Examinar
Envíos recientes
Publicación Acquérir le savoir lexical pour une meilleure compréhension des langues romanes(ATILF/SLR, 2016-01-01) Gómez Fernández, AraceliPublicación Inferencias léxicas y colocaciones en la intercomprensión plurilingüe(Universitat Jaume I, 2011-01-01) Gómez Fernández, AraceliPublicación Las damas nobles en la ciudad en la Edad Moderna: espacios y relatos de vidaLas damas nobles en la ciudad en la Edad Moderna: espacios y relatos de vida(Ediciones Doce Calles/ Universidad Complutense, 2025-01-01) Antorán Pilar, Guiomar; Agencia Estatal de Investigación (España)Publicación An Effectiveness Metric for Ordinal Classification: Formal Properties and Experimental Results(Association for Computational Linguistics Note:, 2020-07-01) Amigo Cabrera, Enrique; Gonzalo Arroyo, Julio Antonio; Mizzarro, Stefano; Carrillo de Albornoz Cuadrado, Jorge AmandoIn Ordinal Classification tasks, items have to be assigned to classes that have a relative ordering, such as positive, neutral, negative in sentiment analysis. Remarkably, the most popular evaluation metrics for ordinal classification tasks either ignore relevant information (for instance, precision/recall on each of the classes ignores their relative ordering) or assume additional information (for instance, Mean Average Error assumes absolute distances between classes). In this paper we propose a new metric for Ordinal Classification, Closeness Evaluation Measure, that is rooted on Measurement Theory and Information Theory. Our theoretical analysis and experimental results over both synthetic data and data from NLP shared tasks indicate that the proposed metric captures quality aspects from different traditional tasks simultaneously. In addition, it generalizes some popular classification (nominal scale) and error minimization (interval scale) metrics, depending on the measurement scale in which it is instantiated.Publicación Pron vs Prompt: Can Large Language Models already Challenge a World-Class Fiction Author at Creative Text Writing?(Association for Computational Linguistics, 2024-11-01) Marco Remón, Guillermo; Gonzalo Arroyo, Julio Antonio; Mateo-Girona, M. Teresa; Castillo Santos, Ramón José del; European CommissionAre LLMs ready to compete in creative writing skills with a top (rather than average) novelist? To provide an initial answer for this question, we have carried out a contest between Patricio Pron (an awarded novelist, considered one of the best of his generation) and GPT-4 (one of the top performing LLMs), in the spirit of AIhuman duels such as DeepBlue vs Kasparov and AlphaGo vs Lee Sidol. We asked Pron and GPT-4 to provide thirty titles each, and then to write short stories for both their titles and their opponent’s. Then, we prepared an evaluation rubric inspired by Boden’s definition of creativity, and we collected several detailed expert assessments of the texts, provided by literature critics and scholars. The results of our experimentation indicate that LLMs are still far from challenging a top human creative writer. We also observed that GPT-4 writes more creatively using Pron’s titles than its own titles (which is an indication of the potential for human-machine co-creation). Additionally, we found that GPT-4 has a more creative writing style in English than in Spanish.Publicación Evaluating Sequence Labeling on the basis of Information Theory(Association for Computational Linguistics, 2025-07-01) Amigo Cabrera, Enrique; Álvarez Mellado, Elena; Carrillo de Albornoz Cuadrado, Jorge Amando; European Commission; Agensia Estatal de Investigación (España)Various metrics exist for evaluating sequence labeling problems (strict span matching, token oriented metrics, token concurrence in sequences, etc.), each of them focusing on certain aspects of the task. In this paper, we define a comprehensive set of formal properties that captures the strengths and weaknesses of the existing metric families and prove that none of them is able to satisfy all properties simultaneously. We argue that it is necessary to measure how much information (correct or noisy) each token in the sequence contributes depending on different aspects such as sequence length, number of tokens annotated by the system, token specificity, etc. On this basis, we introduce the Sequence Labelling Information Contrast Model (SL-ICM), a novel metric based on information theory for evaluating sequence labeling tasks. Our formal analysis and experimentation show that the proposed metric satisfies all properties simultaneously.Publicación Hybrid Neural Network-Based Indoor Localisation System for Mobile Robots Using CSI Data in a Robotics Simulator(Springer, 2025-10-01) Ballesteros Jérez, Javier; Martínez Gómez, Jesús; García Varea, Ismael; Orozco Barbosa, Luis; Castillo-Cara, ManuelWe present a hybrid neural network model for inferring the position of mobile robots using Channel State Information (CSI) data from a Massive MIMO system. By leveraging an existing CSI dataset, our approach integrates a Convolutional Neural Network (CNN) with a Multilayer Perceptron (MLP) to form a Hybrid Neural Network (HyNN) that estimates 2D robot positions. CSI readings are converted into synthetic images using the TINTO tool. The localisation solution is integrated with a robotics simulator, and the Robot Operating System (ROS), which facilitates its evaluation through heterogeneous test cases, and the adoption of state estimators like Kalman filters. Our contributions illustrate the potential of our HyNN model in achieving precise indoor localisation and navigation for mobile robots in complex environments. The study follows, and proposes, a generalisable procedure applicable beyond the specific use case studied, making it adaptable to different scenarios and datasets.Publicación Is anisotropy really the cause of BERT embeddings not being semantic?