Hasil untuk "Latin America. Spanish America"

Eldi Paola Robayo

The export of security-related know-how has become a strategic component of Colombia’s foreign policy over the past two decades. This article aims to describe and analyze the role of the National Police of Colombia in this export between 2010 and 2022. Using a mixed methodological approach that combines qualitative analysis and the systematization of quantitative data, it examines the institutional guidelines underpinning this policy, the volume and characteristics of the training programs offered, and the deployment of experts abroad. Additionally, the cases of Mexico and Honduras are analyzed in depth as representative cooperation experiences. The article highlights the functioning of the triangular cooperation model, in which Colombia provides training while the United States contributes funding and logistical support. The analysis identifies the implications of this strategy, including its alignment with a militarized security approach, the limited evaluation of its actual impacts, and the risks of replicating problematic practices in institutionally fragile contexts.

Johan R. Portela, Nicolás Perez, Rubén Manrique

Natural Language Inference (NLI), also known as Recognizing Textual Entailment (RTE), serves as a crucial area within the domain of Natural Language Processing (NLP). This area fundamentally empowers machines to discern semantic relationships between assorted sections of text. Even though considerable work has been executed for the English language, it has been observed that efforts for the Spanish language are relatively sparse. Keeping this in view, this paper focuses on generating a multi-genre Spanish dataset for NLI, ESNLIR, particularly accounting for causal Relationships. A preliminary baseline has been conceptualized and subjected to an evaluation, leveraging models drawn from the BERT family. The findings signify that the enrichment of genres essentially contributes to the enrichment of the model's capability to generalize. The code, notebooks and whole datasets for this experiments is available at: https://zenodo.org/records/15002575. If you are interested only in the dataset you can find it here: https://zenodo.org/records/15002371.

Cecilia Mafla

Archivo histórico de Kipus: Revista Andina de Letras y Estudios Culturales, 2005.

Stefan Bott, Horacio Saggion, Nelson Peréz Rojas et al.

Automatic lexical simplification is a task to substitute lexical items that may be unfamiliar and difficult to understand with easier and more common words. This paper presents the description and analysis of two novel datasets for lexical simplification in Spanish and Catalan. This dataset represents the first of its kind in Catalan and a substantial addition to the sparse data on automatic lexical simplification which is available for Spanish. Specifically, it is the first dataset for Spanish which includes scalar ratings of the understanding difficulty of lexical items. In addition, we present a detailed analysis aiming at assessing the appropriateness and ethical dimensions of the data for the lexical simplification task.

Krzysztof Nowak, Jędrzej Ziębura, Krzysztof Wróbel et al.

This study introduces the eFontes models for automatic linguistic annotation of Medieval Latin texts, focusing on lemmatization, part-of-speech tagging, and morphological feature determination. Using the Transformers library, these models were trained on Universal Dependencies (UD) corpora and the newly developed eFontes corpus of Polish Medieval Latin. The research evaluates the models' performance, addressing challenges such as orthographic variations and the integration of Latinized vernacular terms. The models achieved high accuracy rates: lemmatization at 92.60%, part-of-speech tagging at 83.29%, and morphological feature determination at 88.57%. The findings underscore the importance of high-quality annotated corpora and propose future enhancements, including extending the models to Named Entity Recognition.

David Gimeno-Gómez, Carlos-D. Martínez-Hinarejos

Speech is considered as a multi-modal process where hearing and vision are two fundamentals pillars. In fact, several studies have demonstrated that the robustness of Automatic Speech Recognition systems can be improved when audio and visual cues are combined to represent the nature of speech. In addition, Visual Speech Recognition, an open research problem whose purpose is to interpret speech by reading the lips of the speaker, has been a focus of interest in the last decades. Nevertheless, in order to estimate these systems in the currently Deep Learning era, large-scale databases are required. On the other hand, while most of these databases are dedicated to English, other languages lack sufficient resources. Thus, this paper presents a semi-automatically annotated audiovisual database to deal with unconstrained natural Spanish, providing 13 hours of data extracted from Spanish television. Furthermore, baseline results for both speaker-dependent and speaker-independent scenarios are reported using Hidden Markov Models, a traditional paradigm that has been widely used in the field of Speech Technologies.

Frederick Riemenschneider, Anette Frank

Intertextual allusions hold a pivotal role in Classical Philology, with Latin authors frequently referencing Ancient Greek texts. Until now, the automatic identification of these intertextual references has been constrained to monolingual approaches, seeking parallels solely within Latin or Greek texts. In this study, we introduce SPhilBERTa, a trilingual Sentence-RoBERTa model tailored for Classical Philology, which excels at cross-lingual semantic comprehension and identification of identical sentences across Ancient Greek, Latin, and English. We generate new training data by automatically translating English texts into Ancient Greek. Further, we present a case study, demonstrating SPhilBERTa's capability to facilitate automated detection of intertextual parallels. Our models and resources are available at https://github.com/Heidelberg-NLP/ancient-language-models.

