Jeremy Barnes, Naiara Perez, Alba Bonet-Jover
et al.
Studies on evaluation metrics and LLM-as-a-Judge models for automatic text summarization have largely been focused on English, limiting our understanding of their effectiveness in other languages. Through our new dataset BASSE (BAsque and Spanish Summarization Evaluation), we address this situation by collecting human judgments on 2,040 abstractive summaries in Basque and Spanish, generated either manually or by five LLMs with four different prompts. For each summary, annotators evaluated five criteria on a 5-point Likert scale: coherence, consistency, fluency, relevance, and 5W1H. We use these data to reevaluate traditional automatic metrics used for evaluating summaries, as well as several LLM-as-a-Judge models that show strong performance on this task in English. Our results show that currently proprietary judge LLMs have the highest correlation with human judgments, followed by criteria-specific automatic metrics, while open-sourced judge LLMs perform poorly. We release BASSE and our code publicly, along with the first large-scale Basque summarization dataset containing 22,525 news articles with their subheads.
Akhilesh Kakolu Ramarao, Kevin Tang, Dinah Baer-Henney
Do transformer models generalize morphological patterns like humans do? We investigate this by directly comparing transformers to human behavioral data on Spanish irregular morphomic patterns from \citet{Nevins2015TheRA}. We adopt the same analytical framework as the original human study. Under controlled input conditions, we evaluate whether transformer models can replicate human-like sensitivity to the morphome, a complex linguistic phenomenon. Our experiments focus on three frequency conditions: natural, low-frequency, and high-frequency distributions of verbs exhibiting irregular morphomic patterns. Transformer models achieve higher stem-accuracy than human participants. However, response preferences diverge: humans consistently favor the "natural" inflection across all items, whereas models preferred the irregular forms, and their choices are modulated by the proportion of irregular verbs present during training. Moreover, models trained on the natural and low-frequency distributions, but not the high-frequency distribution, exhibit sensitivity to phonological similarity between test items and Spanish L-shaped verbs, mirroring a limited aspect of human phonological generalization.
Pedro Azevedo, Emanuella Araújo, Gabriel Pierre
et al.
Current datasets for vehicular applications are mostly collected in North America or Europe. Models trained or evaluated on these datasets might suffer from geographical bias when deployed in other regions. Specifically, for scene classification, a highway in a Latin American country differs drastically from an Autobahn, for example, both in design and maintenance levels. We propose VWise, a novel benchmark for road-type classification and scene classification tasks, in addition to tasks focused on external contexts related to vehicular applications in LatAm. We collected over 520 video clips covering diverse urban and rural environments across Latin American countries, annotated with six classes of road types. We also evaluated several state-of-the-art classification models in baseline experiments, obtaining over 84% accuracy. With this dataset, we aim to enhance research on vehicular tasks in Latin America.
ChatGPT is a popular information system (over 1 billion visits in August 2023) that can generate natural language responses to user queries. It is important to study the quality and equity of its responses on health-related topics, such as vaccination, as they may influence public health decision-making. We use the Vaccine Hesitancy Scale (VHS) proposed by Shapiro et al.1 to measure the hesitancy of ChatGPT responses in English, Spanish, and French. We find that: (a) ChatGPT responses indicate less hesitancy than those reported for human respondents in past literature; (b) ChatGPT responses vary significantly across languages, with English responses being the most hesitant on average and Spanish being the least; (c) ChatGPT responses are largely consistent across different model parameters but show some variations across the scale factors (vaccine competency, risk). Results have implications for researchers interested in evaluating and improving the quality and equity of health-related web information.
The Norman Conquest of 1066 C.E. brought profound transformations to England's administrative, societal, and linguistic practices. The DEEDS (Documents of Early England Data Set) database offers a unique opportunity to explore these changes by examining shifts in word meanings within a vast collection of Medieval Latin charters. While computational linguistics typically relies on vector representations of words like static and contextual embeddings to analyze semantic changes, existing embeddings for scarce and historical Medieval Latin are limited and may not be well-suited for this task. This paper presents the first computational analysis of semantic change pre- and post-Norman Conquest and the first systematic comparison of static and contextual embeddings in a scarce historical data set. Our findings confirm that, consistent with existing studies, contextual embeddings outperform static word embeddings in capturing semantic change within a scarce historical corpus.
J. García-Díaz, Mar Cánovas-García, R. Colomo-Palacios
et al.
