Examinando por Autor "Carrillo de Albornoz Cuadrado, Jorge Amando"

Mostrando 1 - 8 de 8
  • Miniatura
    Publicación
    Authority and Priority Signals in Automatic Summary Generation for Online Reputation Management
    (Wiley, 2021-05-01) Rodríguez Vidal, Javier; Carrillo de Albornoz Cuadrado, Jorge Amando; Gonzalo Arroyo, Julio Antonio; Plaza Morales, Laura
    Online reputation management (ORM) comprises the collection of techniques that help monitoring and improving the public image of an entity (companies, products, institutions) on the Internet. The ORM experts try to minimize the negative impact of the information about an entity while maximizing the positive material for being more trustworthy to the customers. Due to the huge amount of information that is published on the Internet every day, there is a need to summarize the entire flow of information to obtain only those data that are relevant to the entities. Traditionally the automatic summarization task in the ORM scenario takes some in-domain signals into account such as popularity, polarity for reputation and novelty but exists other feature to be considered, the authority of the people. This authority depends on the ability to convince others and therefore to influence opinions. In this work, we propose the use of authority signals that measures the influence of a user jointly with (a) priority signals related to the ORM domain and (b) information regarding the different topics that influential people is talking about. Our results indicate that the use of authority signals may significantly improve the quality of the summaries that are automatically generated.
  • Miniatura
    Publicación
    Automatic classification of sexism in social networks
    (Universidad Nacional de Educación a Distancia (España). Escuela Internacional de Doctorado. Programa de Doctorado en Sistemas Inteligentes, 2025) Rodríguez Sánchez, Francisco Miguel; Carrillo de Albornoz Cuadrado, Jorge Amando; Plaza Morales, Laura
  • Miniatura
    Publicación
    Automatic Generation of Entity-Oriented Summaries for Reputation Management
    (Springer, 2020-04-01) Rodríguez Vidal, Javier; Verdejo, Julia; Carrillo de Albornoz Cuadrado, Jorge Amando; Amigo Cabrera, Enrique; Plaza Morales, Laura; Gonzalo Arroyo, Julio Antonio
    Producing online reputation summaries for an entity (company, brand, etc.) is a focused summarization task with a distinctive feature: issues that may affect the reputation of the entity take priority in the summary. In this paper we (i) present a new test collection of manually created (abstractive and extractive) reputation reports which summarize tweet streams for 31 companies in the banking and automobile domains; (ii) propose a novel methodology to evaluate summaries in the context of online reputation monitoring, which profits from an analogy between reputation reports and the problem of diversity in search; and (iii) provide empirical evidence that producing reputation reports is different from a standard summarization problem, and incorporating priority signals is essential to address the task effectively.
  • Miniatura
    Publicación
    EvALL: Open Access Evaluation for Information Access Systems
    (Association for Computing Machinery (ACM), 2017) Almagro Cádiz, Mario; Rodríguez Vidal, Javier; Verdejo, M. Felisa; Amigo Cabrera, Enrique; Carrillo de Albornoz Cuadrado, Jorge Amando; Gonzalo Arroyo, Julio Antonio
    The EvALL online evaluation service aims to provide a unified evaluation framework for Information Access systems that makes results completely comparable and publicly available for the whole research community. For researchers working on a given test collection, the framework allows to: (i) evaluate results in a way compliant with measurement theory and with state-of-the-art evaluation practices in the field; (ii) quantitatively and qualitatively compare their results with the state of the art; (iii) provide their results as reusable data to the scientific community; (iv) automatically generate evaluation figures and (low-level) interpretation of the results, both as a pdf report and as a latex source. For researchers running a challenge (a comparative evaluation campaign on shared data), the framework helps them to manage, store and evaluate submissions, and to preserve ground truth and system output data for future use by the research community. EvALL can be tested at http://evall.uned.es.
