Argumentation Strategies: We model different argument attributes and identify them automatically in order to study their usage in persuasive and deliberative discourses. The persuasion strategies have been investigated in a couple of papers ( COLING‐18, INLG‐2019, ACL‐20‐1), considering the important impact of the target audience (CoNLL‐18, ACL‐20‐3). The deliberation strategies, which are of equal importance to persuasion, have been targeted by modeling the interaction between users on more than five million discussions in Wikipedia (ACL‐18).
Argument Mining and Generation :Computational Argumentation is an emergent area in NLP. It studies the automatic understanding and generation of argumentation in natural language. The most studied challenge in this area is argument mining; Given a source of argumentative texts, identify the argumentative units, their roles (e.g. premise or conclusion), and their relations (e.g., support or attack). Over the last few years, we have developed different robust argument mining algorithms that can be applied to various forms of web argumentation successfully. In particular, we proposed a distant supervision approach for identifying argumentative texts (NAACL‑16), and a supervised model for identifying several types of evidence (COLING‑16‑11), examining their distribution across topics (EMNLP‑17). We also have worked on the promising topic of argumentation knowledge graph construction. The idea here is to encode the knowledge in arguments in a graph in order to employ this graph for various computational argumentation tasks such as stance detection and argument generation. A new argumentation graph has been successfully constructed (AAAI‑20) and exploited for generating arguments (ACL‑21).
Bias Anlayis : The detection of bias in media as well as in machine learning models is crucial for addressing the ethical side of artificial intelligence. Following this line of research, we have proposed algorithms for detecting bias in Wikipedia (COLING‑12) and tackling the task of abusive language in online user‑generated discussions (NLP4IF‑19).
Scholalrly Document Processing :
AKASE: Argumentation Knowledge-Graphs for Advanced Search Engines (2024- ) This project seeks to construct argumentation knowledge graphs that encode structured, multi- perspective arguments, providing search engines with enriched, balanced, and credible content. The project holds significant value in advancing the reliability, transparency, and objectivity of search engines. In an era of information overload, misinformation, and biased narratives, constructing argumentation knowledge graphs enables search engines to highlight credible arguments across diverse perspectives, ensuring well-informed decision-making. Collaboration with OpenWebSearch.EU grants access to extensive, high-quality open data essential for building these comprehensive graphs. Further, it enables the efficient implementation and integration of the graphs within search interfaces, fostering a tool that enhances user engagement by delivering reliable, diverse perspectives on complex topics.
Narratives in Argumentation: Exploring the Interplay of Stories and Persuasion in Communication (2024-2028) This research project delves into the intricate relationship between narratives and argumentation in persuasive communication. By categorizing narratives into event-based and effect-based types and exploring structured argumentative schemes, the study aims to establish connections between these storytelling techniques and modes of persuasion, among others. The methodology involves detailed annotation and pattern detection to identify narrative elements and their impact on persuasive effects. The research contributes significantly by enriching rhetorical theories and providing practical insights for applications such as conversational AI systems and text generation tools in news contexts.
Cross-domain Arabic Argument Mining (2025-2026): The lack of annotated datasets limits Arabic argument mining, especially across diverse domains like politics, education, and social media, due to linguistic complexity and varying argumentation structures. This study proposes a cross-domain annotated dataset focusing on claims and evidence types, serving as a benchmark for research and practical use. Leveraging advanced AI techniques like transfer learning and few-shot learning, it aims to enhance cross-domain analysis and promote critical thinking, while supporting applications such as media analysis, debate systems, and misinformation detection.
Sed maximus est nibh, eget pretium turpis dignissim iaculis. Nulla cursus, dolor vel rutrum blandit, libero quam dictum massa, tempus iaculis lectus diam id ligula. Maecenas ac consectetur ipsum. Duis euismod porttitor quam, a tristique lorem congue et. Nullam ultricies varius eros nec mattis. Sed feugiat congue nulla vel consectetur. Mauris scelerisque vehicula tellus, quis ullamcorper elit feugiat eget.