04/06/2026 12:00

Sala Descubre/Online Teams. Edificio 8E. Acceso J - 4ª Planta | Universitat Politècnica de València. 46022

Abstract:

The latest developments in AI and NLP have placed the focus on the reasoning capabilities of state-of-the-art systems. Reasoning models (e.g., GPT-5, Gemini 3, DeepSeek-R1) can achieve outstanding performance in a wide range of benchmarks covering tasks of different nature by producing intermediate chains of tokens that simulate step-by-step reasoning (i.e., chain-of-thought). These advances immediately led to a debate about whether these models actually have the ability to reason or whether they simply mimic reasoning-like natural language based on the large amounts of data seen during training.

In this talk, I will introduce the fundamentals to understand natural language reasoning from the perspective of argumentation, a discipline that has been dedicated to the study of reasoning and communication since ancient Greece. Then, I will cover some of my recent work co-developed with my colleagues in which we design new evaluation tools and techniques for argumentative reasoning analysis and assessment. To conclude, I will provide a comprehensive overview of the main limitations identified in AI models regarding their reasoning capabilities.

Bio:

Ramon Ruiz-Dolz is a Lecturer in Computing at the University of Dundee (United Kingdom). Before that, he was PDRA at the Centre for Argument Technology, and a research assistant at the Universitat Politècnica de València. He also held visiting fellowships at the Hong Kong University of Science and Technology (China), ELLIS Alicante (Spain), and the National Institute of Informatics (Japan). Ramon obtained his PhD in Computer Science from the Universitat Politècnica de València with a Cum Laude distinction awarded with the Extraordinary Prize. Ramon’s main research topics are Computational Argumentation and Natural Language Processing. His current research focuses on the analysis of the natural language reasoning capabilities of LLMs from an argumentative perspective, and the integration of concepts from argumentation theory into NLP algorithms to improve reasoning and create tools for developing critical thinking skills.