  • Miniatura
    Publicación
    Feature engineering for sentiment analysis in e-health forums
    (Public Library of Science, 2018-11-29) Rodríguez Vidal, Javier; Carrillo de Albornoz Cuadrado, Jorge Amando; Plaza Morales, Laura; Teomiro García, Ismael Iván
    Introduction Exploiting information in health-related social media services is of great interest for patients, researchers and medical companies. The challenge is, however, to provide easy, quick and relevant access to the vast amount of information that is available. One step towards facilitating information access to online health data is opinion mining. Even though the classification of patient opinions into positive and negative has been previously tackled, most works make use of machine learning methods and bags of words. Our first contribution is an extensive evaluation of different features, including lexical, syntactic, semantic, network-based, sentiment-based and word embeddings features to represent patient-authored texts for polarity classification. The second contribution of this work is the study of polar facts (i.e. objective information with polar connotations). Traditionally, the presence of polar facts has been neglected and research in polarity classification has been bounded to opinionated texts. We demonstrate the existence and importance of polar facts for the polarity classification of health information. Material and methods We annotate a set of more than 3500 posts to online health forums of breast cancer, crohn and different allergies, respectively. Each sentence in a post is manually labeled as “experience”, “fact” or “opinion”, and as “positive”, “negative” and “neutral”. Using this data, we train different machine learning algorithms and compare traditional bags of words representations with word embeddings in combination with lexical, syntactic, semantic, network-based and emotional properties of texts to automatically classify patient-authored contents into positive, negative and neutral. Beside, we experiment with a combination of textual and semantic representations by generating concept embeddings using the UMLS Metathesaurus. Results We reach two main results: first, we find that it is possible to predict polarity of patient-authored contents with a very high accuracy (≈ 70 percent) using word embeddings, and that this considerably outperforms more traditional representations like bags of words; and second, when dealing with medical information, negative and positive facts (i.e. objective information) are nearly as frequent as negative and positive opinions and experiences (i.e. subjective information), and their importance for polarity classification is crucial.
  • Miniatura
    Publicación
    Leveraging Unsupervised Task Adaptation and Semi-Supervised Learning With Semantic-Enriched Representations for Online Sexism Detection
    (Wiley, 2024-10-25) Rodríguez-Sánchez, Francisco; Carrillo de Albornoz Cuadrado, Jorge Amando; Plaza Morales, Laura; https://orcid.org/0000-0002-4669-5261
    Over the past decade, the proliferation of hateful and sexist content targeting women on social media has become a concerning issue, adversely affecting women's lives and freedom of expression. Previous efforts to detect online sexism have utilized monolingual ensemble transformers combined with data augmentation techniques that incorporate related-domain data, such as hate speech. However, these approaches often struggle to capture the full diversity and complexity of sexism due to limitations in the size and quality of training data. In this study, we introduce a novel sexism detection system that employs in-domain unlabeled data through unsupervised task-adaptation techniques and semi-supervised learning, using an efficient single multilingual transformer model. Additionally, we incorporate a Sentence-BERT layer to enhance our system with semantically meaningful sentence embeddings. Our proposed system outperforms existing state-of-the-art methods across all tasks and datasets, demonstrating its effectiveness in detecting and addressing sexism in social media text. These results underscore the potential of our approach, providing a foundation for further research and practical applications.
  • Miniatura
    Publicación
    A systematic review on media bias detection: What is media bias, how it is expressed, and how to detect it
    (Elsevier, 2023-09-26) Rodrigo Ginés, Francisco Javier; Carrillo de Albornoz Cuadrado, Jorge Amando; Plaza Morales, Laura; https://orcid.org/0000-0001-6235-6860
    Media bias and the intolerance of media outlets and citizens to deal with opposing points of view pose a threat to the proper functioning of democratic processes. In this respect, we present a systematic review of the literature related to media bias detection, in order to characterize and classify the different types of media bias, and to explore the state-of-the-art of automatic media bias detection systems. The main objectives of this paper were twofold. First, we framed information, misinformation and disinformation within a theoretical framework that allows us to differentiate the different existing misinformation problems such as us media bias, fake news, or propaganda. Second, we studied the state of the art of automatic media bias detection systems: analyzing the most recently used techniques and their results, listing the available resources and the most relevant datasets, and establishing a discussion about how to increase the maturity of this area. After doing a comprehensive literature review, we have identified and selected a total of 17 forms of media bias that can be classified depending on the context (e.g., coverage bias, gatekeeping bias, or statement bias), and on the author’s intention (e.g., spin bias, or ideology bias). We also reviewed, following the PRISMA methodology, the main automatic media bias detection systems that have been developed so far, selecting 63 relevant articles, from which we extracted the most used techniques; including non-deep learning methods (e.g., linguistic-based methods, and reported speech-based methods), and deep learning methods (e.g., RNNs-based methods, and transformers-based methods). Additionally, we listed and summarized 18 available datasets for the task of automatic media bias detection. In conclusion, the current methods for automatic media bias detection are still in their infancy and there is still a lot of potential for improvement in terms of accuracy and robustness. We have proposed some future research lines that could potentially contribute to the development of more advanced techniques.