(Association for Computational Linguistics, 2022-01-01) Fuster Baggetto, Alejandro; Fresno Fernández, Víctor DiegoIn this paper we conduct a set of experiments aimed to improve our understanding of the lack of semantic isometry in BERT, i.e. the lack of correspondence between the embedding and meaning spaces of its contextualized word representations. Our empirical results show that, contrary to popular belief, the anisotropy is not the root cause of the poor performance of these contextual models’ embeddings in semantic tasks. What does affect both the anisotropy and semantic isometry is a set of known biases: frequency, subword, punctuation, and case. For each one of them, we measure its magnitude and the effect of its removal, showing that these biases contribute but do not completely explain the phenomenon of anisotropy and lack of semantic isometry of these contextual language models.Publicación The Reader is the Metric: How Textual Features and Reader Profiles Explain Conflicting Evaluations of AI Creative Writing(Association for Computational Linguistics, 2025-01-01) Marco Remón, Guillermo; Gonzalo Arroyo, Julio Antonio; Fresno Fernández, Víctor DiegoRecent studies comparing AI-generated and human-authored literary texts have produced conflicting results: some suggest AI already surpasses human quality, while others argue it still falls short. We start from the hypothesis that such divergences can be largely explained by genuine differences in how readers interpret and value literature, rather than by an intrinsic quality of the texts evaluated. Using five public datasets (1,471 stories, 101 annotators including critics, students, and lay readers), we (i) extract 17 reference-less textual features (e.g., coherence, emotional variance, average sentence length...); (ii) model individual reader preferences, deriving feature importance vectors that reflect their textual priorities; and (iii) analyze these vectors in a shared “preference space”. Reader vectors cluster into two profiles: _surface-focused readers_ (mainly non-experts), who prioritize readability and textual richness; and _holistic readers_ (mainly experts), who value thematic development, rhetorical variety, and sentiment dynamics. Our results quantitatively explain how measurements of literary quality are a function of how text features align with each reader’s preferences. These findings advocate for reader-sensitive evaluation frameworks in the field of creative text generation.Publicación Robust Estimation of Population-Level Effects in Repeated-Measures NLP Experimental Designs(Association for Computational Linguistics, 2025-01-01) Benito Santos, Alejandro; Ghajari Espinosa, Adrián; Fresno Fernández, Víctor DiegoNLP research frequently grapples with multiple sources of variability—spanning runs, datasets, annotators, and more—yet conventional analysis methods often neglect these hierarchical structures, threatening the reproducibility of findings. To address this gap, we contribute a case study illustrating how linear mixed-effects models (LMMs) can rigorously capture systematic language-dependent differences (i.e., population-level effects) in a population of monolingual and multilingual language models. In the context of a bilingual hate speech detection task, we demonstrate that LMMs can uncover significant population-level effects—even under low-resource (small-N) experimental designs—while mitigating confounds and random noise. By setting out a transparent blueprint for repeated-measures experimentation, we encourage the NLP community to embrace variability as a feature, rather than a nuisance, in order to advance more robust, reproducible, and ultimately trustworthy results.Publicación Four distinctive features in the early dawn of vernacular Spanish Poetics: between Classical influence and the Italian style ruling(2020-09-01) Rodríguez Cachón, IreneDuring most of the 16th century, Spanish Poetics was understood as something difficult to figure it out. Therefore, the first poetic standards that organized and fixed the new way of doing poetry — both in their principles and in their uses - did not appear in the lberian Peninsula until almost the last two decades of the 16th century. However, this fact does not imply that the Spanish poets were not caught up in the new winds of change, which had started early in the same century in Italy. Therefore, some treatises, which seek to explain and systematize a new poetic regulation just all in