Julio Pinto Vallejos

Como homenaje a Jorge Gelman, este artículo resume los resultados de una investigación comparativa sobre la temprana formación de estado en tres países sudamericanos: Argentina, Perú y Chile. Tomando como eje analítico los esfuerzos de recomposición de hegemonía desplegados por tres liderazgos caudillistas (Juan Manuel de Rosas, Ramón Castilla y Diego Portales), se exploran los procesos de negociación, incorporación y exclusión de sectores populares que el artículo denomina “construcción social del estado”. Se concluye que, a partir de una tarea común de restablecimiento del orden, los tres casos estudiados aplicaron estrategias y dosis diferentes de imposición o seducción, derivando en resultados igualmente disímiles en materia de adhesiones plebeyas e inclusión social.

Frank D. Zamora-Reina, Felipe Bravo-Marquez, Dominik Schlechtweg

We present the first shared task on semantic change discovery and detection in Spanish and create the first dataset of Spanish words manually annotated for semantic change using the DURel framework (Schlechtweg et al., 2018). The task is divided in two phases: 1) Graded Change Discovery, and 2) Binary Change Detection. In addition to introducing a new language the main novelty with respect to the previous tasks consists in predicting and evaluating changes for all vocabulary words in the corpus. Six teams participated in phase 1 and seven teams in phase 2 of the shared task, and the best system obtained a Spearman rank correlation of 0.735 for phase 1 and an F1 score of 0.716 for phase 2. We describe the systems developed by the competing teams, highlighting the techniques that were particularly useful and discuss the limits of these approaches.

Adriana Guevara-Rukoz, Isin Demirsahin, Fei He et al.

Michael Janoschka, J. Sequera, Luis Salinas

V. Gongora, Verónica Cruz Licea, M. Mebarak Chams et al.

In order to advance in the study of positive body image among different cultures, it is important to create culturally appropriate measures. We examined the psychometric properties of a Latin-American Spanish translation of the Body Appreciation Scale-2 (BAS-2; Tylka & Wood-Barcalow, 2015a), specifically assessing measurement invariance using a large sample of 3845 male and female adolescents from Argentina, Mexico, and Colombia. Participants completed the BAS-2, Eating Disorder Inventory-2, Male Body Attitude Scale and the Sociocultural Attitudes Toward Appearance Questionnaire-3. The BAS-2 had a unidimensional factor structure in each of the three samples. We confirmed the structural, metric, and scalar invariance of the scale regardless of gender or country. Adolescents in Argentina had lower body appreciation compared with those in Mexico and Colombia. Overall, females had lower body appreciation than males, with the greatest gender difference found in the Argentinean sample. Our findings strongly support the validity and reliability of this Latin-American translation of the BAS-2 in measuring positive body image in adolescents in Argentina, Mexico, and Colombia.

Christine Focquenoy Simonnet

Álvaro Luque Amo

El establecimiento del diario personal como género autobiográfico en el sistema literario posee diferentes periodizaciones de acuerdo al contexto: mientras que en el ámbito francés tiene lugar a finales del siglo xix y primera mitad del xx, en la península ibérica no se produce hasta las últimas décadas del siglo xx. Este trabajo tiene como objetivo analizar el diario personal de tres autores, Rosa Chacel, Miguel Torga y Josep Pla, cada uno de los cuales pertenece a una tradición lingüística del contexto ibérico: la española, la portuguesa y la catalana, respectivamente. El estudio de estas tres obras desde una perspectiva conjunta –influida por los Estudios Ibéricos– permitirá entender el surgimiento de un nuevo género literario en estos ámbitos: el diario literario.

Olivier Goudet, Jin-Kao Hao

The partial Latin square extension problem is to fill as many as possible empty cells of a partially filled Latin square. This problem is a useful model for a wide range of applications in diverse domains. This paper presents the first massively parallel evolutionary algorithm algorithm for this computationally challenging problem based on a transformation of the problem to partial graph coloring. The algorithm features the following original elements. Based on a very large population (with more than $10^4$ individuals) and modern graphical processing units, the algorithm performs many local searches in parallel to ensure an intensive exploitation of the search space. The algorithm employs a dedicated crossover with a specific parent matching strategy to create a large number of diversified and information-preserving offspring at each generation. Extensive experiments on 1800 benchmark instances show a high competitiveness of the algorithm compared to the current best performing methods. Competitive results are also reported on the related Latin square completion problem. Analyses are performed to shed lights on the roles of the main algorithmic components. The code of the algorithm will be made publicly available.

Juan Manuel Pérez, Damián A. Furman, Laura Alonso Alemany et al.