  • Miniatura
    Publicación
    Using pre-trained language models to automatically identify research phases in biomedical publications
    (Universidad Nacional de Educación a Distancia (España). Escuela Técnica Superior de Ingeniería Informática. Departamento de Lenguajes y Sistemas Informáticos, 2022-07-08) Duran Silva, Nicolau; Plaza Morales, Laura; Carrillo de Albornoz Cuadrado, Jorge Amando
    La ciencia, la investigación y la innovación buscan resolver retos complejos, como por ejemplo abordar un tipo de cáncer o, como recientemente, desarrollar la vacuna del COVID-19. La resolución de estos problemas complejos, especialmente en la investigación biomédica, puede ser costosa, inficiente e insostenible. Suele implicar la colaboración de un amplio conjunto de sectores y actores, puesto que generalmente una sola institución no dispone de los recursos necesarios para desarrollar una innovación de principio a fin, algunos actores se apoyan en otros para combinar sus descubrimientos y lograr una mayor contribución al individuo. De hecho, el número de publicaciones científicas disponibles crece año tras año, especialmente en el ámbito biomédico. Las agencias de financiación, los gobiernos y las universidades están cada vez más interesados en comprender que actividades de investigaci ón se _financian o se llevan a cabo en el ecosistema de investigación, cómo contribuye la ciencia a estas misiones y desafíos, y si existen lagunas de financiación e investigación en diferentes áreas o dominios. La comprensión de los temas abordados por las publicaciones cientificas ha atraído la atención de los investigadores en procesamiento del lenguaje natural (PLN), des de hace varias décadas. Sin embargo, los \dominios específicps", como la biomedicina, se enfrentan a retos y complejidades adicionales. Los modelos neuronales del lenguaje basados en el Transformer han supuesto un gran avance para diversas tareas de PLN, ya que estan preentrenados sobre grandes conjuntos de documentos sin etiquetar y son capaces de aprender una representación universal del lenguaje que se adapta a las tareas posteriores. La mayoría de estos modelos están preentrenados sobre textos de dominio general, aunque hay algunos preentrenados o adaptados a los dominios biomédico y clínico, que son especialmente prometedores para abordar el procesamiento y comprensión de textos en el dominio que nos ocupa. En el presente trabajo, y para dar respuesta a la creciente necesidad de conocer el estado de la investigación en el dominio biomédico, presentamos BATRACIO (BAsic-TRAnslational-Clinical research phases classification in bIOmedical publications), un conjunto de datos para clasificar publicaciones científicas del dominio biomédico en fases de investigación. Exploramos si los modelos lingüisticos preentrenados específicos del dominio superan a los modelos del lenguaje preentrenados en el dominio general, y cómo los adaptamos para enfrentarnos a un conjunto de datos desequilibrado en el dominio biomédico y con categorías adyacentes. Finalmente, en los resultados observamos que los modelos preentrenados del lenguaje basados en BERT, específicamente los modelos preentrenados en el dominio biomédico o científico, ofrecen una gran oportunidad para resolver esta tarea satisfactoriamente. Además, también hemos explorado cómo utilizarlos para la clasificación de textos y que estrategias pueden ser favorables para la clasificación de artículos de investigación biomédica, como la limpieza del texto y el ajuste de hiperparámetros. No obstante, los principales retos específicos de nuestro conjunto de datos son el desequilibrio de clases y que las categorías no son mutuamente independientes, sino que tienen relaciones semánticas de adyacencia entre ellas. Este no era un objetivo principal del proyecto, pero tambien hemos explorado si ligeras modificaciones en la función de pérdida pueden hacer frente a las categorías desequilibradas y adyacentes, aunque los resultados de estos experimentos son parcialmente satisfactorios, apuntan a futuras líneas de investigación.