domain. This study seeks to address this gap by creating a specialized dataset designed to detect smishing attacks aimed at Mobile Money Transfer users. The data set consists of crowd-sourced text messages from Mozambican mobile users, meticulously annotated into two categories: legitimate messages (ham) and fraudulent smishing attempts (spam). The messages are written in Portuguese, often incorporating microtext styles and linguistic nuances unique to the Mozambican context.We also investigate the effectiveness of LLMs in detecting smishing. Using in-context learning approaches, we evaluate the models' ability to identify smishing attempts without requiring extensive task-specific training. The data set is released under an open license at the following link: huggingface-Anonymous.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/172065 2026-01-10T07:34:18Z 2025-01-01T00:00:00Z

Title: Evidências de neotectônica na formação Boa Vista e suas implicações na dinâmica do gráben do Tacutu, centro-norte de Roraima Description: Apesar da reconhecida estabilidade do Cráton Amazônico, evidências recentes de atividade neotectônica têm sido objeto de estudo, especialmente em sua porção norte. Diante do crescente interesse geológico na exploração de hidrocarbonetos na Bacia do Tacutu e no litoral da Guiana, este estudo investigou variações topográficas sutis na Formação Boa Vista, localizada no centro-norte de Roraima. O objetivo foi correlacionar essas variações com a tectônica da bacia sedimentar do Tacutu. A metodologia empregada envolveu a fotointerpretação geológica em uma base cartográfica, seguindo um método lógico-sistemático. Foi utilizado um Modelo Digital de Elevação (MDE) de 10x10 metros de resolução espacial, obtido por interferometria SAR, para classificar intervalos altimétricos por meio da análise estatística do desvio padrão. Os procedimentos metodológicos identificaram um intenso controle estrutural da rede de drenagem e feições geomorfológicas como o alinhamento de corpos lacustres e descontinuidades topográficas sutis com orientação preferencial NW-SE. Essas feições estruturais e geomorfológicas observadas sugerem a reativação de antigas estruturas geológicas na porção norte do Cráton Amazônico. Tais processos podem estar relacionados à evolução tectônica da Bacia do Tacutu, bem como à formação de reservatórios de hidrocarbonetos e depósitos minerais. Dessa forma, os resultados são relevantes para a exploração econômica e para a avaliação de riscos geológicos na região.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171990 2026-01-10T07:31:11Z 2025-01-01T00:00:00Z

Title: Os valores da geodiversidade da Serra do Tepequém - RR e sua relevância para o geoturismo Description: A serra do Tepequém, localizada na porção norte do estado de Roraima, destaca-se no contexto da Amazônia setentrional pela relevância de sua beleza cênica e de seu histórico de ocupação, ligado à atividade garimpeira aurífera e diamantífera e pela conjuntura turística. As trilhas usadas pelos garimpeiros entre as décadas de 1930 e 1990 são hoje os principais roteiros turísticos, levando a cachoeiras, morros, cavernas e outros atrativos naturais. Sua geodiversidade, composta por formações rochosas, sistemas lacustres e feições erosivas, cria uma paisagem única, ideal para o geoturismo. Este estudo buscou identificar os valores científicos, culturais, estéticos e econômicos da geodiversidade da serra, correlacionando-os com o desenvolvimento do geoturismo. A metodologia incluiu revisão bibliográfica, análise de mapas, imagens e dados de campo. Identificou-se o valor intrínseco na preservação ambiental promovida por ex-garimpeiros, o valor cultural em lendas locais, como a do "Vulcão do Deus do Fogo" e em artefatos indígenas, o valor estético da paisagem exuberante que atrai muitos visitantes. O valor científico é evidenciado por pesquisas em geociências, desde iniciação científica até pós-doutorado. Já o valor econômico está ligado ao turismo, gerando renda para pousadas, restaurantes e guias, além do artesanato em pedra-sabão e da venda de diamantes artesanais. O valor funcional refere-se à infraestrutura turística e aos esforços para equilibrar conservação e uso público, um desafio constante. Conclui-se que a geodiversidade da serra do Tepequém possui grande potencial para o geoturismo, integrando conservação, cultura e desenvolvimento local. A gestão sustentável desses recursos é essencial para preservar sua relevância ambiental e econômica