Spanish, rather than in Latin, begin to appear timidly from 1580 onwards. This article will attempt to propose and summarize four distinctive features that explain why the modern Spanish poetic treatises come so chronologically late regarding the rest of Europe. In addition, it will point out the fundamental aspects of the new modern Spanish Poetics, considering the most significant titles that were published in Spain at the dawn ofthe Early Modern Age.Publicación Las cornisas de estuco del I Estilo del yacimiento de Can Tacó-Turó d’en Roïna (Montmeló-Montornès Del Valles, Barcelona)(Edizioni Quasar, 2024-01-01) Mostalac Carrillo, Antonio; Guiral Pelegrín, Carmen; Rodrigo Requena, EstherThe aim of this study is to present new types of Roman-Republican parietal mouldings found in the province of Barcelona. The new profiles and their chronology, derived from stratigraphic and stylistic criteria, allow us to place them in the second quarter of the 2nd century BC. They are stylistically related to some examples from localities in Campania, in the Italic Peninsula and also from Sicily and North Africa.Publicación Horti picti provenienti da un monumento funerario del sobborgo occidentale di Tarraco (Tarragona, Spagna)(Edizioni Quasar, 2024-01-01) Guiral Pelegrín, Carmen; Íñiguez Berrozpe, Lara; Remolà Vallverdú, Josep AntonThree large pictorial panels have been preserved from ancient excavations carried out in the 1930s in a funerary monument located in the western suburb of Tarraco (Tarragona). They correspond to the lower part of the wall and depict horti picti decorated with animals and hunting scenes. Although the date of the excavation prevents a stratigraphic study, the materials indicate a date between the last third of the 1st century and the 2nd century.Publicación II Congreso Internacional "Educación y Ámbito Penitenciario. I Foro Global de Educación Social en medio abierto con personas (ex) privadas de la libertad. 4 de noviembre de 2025. Actas de resúmenes(2025-10-26) Pozo Serrano, Francisco José delEvolución de la protección penal de los animales frente a los malos tratos Este Congreso Internacional y Foro Global pretenden analizar las políticas, normativas y programas públicos y privados sobre Educación, Ámbito penitenciario y Medio Abierto con (Ex) Personas privadas de la libertad. En el marco del evento, se busca poder compartir buenas prácticas y experiencias a nivel internacional, nacional, regional y local de Educación (Social) para la vida en Libertad y la transferencia de conocimiento sobre la Educación y la Educación Social con Personas (ex) privadas de la libertad. Todo para el fortalecimiento de la RED INN-GFS. Capítulo España, dentro de la Red Internacional. Incarceration Nations Network - Global Freedom Scholars Network (el GFS) como alianza de entidades y universidades con las personas (ex) privadas de la libertad como agentes activos de los procesos de cambio y participación social. Del mismo modo, se focaliza la necesidad de potenciar la Educación Social como práctica, disciplina y profesión a nivel global en el medio penitenciario como derecho para la reinserción y la reincorporación comunitaria.Publicación Ensemble of Constructive Techniques in the Castle of Ricote (Murcia, Spain)(ISPRS Foundation, 2020-07-24) Almela, Íñigo; Martínez, LucíaThe Castle of Ricote, also known as Los Peñascales, is a fortification on a steep hill of the Ricote Valley overlooking the Vega Media of the Segura River, to the east, and the village of Ricote to the west. According to written sources, the history of this castle dates back from the ninth century. However, its military and administrative weight persisted even after the Christian conquest, when it became the headquarters of the Order of Santiago, until the fifteenth century. Despite its poor state of repair, the use of the castle overtime can be established on the site by means of a rather complex sequence of phases and a very heterogeneous set of construction techniques. Although it has been hard to accomplish a complete analysis, in this paper we have attempted a stratigraphic analysis and a synthesis of the techniques used in the medieval interventions, which are highly relevant due to their diversity and special features. Among them, the following have been covered: stonework with lime mortar built through shuttering, rammed earth, and lime-crusted rammed earth. In addition, the two main phases detected, and their respective techniques will also be underlined, since they are present consistently throughout the whole castle.Publicación Estructuras de conservación y distribución de vino en la domus de las Columnas Rojas de Sisapo-La Bienvenida (Almodóvar del Campo, Ciudad Real)(Museu de Badalona, 2024) Pina Burón, María Rosa; Zarzalejos Prieto, María del Mar; Esteban Borrajo, Germán; Hevia Gómez, PatriciaThe domus of the Red Columns is a peristyle house built in the central area of the ancient town of Sisapo-La Bienvenida. In its first layout, dated at the beginning of the 1st century AD, a part of the north wing was intended for a craft use, as can be deduced from the architecture of room 4/7, large in size and provided with a