Pasar al contenido principal

MALICIOUS ACTORS PROFILING AND DETECTION IN ONLINE SOCIAL NETWORKS THROUGH ARTIFICIAL INTELLIGENCE

PCI2022-135008-2

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Afrontar las Prioridades de Nuestro Entorno
Subprograma Subprograma Estatal de Internacionalización
Convocatoria Proyectos Programación Conjunta Internacional
Año convocatoria 2022
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 8
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Identification of Racial and Sexist Stereotypes in Spanish: A Learning with Disagreements Approach, Identificación de Estereotipos Raciales y Sexistas en Español: Un Enfoque Basado en Aprendizaje con Desacuerdos

  • Urios Alacreu, Elias
  • Rosso, Paolo|||0000-0002-8922-1242
[ES] El discurso del odio ha proliferado significativamente en los ¿ultimos a¿nos, en gran medida impulsado por la adopci¿on generalizada de plataformas de redes sociales. El discurso de odio a menudo opera de manera impl¿¿cita, aprovechando estereotipos sutiles para propagar pensamientos discriminatorios. Estos mecanismos encubiertos han permitido que el contenido odioso se oculte a si mismo, haciendo que su detecci¿on resulte cada vez m¿as compleja. Como resultado, la lucha contra los discursos de odio se ha convertido en una prioridad urgente, impulsando la adopci¿on generalizada de modelos de aprendizaje profundo para detectar y combatir contenidos nocivos. Dada la naturaleza inherentemente subjetiva de los discursos del odio, es necesario desarrollar modelos que sean lo m¿as generalizables posible. Esto ha llevado a la aparici¿on del paradigma de aprendizaje con desacuerdos, que tiene como objetivo introducir desacuerdos dentro de la propia tarea para mejorar la generalizabilidad del modelo. Este trabajo investiga este ¿ultimo paradigma a trav¿es de dos shared tasks. La primera tarea, DETEST-Dis, explora los estereotipos contra los inmigrantes en comentarios en l¿¿nea y fue organizada en IberLEF 2024. Nuestros resultados se encuentran entre los mejores de todos los equipos participantes, superando los enfoques tradicionales. La segunda tarea, EXIST, se centra en el sexismo en los memes y fue organizada en CLEF 2024. En este caso, nuestro rendimiento presenta una mejor¿¿a a¿nadiendo caracter¿¿sticas de un modelo externo as¿¿ como tambi¿en data augmentation. Nuestro c¿odigo fuente se puede encontrar en https://github.com/Buzzeitor30/DETESTS-DIS y https://github.com/Buzzeitor30/EXIST-TFM, [EN] Hate speech has proliferated significantly in recent years, largely driven by the widespread adoption of social media platforms. Hate speech often operates implicitly, leveraging subtle stereotypes to propagate discriminatory views. These covert mechanisms allow harmful content to disguise itself, making detection increasingly complex. As a result, tackling hate speech has become an urgent priority, driving the widespread adoption of deep learning models to detect and combat harmful content. Given the inherently subjective nature of hate speech and its nuanced manifestations, there is a need to develop models that are as generalizable as possible. This has led to the emergence of the learning with disagreements paradigm, which aims to introduce disagreements within the task itself to enhance model generalizability. This paper investigates the latter paradigm through two shared tasks. The first task, DETEST-Dis, explores stereotypes against immigrants in online comments and was organized at IberLEF 2024. Our results are among the best of all participating teams, surpassing traditional approaches. The second task, EXIST, focuses on sexism in memes and was organized at CLEF 2024. Here, our performance is enhanced by adding features from an external model as well as data augmentation. Our source code can be found on https://github.com/Buzzeitor30/DETESTS-DIS and https://github.com/Buzzeitor30/EXIST-TFM., This work has been developed under the project FAKEnHATE-PdC (PDC2022-133118-I00) with founding from European NextGenerationEU/PRTR. EUR and PR carried out part of thier work also in the framework of the following projects respectively: MARTINI (Grant PCI2022-135008-2) and FairTransNLP-Stereotypes (Grant PID2021-124361OB-C31) both funded by MCIN/AEI/10.13039/501100011033 and by European Union NextGenerationEU/PRTR.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