Abstract Online social networks allow powerless people to gain enormous amounts of control over particular people’s lives and profit from the anonymity or social distance that the Internet provides in order to harass other people. One of the most frequently targeted groups comprise women, as misogyny is, unfortunately, a reality in our society. However, although great efforts have recently been made to identify misogyny, it is still difficult to distinguish as it can sometimes be very subtle and deep, signifying that the use of statistical approaches is not sufficient. Moreover, as Spanish is spoken worldwide, context and cultural differences can complicate this identification. Our contribution to the detection of misogyny in Spanish is two-fold. On the one hand, we apply Sentiment Analysis and Social Computing technologies for detecting misogynous messages in Twitter. On the other, we have compiled the Spanish MisoCorpus-2020, a balanced corpus regarding misogyny in Spanish, and classified it into three subsets concerning (1) violence towards relevant women, (2) messages harassing women in Spanish from Spain and Spanish from Latin America, and (3) general traits related to misogyny. Our proposal combines a classification based on average word embeddings and linguistic features in order to understand which linguistic phenomena principally contribute to the identification of misogyny. We have evaluated our proposal with three machine-learning classifiers, achieving the best accuracy of 85.175%. Finally the proposed approach is also validated with existing corpora for misogyny and aggressiveness detection such as AMI and HatEval obtaining good results
Las investigaciones sobre historia de las juventudes conforman un campo reciente dentro de la historiografía argentina. Si bien de modo predominante estos trabajos se enfocaron en las culturas juveniles a partir de la década de 1960, recientemente se ha comenzado a indagar respecto de las prácticas y representaciones sobre las juventudes en períodos anteriores. Con el objetivo de analizar estas últimas producciones, en el artículo se identifican distintas aproximaciones temáticas y metodológicas del estudio de la historia de las juventudes en Argentina de la primera mitad del siglo XX. De modo complementario, en correspondencia con los procesos que han sido señalados para el ámbito europeo y norteamericano sobre la construcción de las juventudes desde finales del siglo XIX, el trabajo plantea interrogantes y posibles temas de análisis que contribuyan a pensar las juventudes en clave local y transnacional.
O artigo analisa a política externa dos governos militares de Médici, Geisel e Figueiredo (1969-85), quando o regime adota uma postura pragmática e autonomista, em lugar da ênfase ideológica. Os governos Castelo Branco e Costa e Silva (1964-69) vivenciaram a dialética ruptura x continuidade, com a frustração do alinhamento ideológico com os EUA. Uma vez adotada uma política econômica desenvolvimentista, com forte apoio do Estado, a busca de autonomia diplomática se afirmou. Se o governo Geisel constitui o paradigma de tal estratégia, observa-se que ela teve início no governo anterior e, apesar das dificuldades, prosseguiu até 1985.
Eliana Gómez-Rodríguez, Jorge Márquez-Valderrama, Victoria Estrada-Orrego
Este artículo aborda las prácticas, así como los discursos médicos y administrativos especializados en los niños y niñas abandonados y huérfanos, atendidos por la sección Amas de Oriente del Hospicio de Bogotá, entre 1918 y 1936. Mediante la crítica documental, se analizó un corpus elaborado a partir del Archivo de la Beneficencia de Cundinamarca. La documentación permitió una aproximación al funcionamiento de la asistencia pública en general y, específicamente, al del Hospicio. Además, facilitó el estudio de las estrategias de atención, vigilancia y control dirigidas a la primera infancia y a las amas de campo. Esto último contribuyó al hallazgo de un sistema oficial de nodrizas para la infancia abandonada. Para comprenderlo, se reconstruyó el contexto histórico de la pobreza, la desigualdad, el trabajo femenino, la nutrición y la economía de la salud. Los niños más pequeños fueron atendidos por nodrizas asalariadas, mujeres campesinas mal pagas, utilizadas por las autoridades, tildadas de ignorantes, sucias y enfermas. Fue un sistema de crianza duradero, quizás porque era más viable y menos oneroso que el modelo de atención intramural del Hospicio de Bogotá. Paradójicamente, este sistema creado como institución especializada en la lactancia delegada, terminó convirtiéndose en una institución de maternidad sustituta.
A partial Latin square of order $n$ can be represented by a $3$-dimensional chess-board of size $n\times n\times n$ with at most $n^2$ non-attacking rooks. Based on this representation, we apply a uniform method to prove the M. Hall's, Ryser's and Cruse's theorems for completion of partial Latin squares. With the help of this proof, we extend the scope of Cruse's theorem to compact bricks, which appear to be independent of their environment. Without losing any completion you can replace a dot by a rook if the dot must become rook, or you can eliminate the dots that are known not to become rooks. Therefore, we introduce primary and secondary extension procedures that are repeated as many times as possible. If the procedures do not decide whether a PLSC can be completed or not, a new necessary condition for completion can be formulated for the dot structure of the resulting PLSC, the BUG condition.
Multi-Angled Parallelism (MAP) is a method to recognize lines in binary images. It is suitable to be implemented in parallel processing and image processing hardware. The binary image is transformed into directional planes, upon which, directional operators of erosion-dilation are iteratively applyed. From a set of basic operators, more complex ones are created, which let to extract the several types of lines. Each type is extracted with a different set of operations and so the lines are identified when extracted. In this paper, an overview of MAP is made, and it is adapted to line recognition in Spanish topographical maps, with the double purpose of testing the method in a real case and studying the process of adapting it to a custom application.
