Dynamique des connaissances : Fiche UE : Offre de formation

Nombre de crédits : 3 ECTS

Code Apogée : INF2347M

Contient des enseignements proposés en anglais et des documents totalement disponibles en anglais

Type d'enseignement

Nb heures *

* Ces horaires sont donnés à titre indicatif.

Pré-requis :

M1 in computer science, having completed a traditional computer science program (incluging training in symbolic AI and machine learning).

Compétences attestées (transversales, spécifiques) :

Argumentation is the process of constructing, presenting, and evaluating arguments, typically to persuade or reach a conclusion. It involves the use of logic and reasoning to support or refute claims, often through structured discussions or debates. In the context of computer science, argumentation plays a crucial role in areas such as decision-making, automated reasoning, and human-computer interaction, where the ability to model and analyze arguments is essential.

In this course, we will explore the study of argumentation in computer science from two perspectives: symbolic AI and machine learning. From the symbolic AI perspective, we will examine how argumentation graphs are modeled, the semantics used to derive conclusions, and how these conclusions are connected to knowledge bases. From the machine learning perspective, we will focus on the formalization of natural language arguments and the application of natural language models to tasks such as fallacy detection and relation-based argument mining. By the end of the course, you will have a comprehensive understanding of how these approaches contribute to the advancement of argumentation studies in computing.

Programme de l'UE / Thématiques abordées :

The structure of this course is as follows:
- Session 1
- Argumentation theory and extension-based semantics
- Labelling semantics and links to extension-based semantics
- Session 2
- Advanced notions on admissibility and serialization
- Session 3
- Principled-based approach and ranking-based semantics
- Logic-based argumentation and fallacies
- Session 4
- Assumption-based argumentation and complexity
- ABA+ and comparison to ASPIC+
- Session 5
- Graph neural networks for extension-based semantics
- Learning gradual semantics and enforcement problems
- Session 6
- Dealing with natural language arguments

For an up-to-date version of this syllabus, refer to the online version of this course from my personal website.

Parcours / Spécialité / Filière / Option utilisant cette UE :

M mention : Informatique : M2 Intelligence artificielle

Date de la dernière mise-à-jour : 23/07/2025