From Past Outbreaks to Future Threats: Detecting Medical Conspiracy Theories With LLMs and Limited Labels

  • Baris-Schlicht, Ipek
  • Korencic, Damir
  • Chulvi-Ferriols, María Alberta
  • Flek, Lucie
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] Online dissemination of conspiracy theories (CTs) during epidemics poses significant risks to public health. This paper addresses the problem of detecting CTs in social media posts with an emphasis on the resource-constrained scenarios characterized by the scarcity of labelled datasets and expert annotations, and the lack of computational budget for large-scale LLM inference. To address these challenges, we investigate resource-efficient methods for CT detection across multiple epidemics. We construct a novel dataset of CT-labelled social media posts covering four major epidemics from the past decade: Ebola, Zika, COVID-19 and Monkeypox. We conduct extensive experiments addressing four research questions: (1) the performance of BERT-like models on individual epidemics, (2) the ability to transfer knowledge from past epidemics to new ones, (3) the efficacy of zero-shot classification using Large Language Models (LLMs) and (4) the feasibility of training BERT-like models on LLM-labelled datasets. Our findings indicate that BERT-like models exhibit highly variable performance across epidemics. Transfer learning from prior epidemics can be effective and their performance can be improved with the number of prior datasets. Zero-shot LLM classifiers, including ensemble methods, achieve performance that matches or surpasses that of fine-tuned BERT-like models. Finally, we demonstrate that BERT-like models trained on LLM-labelled datasets achieve results close to the models trained on expert-annotated data, offering a practical alternative when expert labelling is infeasible. While automated methods can be useful for data analysis, we caution against automatization of content filtering due to the inherent difficulty of CT detection and the potential biases of language models., This work was supported by Bundesministerium fur Bildung und Forschung; Lamarr Institute for Machine Learning and Artificial Intelligence; European Union NextGenerationEU/PRTR, PCI2022-135008-2; Croatian Science Foundation, UIP-2025-02-7498 (AGENTS); MCIN/AEI/10.13039/501100011033, PCI2022-135008-2.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Computational Multimodal Analysis of Polarized Political Discourse After the DANA in Valencia

  • Aloy-Mayo, Maria
  • Rosso, Paolo|||0000-0002-8922-1242
  • Soler-Sanz, Guillem
[EN] This paper aims to analyze sentiment and emotions in news articles and user comments following the devastating floods that took place in Valencia on October 29, 2024, which caused 229 deaths, extensive damage, and significant public concern. The study focuses on news articles about two key politicians: Carlos Maz & oacute;n, leader of the Valencian Popular Party (PP) and President of the Generalitat Valenciana, and Teresa Ribera, Third Vice President of Spain and former Minister for the Ecological Transition (PSOE). Our corpora comprise news articles and their user comments published during the 15-day window following the disaster (2024-10-30 to 2024-11-13) across four Spanish newspapers with different political orientations: El Pa & iacute;s, elDiario.es, El Mundo, and El Confidencial. The news articles were automatically extracted and preprocessed to standardize their format and prepare them for computational analysis. The images of these news articles were also included for a preliminary multimodal analysis approach to explore the semantic relations between linguistic structures and images. Emotion analysis was carried out employing Artificial Intelligence models to quantify the emotional load of news articles and comments from online readers. Large Language Models were used to automatically detect emotions expressed through linguistic expressions targeting both politicians.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Exploring LLMs Applications in Law: A Literature Review on Current Legal NLP Approaches