cocciopesto floor with a number of negative structures carved on its surface whose function was difficult to ascertain in the absence of further archaeological evidence of any kind. The latest research carried out in this space, based on the 3D virtual recreation and the analysis of organic waste from the construction material, have allowed us to deepen its functional analysis and to propose an interpretative hypothesis for the set of dependencies located in this part of the domus corresponding to the original project (augustan-tiberian period), in relation to the storage and aging of wine for its possible retail sale.Publicación Supporting growers with recommendations in redvides: some human aspects involved(Springer Nature, 2014-10-10) Santos, Olga C.; Salmeron Majadas, Sergio; González Boticario, JesúsThis paper discusses some human aspects that are to be considered when designing recommendations for RedVides, a cloud based networking environment that collects the status of the crop with sensors and can take decisions through corresponding actuators. The goal behind is to support growers in decision making processes, which can be benefited from collaborations among growers and with other stakeholders.Publicación Comparison of physiological data acquisition for modeling of drivers in autonomous vehicles(Springer Nature, 2025-04-24) Fernandez Matellan, Raul; Puertas Ramírez, David; Martín Gómez, David; González Boticario, JesúsHumans can undergo rapid emotional changes and these changes can significantly affect their ability to perform tasks. Consequently, when we develop Human-Centred Symbiotic Artificial Intelligence (HCSAI) systems to support the interaction between autonomous systems and drivers, the intelligent system controlling the vehicle must adapt to the state of the user. This symbiotic relationship highlights the importance of collaboration and cooperation between humans and agents of artificial intelligence (AI). In the field of Autonomous Vehicles (AV), measurements must be made using non-invasive devices that do not interfere with the driving task. We have therefore used wristbands to measure physiological signals. This comparison is used to select the right equipment for setting up user modelling in different levels of autonomous vehicles. We compared the accuracy, precision and ease of use of three different wristbands: Fitbit Sense2, Empatica E4 and Emotibit. We tested the performance of the bands in two different driving scenarios: SAE Level 4 environment using autonomous golf carts (iCab), and a real-world SAE Level 2 driving environment in Scotland using a Toyota Prius equipped with Comma OpenPilot technology. The Fitbit Sense2 does not allow researchers to access raw data. The Emotibit and the Empatica E4 are designed for research, so they provide access to raw data, while the Empatica E4 is easier to use than the Emotibit. The comparison calls for the development of open source codes that will facilitate integration with different operating systems and other devices, as well as an easy way to use the devices in real time.Publicación Towards multimodal affective detection in educational systems through mining emotional data sources(Springer Nature, 2015) Salmeron Majadas, Sergio; Santos, Olga C.; González Boticario, JesúsThis paper introduces the work being carried out in an ongoing PhD research focused on the detection of the learners’ affective states by combining different available sources (from physiological sensors to keystroke analysis). Different data mining algorithms and data labeling techniques have been used generating 735 prediction models. Results so far show that predictive models on affective state detection from multimodal-based approaches provide better accuracy rates than single-based.Publicación Fusion of physiological signals for modeling driver awareness levels in conditional autonomous vehicles using semi-supervised learning(IEEE, 2024-10-11) Fernandez Matellan, Raul; Puertas Ramírez, David; Martín Gómez, David; González Boticario, JesúsThe evolution of autonomous vehicles (AVs) requires a paradigm shift towards the integration of human factors to improve safety and efficiency at levels 2,3 and 4 of automation. This paper presents a comparison of three different fusion technologies (Low-Level fusion, Medium-Level fusion, and a hybrid fusion), highlighting the critical role of multimodal data integration and semi-supervised learning in predicting and adapting to levels of driver awareness. Our approach uses semi-supervised learning to deal with the data labelling problem, using unlabelled data to train an autoencoder and sparsely labelled data to train a 4-state classifier. Our model facilitates the fusion of data from different physiological signals, including skin electrodermal activity, heart rate, body temperature and acceleration. Using real driving data, the Medium-Level fusion approach gives the best performance, achieving 84% accuracy in predicting situations where the user may not be aware enough to take control of the vehicle. This research highlights the essential nature of fusion technologies to create adaptive and user-centred AV systems.