Address:
Wided Ben Marzouka from the Data Science department (DSD) and from the Latim Laboratory, will present her research about :
"Joint modelling of human and machine knowledge for decision-making"
Abstract: Case-based reasoning (CBR) is a powerful methodology for advanced classification in decision- making systems across various fields, including fault diagnostics (FD) in industrial systems. Traditional CBR systems for FD adopt a pattern recognition perspective, considering all observable features and treating different faults as semantic classes. However, these approaches rely on case bases with fully populated observation vectors, which over look the sequential reasoning process that experts employ based on observed features; a crucial aspect of human experiential knowledge. To address these chalenges, this thesis presents an interpretable approach based on possibilistic Hypothetical Case-Based Reasoning (PH-CBR) for FD, modeling a Decision Support System (DSS). This modeling aims to capture the experience of a human expert performing the complex task of FD, thereby emulating human diagnosing expertise. This thesis presents three contributions. [...]
Organizer(s)
Thesis co-acreditation from IMT Atlantique with the doctoral school SPIN and with "Ecole Nationale des Sciences de l'Informatique de Manouba"
Keywords : Fault Diagnosis, Solving-Problem, Decision-Making, Case-Based Reasoning, Hy- pothetical Case, Possibilistic Similarity