  • Siino, Marco
  • Falco, Mariana
  • Croce, Daniele
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] Artificial Intelligence (AI) is reshaping the legal landscape, with software tools now impacting various aspects of legal work. The intersection of Natural Language Processing (NLP) and law holds potential to transform how legal professionals, including lawyers and judges, operate, resolve disputes, and retrieve case information to formulate their decisions. To identify the current state of the applications of Transformers (also known as Large Language Models or LLMs) in the legal domain, we analysed the existing literature from 2017 to 2023 through a database search and snowballing method. From 61 selected publications, we identified key application categories such as legal document analysis, case prediction, and contract review, along with their main characteristics. We observed a discernible upsurge in the volume of scholarly publications, a diversification of tasks undertaken (e.g., legal research, contract analysis, and regulatory compliance), and an increased range of languages considered. There has been a notable enhancement in the methodological sophistication employed by researchers in practical applications. The performance of models grounded in the Generative Pre-trained Transformer (GPT) architecture has consistently improved across various legal domains, including contract review, legal document summarization, and case outcome prediction. This paper makes several significant contributions to the field. Firstly, it identifies emerging trends in the application of LLMs within the legal domain, highlighting the growing interest and investment in this area. Secondly, it pinpoints methodological gaps in current research, suggesting areas where further development and refinement are needed. Lastly, it discusses the broader implications of these advancements for real-world legal tasks, offering insights into how LLM-based AI can enhance legal practice while addressing the associated challenges., This work was supported in part by the European Union through the Italian Ministero dell'Universita e della Ricerca (MUR) National Recovery and Resilience Plan (NRRP) of NextGenerationEU, partnership on "Telecommunications of the Future"(program "Research and Innovation on Future Telecommunications Systems and Networks, to make Italy more Smart (RESTART)"), within the project Net4Future under Grant PE0000001; in part by the research project Malicious Actors Profiling and Detection in Online Social Networks through Artificial Intelligence (MARTINI) funded by Ministerio de Ciencia e Innovacion (MCIN)/Agencia Estatal de Investigacion (AEI)/10.13039/501100011033 under Grant PCI2022-135008-2; in part by the European Union NextGenerationEU/Plan de Recuperacion, Transformacion y Resiliencia (PRTR) and FairTransNLP-Stereotypes-Fairness and transparency for equitable NLP applications in social media: Identifying stereotypes and prejudices and developing equitable systems funded by MCIN/AEI/10.13039/501100011033 under Grant PID2021-124361OB-C31; and in part by the European Regional Development Fund (ERDF), EU, A way of making Europe.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Backtranslate what you are saying and I will tell who you are

  • Siino, Marco
  • Lomonaco, Francesco
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] With this work, we hypothesize that semantically enriching a user's text corpus using backtranslation and expansion modules can improve performance for author profiling tasks. To perform this textual enrichment, we translate an author's representative text. Translations are made from one language-the source language-into another-the target language-and then back to the original one. Finally, we expand an author's text by integrating the original version with the back-translated one. Our framework includes these backtranslation and expansion modules followed by a SOTA classifier successfully employed for text classification. The framework is tested on three author profiling datasets from the last three years' shared tasks on fake news, hate speech, irony and stereotypes detection hosted at the CLEF conference for the PAN Lab. This work is an extension of our previous one where we just presented our main idea. Here we improve our framework, and we also investigate more languages and more datasets. Finally, a qualitative analysis is provided. The results confirm that the backtranslation and expansion add-on modules improve model performance on all three datasets evaluated., European Union NextGenerationEU/PRTR, Grant/Award Number: PCI2022-135008-2; ERDF, Grant/Award Number:PID2021-124361OB-C31
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Emotions and News Structure: An Analysis of the Language of Fake News in Spanish