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171917 2026-01-10T07:27:10Z 2025-01-01T00:00:00Z

Title: Cartografia geomorfológica da Serra do Tepequém, Amazónia setentrional: um estudo das formas erosivas e deposicionais associadas a dinâmica antrópica Description: A paisagem da Serra do Tepequém, no norte da Amazônia brasileira, exibe uma geomorfologia singular, onde feições erosivas e deposicionais revelam a história das intervenções antrópicas, a exemplo do garimpo de diamantes, atividade que perdurou de 1937 até a década de 1990. A pesquisa direcionou-se à cartografia geomorfológica detalhada das formas de relevo e padrões de canais alteradas pelo garimpo diamantífero histórico. A análise geomorfológica focou-se em três áreas com intensa atividade de mineração: as planícies e fundos de vale dos rios Cabo Sobral, Barata e Paiva. A metodologia combinou análise de imagens do satélite Sentinel 2, imagens e modelos digitais de terreno obtidos com drone, trabalhos de campo e técnicas de geoprocessamento. As análises revelaram a predominância de leito aluvial assoreado nos três rios, com variações na composição dos sedimentos. As análises demonstram que nas três áreas foco, predomina leito aluvial fortemente assoreado, essencialmente arenoso, para o rio Paiva e de forma intercalada (arenosos com trechos rochosos) para os rios Cabo Sobral e Barata. Nas áreas marginais, predomina a planície de inundação, com presença de barras laterais, para a margem esquerda do rio Cabo Sobral. Enquanto extensas coberturas arenosas, são destacadas nas margens do rio Paiva, indicando subtração da mata ciliar e revolvimento superficial. Destacam-se nesse ambiente, pequenas lagoas marginais com áreas variando de 485,71 a 3.362,54 m2, resultantes das escavações do garimpo. As feições erosivas lineares (ravinas e voçorocas) concentram-se nas planícies e fundo de vales e são de grandes dimensões, alcançam áreas de até 44.143,8 m² e um volume erodido de 106.026,80 m³. Indícios da ação antrópica, como desmonte dos taludes, presença de cavidades marginais para acúmulo da água, além de empilhamento de seixos (conhecidos localmente por suruca), foram observados nas feições erosivas. Os resultados evidenciaram a influência antrópica nos processos de agradação e degradação da paisagem.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171899 2026-01-10T07:26:22Z 2025-01-01T00:00:00Z

Title: Análise espacial quantitativa da geodiversidade de Roraima: uma abordagem multimetodológica com densidade kernel Description: A geodiversidade, essencial para a gestão dos recursos naturais, tem sido avaliada de forma predominantemente qualitativa na Amazônia brasileira, incluindo Roraima. A complexa composição abiótica roraimense, contudo, exige métodos quantitativos para uma avaliação precisa. Este estudo objetivou quantificar a geodiversidade de Roraima utilizando um índice abrangente, integrando dados geológicos, geomorfológicos, pedológicos e hidrológicos. Para isso, a metodologia empregou uma função de densidade de Kernel, permitindo estimar a densidade de pontos representativos da geodiversidade no território estadual. A partir da elaboração de mapa integrado que representa a sobreposição espacial de diferentes atributos geológicos, geomorfológicos, pedológicos e da hidrografia. A análise da densidade Kernel revelou um mapa de intensidade de geodiversidade para Roraima, com variações sutis entre os modelos de reclassificação. Observou-se uma vasta distribuição de áreas com alta intensidade de geodiversidade por todo o estado, reflexo da diversidade paisagística e da configuração geotectônica do cráton Amazônico. As regiões setentrionais, influenciadas pelo Cinturão Ígneo Orocaima, e as áreas centrais, marcadas por formas de relevo residuais e pela bacia hidrográfica do rio Cauamé, apresentaram elevados índices. Trechos ao longo do Rio Branco também exibiram alta geodiversidade, atribuída às feições geomorfológicas fluviais. Essas zonas de alta intensidade indicam significativo potencial para a identificação de geossítios, como a Pedra Pintada, já reconhecida pelo GEOSSIT, e a região do graben do Tacutu, que se destaca pela sua relevância para o geopatrimônio. O estudo permitiu a distinção de padrões espaciais da geodiversidade em Roraima, com áreas de alta intensidade ligadas à complexa configuração geotectônica do cráton Amazônico, especialmente no cinturão Orocaima, graben do Tacutu e Rio Branco. Os resultados fornecem uma base sólida para a identificação de geossítios e futuras pesquisas em geopatrimônio, destacando o potencial do estado para conservação e planejamento territorial.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171882 2026-01-10T07:25:38Z 2025-01-01T00:00:00Z

