We study language technologies that help people reason, deliberate, and engage critically with information. Our work connects NLP, computational argumentation, discourse analysis, structured representations, and human–AI interaction in systems that are transparent, useful, and grounded in real-world settings.
We focus on settings where perspectives matter, disagreement is natural, and decisions benefit from reflection rather than speed alone. That means making reasoning visible, connecting claims to evidence, and designing systems that help people compare viewpoints more responsibly.
Our projects combine argumentation theory, discourse studies, NLP, and human-centered evaluation. Depending on the project, we create datasets, build models and knowledge graphs, develop interactive systems, and study how people actually use them in reasoning-heavy tasks.
Interactive systems that make reasoning visible, inspectable, and useful in high-stakes discussion.
Models and resources for claims, evidence, persuasion, argument search, and deliberative strategies.
Tools for fallacy detection, critical-question generation, bias awareness, and reflective AI use.
Narratives, discourse structure, knowledge graphs, multilingual resources, and controllable generation.
A compact view of the group's main research directions. Open a theme for keywords, motivation, and representative publications.
We study how AI systems can participate in structured reasoning with people instead of only producing final answers. This line combines multi-agent discussion, moderation, real-time summaries, argument maps, critical-question generation, and hybrid human–AI workflows for reasoning-intensive tasks.
This is the group’s core line of work on how arguments are expressed, structured, and evaluated in natural language. It spans argument components and relations, evidence types, persuasion strategies, argument arrangement, deliberative strategies, causality, and search over arguments at web scale.
This theme asks whether AI helps people think better: identify fallacies, ask the right critical questions, compare alternatives, and notice bias or audience effects. It connects reasoning support with the evaluation of persuasiveness, stance balance, and reflective use of AI.
We examine how discourse organization, narrative, framing, and structure shape communication, and we build representations and tools that make these patterns computationally useful. This includes narrative use in online discussions and Arabic discourse, argumentation knowledge graphs, controllable generation, and structure-aware exploration tools.
Our current projects connect structured language analysis with usable systems for search, deliberation, generation, and multilingual discourse research.
AKASE developed argumentation knowledge graphs that encode structured, multi-perspective arguments for advanced search settings. In collaboration with OpenWebSearch.EU, the project used large-scale open web data to support more balanced retrieval, richer justification, and search interfaces that surface credible arguments across perspectives. The project was completed in 2026.
This project explores the relationship between narratives and argumentation in persuasive communication. It examines how storytelling patterns, argumentative schemes, and discourse organization interact in online discussions and other forms of public reasoning.
This project develops resources and methods for Arabic argument mining across domains such as debate, editorials, and public discourse, with attention to claims, evidence, and cross-domain generalization.
We welcome thesis students who want to work on language, reasoning, and society. Topics can be scoped for bachelor’s or master’s level and connected to ongoing projects, datasets, and tool development.
We regularly supervise bachelor’s and master’s theses connected to ongoing group research.
We welcome students who are curious about language, reasoning, AI, and society.