PhD Course: Argumentation in Artificial Intelligence

An important question that has deserved the attention of the AI field over the last twenty years and which leads to the development of a wide research area called formal argumentation has been the integration of human capacities to convince people and draw conclusions within smart and sophisticated systems. In fact, in the last twenty years argumentation has emerged as a distinct field within Artificial Intelligence by reflecting in philosophy, law and formal logic. Argumentation in AI differs from classical logic, where, if a proof exists then the statement is not refutable (i.e., what is proven correct once, remains correct). In contrast, the aim of arguments is to persuade, not to formally prove statements. Moreover, arguments are defeasible.

Thus, what made an argument convincing, might not be convincing anymore in the light of new information, and thus retracting conclusions is common when arguing, yielding argumentation non- monotonic.
After providing an introduction to some of the most relevant argumentation frameworks available in literature, the course will focus on both practical and theoretical aspects of argumentation, also concentrating on how incremental techniques can reduce, after adding or removing new pieces of knowledge, the high computational cost of the main argumentation semantics, namely, ways to conclude something.

seminar room - 5th floor, cube 42C

22/01/2024 - (09:00-12:00)
23/01/2024 - (09:00-12:00)
24/01/2024 - (09:00-12:00)
25/01/2024- (09:00-12:00)

Codice Teams :   fe8flr7

12 h - 3 CFU