Title: Urban growth in metropolitan regions using dynamic modeling by cellular automata: a comparative analysis between Brazil and Portugal Abstract: This study synthesizes the outcomes of land use changes obtained through the implementation of dynamic modeling by cellular automata across two metropolitan regions in Portugal and Brazil. The purpose is to analyze the primary findings acquired, considering the particularities of each nation and evaluate the potentialities of the used data. The study examined the metropolitan regions of MRRJ (Rio de Janeiro, Brazil) and AMP (Porto, Portugal). Modifications were implemented in the DinamicaEgo software to the fundamental data representing static and dynamic variables for each context. The findings revealed a substantial increase of urban areas in the MRRJ, and the modeling demonstrated its applicability across the two contexts, considering the requisite modifications for the data accessible in each country.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171521 2026-01-04T07:31:50Z 2025-01-01T00:00:00Z

Title: PolyNarrative: a multilingual, multilabel, multi-domain dataset for narrative extraction from news articles Abstract: We present PolyNarrative, a new multilingual dataset of news articles, annotated for narra- tives. Narratives are overt or implicit claims, recurring across articles and languages, promot- ing a specific interpretation or viewpoint on an ongoing topic, often propagating mis/disinfor- mation. We developed two-level taxonomies with coarse- and fine-grained narrative labels for two domains: (i) climate change and (ii) the military conflict between Ukraine and Russia. We collected news articles in four languages (Bulgarian, English, Portuguese, and Russian) related to the two domains and manually anno- tated them at the paragraph level. We make the dataset publicly available, along with experi- mental results of several strong baselines that assign narrative labels to news articles at the paragraph or the document level. We believe that this dataset will foster research in narrative detection and enable new research directions to- wards more multi-domain and highly granular narrative related tasks.

2025-01-01T00:00:00Z https://hdl.handle.net/10216/171599 2026-01-04T07:31:51Z 2025-01-01T00:00:00Z

Title: SemEval 2025 task 10: multilingual characterization and extraction of narratives from online news Abstract: We introduce SemEval-2025 Task 10 on Multilingual Characterization and Extraction of Narratives from Online News, which focuses on the identification and analysis of narratives in online news media. The task is structured into three subtasks: (1) Entity Framing, to identify the roles that relevant entities play within narratives, (2) Narrative Classification, to assign documents fine-grained narratives according to a given, topic-specific taxonomy of narrative labels, and (3) Narrative Extraction, to provide a justification for the dominant narrative of the document. To this end, we analyze news articles across two critical domains, Ukraine-Russia War and Climate Change, in five languages: Bulgarian, English, Hindi, Portuguese, and Russian. This task introduces a novel multilingual and multifaceted framework for studying how online news media construct and disseminate manipulative narratives. By addressing these challenges, our work contributes to the broader effort of detecting, understanding, and mitigating the spread of propaganda and disinformation. The task attracted a lot of interest: 310 teams registered, with 66 submitting official results on the test set.

2025-01-01T00:00:00Z