Since BERT appeared, Transformer language models and transfer learning have become state-of-the-art for Natural Language Understanding tasks. Recently, some works geared towards pre-training specially-crafted models for particular domains, such as scientific papers, medical documents, user-generated texts, among others. These domain-specific models have been shown to improve performance significantly in most tasks. However, for languages other than English such models are not widely available. In this work, we present RoBERTuito, a pre-trained language model for user-generated text in Spanish, trained on over 500 million tweets. Experiments on a benchmark of tasks involving user-generated text showed that RoBERTuito outperformed other pre-trained language models in Spanish. In addition to this, our model achieves top results for some English-Spanish tasks of the Linguistic Code-Switching Evaluation benchmark (LinCE) and has also competitive performance against monolingual models in English tasks. To facilitate further research, we make RoBERTuito publicly available at the HuggingFace model hub together with the dataset used to pre-train it.

Nicolás Urbina-Cardona, Mary E. Blair, M. Londoño et al.

Species distribution modeling (SDM) is a booming area of research that has had an exponential increase in use and development in recent years. We performed a search of scientific literature and found 5,533 documents published from 1993 to 2018 using SDM, representing a global network of 4,329 collaborating institutions from 155 countries, with Brazil and Mexico being in the top 10 of the most prolific countries globally. National Autonomous University of Mexico, Chinese Academy of Sciences, University of Kansas, and U.S. Geological Survey are the most prolific institutions worldwide. Latin American institutions (n = 556) participated in 1,000 (18% of global productivity) documents published in collaboration with 591 institutions outside Latin American countries, from which the National Autonomous University of Mexico, Federal University of Goiás, Institute of Ecology A.C., National Scientific and Technical Research Council in Argentina, University of São Paulo, and University of Brasilia were the most productive. From this body of literature, the most frequently modeled taxonomic groups were Chordata and Insecta, and the most common realms of application were conservation planning and management, climate change, species conservation, epidemiology, evolutionary biology, and biological invasions. From the 36 modeling methods identified to generate SDMs, MaxEnt is used in 73.5% of the papers, followed by Genetic Algorithm for Rule-Set Prediction (GARP) with 18.7%, and just 7.4% of the papers compared between 3 and 10 modeling methods. In Latin American countries, productivity in SDM research could be improved as the network of collaborations diversifies and connects with other productive countries (such as United Kingdom, China, Spain, Germany, Australia, and France). The scientific collaboration between Latin American countries should be increased, as the most prolific countries (Brazil, Mexico, Argentina, and Colombia) share less than 10% of its productivity. Some of the main challenges for SDM development in Latin America include bridging the gaps from (a) software use to research productivity and (b) translation to decision-making. To address these challenges, we propose to strengthen communities of practice where modelers, species experts, and decision-makers come together to discuss and develop SDM to shift and enhance current paradigms on how science and decision-making are linked.

B. Nuche-Berenguer, D. Sakellariou

Introduction Cancer incidence and mortality in Latin America are rising. While effective cancer screening services, accessible to the whole population and enabling early cancer detection are needed, existing research shows the existence of disparities in screening uptake in the region. Objective We conducted a systematic review to investigate the socioeconomic determinants for the disparities in the use of breast, cervical and colorectal cancer screening services in Latin America. Methods We searched for studies reporting on socioeconomic determinants impacting on access to breast, cervical and colorectal cancer screening, published from 2009 through 2018. The studies that qualified for inclusion contained original analyses on utilisation of breast, cervical and colorectal cancer screening across socioeconomic levels in Latin America. For each study, paired reviewers performed a quality analysis followed by detailed review and data extraction. Results Twenty-four articles that met the eligibility criteria and were of sufficient quality were included in this review. Thirteen of the included articles were written in English, eight in Portuguese and three in Spanish, and they reported on the use of breast or cervical cancer screening. No studies were found on the socioeconomic determinants regarding the utilisation of colorectal cancer screening in Latin America. Low income, low education level, lack of health insurance and single marital status were all found to be determinants of underuse of breast and cervical cancer screening services. Conclusions Cancer screening programs in the region must prioritize reaching those populations that underuse cancer screening services to ensure equitable access to preventive services. It is important to develop national screening programmes that are accessible to all (including uninsured people) through, for example, the use of mobile units for mammography and self-screening methods.

M. P. Navarro Pérez, M. Marín Gracia, E. Bellosta Diago et al.

INTRODUCTION Migraine is a very disabling disease that has a great impact on patients' quality of life and interferes in their personal, social, work and family spheres. From a historical point of view, the connection between the Iberian Peninsula and Latin America has been very important, and so it seems reasonable to find there are parallels in the epidemiology of this disease, given the role that certain genetic and lifestyle-related determinants have in its natural history. AIM To conduct a detailed review of the descriptive epidemiological studies of migraine in Spain and Latin America. DEVELOPMENT Literature search of epidemiological studies on migraine in our country and in Latin America. The population studied, the methodology, the questionnaire used for diagnosis and the prevalence data were analysed. Altogether 23 studies were evaluated. CONCLUSIONS Not all countries have population-based epidemiological studies of migraine, and most of them were conducted more than 10 years ago. Moreover, a wide range of methodologies were applied. The prevalence data obtained in the selected studies, with the exception of some conducted in Brazil and Peru, are very similar to those found in Spain.

. .

.