  • Togni, Benedetta
  • Coll, Mariona
  • Chulvi-Ferriols, María Alberta
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] Research has repeatedly demonstrated that fake news tend to appeal to the emotions, but it is less common to consider the presence of emotions in relation to the inverted pyramid, a key aspect of journalistic writing. In this work, we adopt an existing neural model for fake news detection, FakeFlow, one of the few approaches to consider the salience of emotions with respect to the structure of news articles, and adapt it for Spanish. We conduct our experiments on the Spanish Fake News Corpus, introduced at the Fake News Detection in Spanish shared task (FakeDeS), with the goal of gaining a better understanding of the characteristics underlying such texts. In our analyses, we show that both the distribution of affective features and the attention mechanism of the model validate the importance of considering the inverted pyramid structure for detecting fake news., This work was funded by the research project FAKEnHATE-PdC: FAKE news and HATE speech Proof of Concept, funded by MCIN/AEI/10.13039/501100011033 and by European Union NextGenerationEU/PRTR. The work of Paolo Rosso was in the framework of the Iberifier Plus project, co-funded by the European Commission under the Call DIGITAL-2023- DEPLOY-04, European Digital Media observatory (EDMO)-National and multinational hubs (reference 101158511), and the MARTINI project (Malicious Actors Profiling and Detection in Online Social Networks Through Arificial Intelligence), grant PCI2022-135008-2, funded by MCIN/AEI/10.13039/501100011033 and by European Union NextGenerationEU/PRTR. The work of Mariona Coll was in the framework of the research project FairTransNLP on Fairness and Transparency for equitable NLP applications in social media, grant PID2021-124361OB-C31, funded by MCIN/AEI/10.13039/501100011033 and by ERDF, EU A way of making Europe. The work of Berta Chulvi was conducted while at Universitat Politècnica de València.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Studying Fake News Spreading, Polarisation Dynamics, and Manipulation by Bots: a Tale of Networks and Language

  • Ruffo, Giancarlo
  • Semeraro, Alfonso
  • Giachanou, Anastasia
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] With the explosive growth of online social media, the ancient problem of information disorders interfering with news diffusion has surfaced with a renewed intensity threatening our democracies, public health, and news outlets¿ credibility. Therefore, thousands of scientific papers have been published in a relatively short period, making researchers of different disciplines struggle with an information overload problem. The aim of this survey is threefold: (1) we present the results of a network-based analysis of the existing multidisciplinary literature to support the search for relevant trends and central publications; (2) we describe the main results and necessary background to attack the problem under a computational perspective; (3) we review selected contributions using network science as a unifying framework and computational linguistics as the tool to make sense of the shared content. Despite scholars working on computational linguistics and networks traditionally belong to different scientific communities, we expect that those interested in the area of fake news should be aware of crucial aspects of both disciplines., The work of Paolo Rosso was carried out in the framework of the IBERIFIER hub on Iberian media research and factchecking funded by the European Digital Media Observatory (INEA/CEF/ICT/A202072381931 n.2020-EU-IA-0252), and the following projects: XAI-Disinfodemics on eXplainable AI for disinformation and conspiracy detection during infodemics (PLEC2021-007681) and MARTINI on Malicious Actors pRofiling and deTection In online social Networks through artificial Intelligence (PCI2022-134990-2) of the CHISTERA IV Cofund 2021 program, funded by MCIN/AEI/10.13039/501100011033, Spain and by the European Union NextGenerationEU/PRTR, Spain .
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Mitigating the Negative Transfer in Multi-Task Learning for Harmful Language Detection in Spanish and Arabic

  • Magnossao de Paula, Angel Felipe
  • Bensalem, Imene
  • Spina, Damiano
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] Negative transfer continues to limit the benefits of multi-task learning (MTL) in harmful language detection, where related tasks must share representations without diluting task-specific nuances. We introduce task awareness (TA), a methodological framework that explicitly conditions MTL models on the task they must solve. TA is instantiated through two complementary mechanisms: Task-aware input (TAI), which augments textual inputs with natural-language task descriptions, and task embedding (TE), which learns task-specific transformations guided by a task identification vector. Together they enable the encoder to disentangle shared and task-dependent signals, reducing interference during joint optimisation. We integrate TA with BETO and AraBERT encoders and evaluate on six Spanish and Arabic datasets covering sexism, toxicity, offensive language, and hate speech. Across cross-validation and official train-test splits, TA consistently mitigates negative transfer, surpasses single-task and conventional MTL baselines, and yields new state-of-the-art scores on EXIST-2021, HatEval-2019, and HSArabic-2023. The proposed methodology therefore combines a principled architectural innovation with demonstrated practical gains for multilingual harmful language detection. The resources to reproduce our experiments are publicly available at https://github.com/AngelFelipeMP/Arabic-MultiTask-Learning., This work was supported by MCIN/AEI/10.13039/501100011033 (PCI2022-135008-2); European Union NextGenerationEU/PRTR.