A. Morlett Paredes, A. Gooding, L. Artiola i Fortuny
et al.
Abstract Objective: The present review paper aimed to identify published neuropsychological test norms developed for Spanish-speakers living in the United States (U.S.). Methods: We conducted a systematic review of the literature via an electronic search on PubMed using keywords “Normative data,” “Neuropsychological test,” “norms”, “Hispanic/Latinos,” “Spanish Speakers,” and “United States.” We added other studies and published manuals as identified by citations in papers from the original search. Results Eighteen sources of normative data for Spanish-speakers in the U.S. were identified. Of the 18 citations identified, only four provide normative data on comprehensive batteries of tests for Spanish-Speakers. Two of these are based on persons living in the southwest of the U.S., who tend to be of Mexican origin. Overall, a number of the studies are focused on older persons and although the majority include participants with wide ranges of education, participants in the ends of the education distribution tend to be underrepresented. Conclusion: Here we provide a detailed description of the neuropsychological normative data currently available for Spanish-speakers living in the U.S. While there has been increased attention towards developing norms for neuropsychological batteries in Spanish-speaking countries (e.g., Latin America and Spain), there is still an urgent need to standardize neuropsychological tests among diverse groups of Spanish-speaking adults living in the U.S. The present review presents a list of norms for U.S.-dwelling Spanish-speakers, thus providing an important tool for clinicians and researchers.
This article establishes essential items introduced by the urban actors in Bogota, capital city of Colombia, with the aim of achieving its internationalization. The first part inscribes this theme in the framework of globalization, considers the new functionalities and elements that are at stake, and are also an object of rivalry between cities on the world stage. The second part develops, in a neo-institutional perspective, the role played by public and private urban elites and the strategy they have chosen for the city internationalization and its development.
Relations between Spaniards and Argentines have been good since the beginning of this shared history. There are hundreds of examples and there is the particular situation that originates in the years of the Civil War that shakes the former mother country and, in the immediate aftermath, when intellectuals, politicians and journalists, doctors and lawyers, persecuted or threatened by the system that put an end to the Second Republic, to which they had adhered, makes their private and public life difficult. From this immense group, we selected a few, privileging those who worked in our environment, who, through the university chairs to which they accessed, professed a marked influence on young people and adults, since they allowed a ferment of renovating proposals. By offering to participate in teaching, one of the most delicate aspects of the exile was covered : job placement. For their part, they helped expand the contents of the programs, the approaches, the bibliography and stimulate a comparative criterion.
In this work we present a new small data-set in Computational Creativity (CC) field, the Spanish Literary Sentences for emotions detection corpus (LISSS). We address this corpus of literary sentences in order to evaluate or design algorithms of emotions classification and detection. We have constitute this corpus by manually classifying the sentences in a set of emotions: Love, Fear, Happiness, Anger and Sadness/Pain. We also present some baseline classification algorithms applied on our corpus. The LISSS corpus will be available to the community as a free resource to evaluate or create CC-like algorithms.
OBJECTIVE To identify and analyze available literature on care transition strategies in Latin American countries. METHODS Integrative literature review that included studies indexed in PubMed, LILACS, Web of Science Core Collection, CINAHL, SCOPUS databases, and the Scientific Electronic Library Online (SciELO), published in Portuguese, Spanish or English, between 2010 and 2017. RESULTS Eleven articles were selected and the strategies were grouped into components of care transition: discharge planning, advanced care planning, patient education and promotion of self-management, medication safety, complete communication of information, and outpatient follow-up. These strategies were carried out by multidisciplinary team members, in which nurses play a leading role in promoting safe care transitions. CONCLUSIONS Care transition activities are generally initiated very close to patient discharge, this differs from recommendations of care transition programs and models, which suggest implementing care transition strategies from the time of admission until discharge.
With the beginning of the 21st century, there has been an acceleration of migratory fl ows from Latin America and the Caribbean (LAC) to Europe. As a result, and despite the negative impact of the economic crisis, 4.6 million Latin American and Caribbean immigrants reside in Europe, half of them in Spain. This article analyses the recent evolution of these migratory fl ows, their territorial distribution, and their demographic pro fi les according to the 2011 European census data disseminated by a new tool — the Census Hub — implemented by the European Statistical System. The analysis shows the existence of a high LAC immigrant concentration in Spain and in certain European cities, a marked young and feminized demographic pro fi le, a great variety of educational levels and a different insertion in each European labour market, although many LAC immigrants work in low-skill occupations, being overquali fi ed and underemployed in most of the countries.