News - Thesis announce

Date : Sept. 29, 2023, 1:30 p.m. - Kenneth Ifeanyi EZUKWOKE - Amphi A22 - Espace Fauriel

Probabilistic Graphical Model with Large Language Architecture for Decision-making in Failure Analysis: Application to the Semiconductor Industry 4.0
Failure Analysis (FA) 4.0 is emerging as an essential part of the digital industry, aimed at ensuring maximum reliability and safety of increasingly complex electronic systems in a variety of applications, including autonomous vehicles and digital industrial production (Industry 4.0). At present, failure analysis is mainly carried out by human expertise, involving laborious processes and focusing on individual tasks from production, reliability testing and field feedback. However, these time-consuming procedures limit the global analysis of data combining electrical, material and metrology tests throughout the analysis flow. To address these challenges, the FA 4.0 project aims to provide innovative artificial intelligence-based tools and methodologies that support expert failure analysis throughout the development and manufacture of electronic components and systems.
This thesis presents a formal method for representing textual failure analysis history data using mathematical representation space learning and decision making for failure root cause analysis throughadvanced linguistic modeling.
 
Keywords: Failure Root Cause Analysis, Natural Language Processing, Representation Learning, Natural Language Generation, Failure Analysis Triplets Generation.
 
The thesis defense committee is composed of
    • Massih-Reza Amini            Professor                                               Universiteé Grenoble Alpes                 Examiner
    • Yannis Haralambous          Professor                                               IMT Atlantique BREST                        Reporter
    • Matthieu PUIGT                 Maitre de conférence HDR                   Université du Littoral Côte d'Opale    Reporter
    • Anis Hoayek                       Maître de conférence                            Ecole des mines de Saint-Etienne       Co-supervisor
    • Mireille Batton-Hubert      Professor                                                Ecole des mines de Saint-Etienne       Director of thesis
    • Xavier Boucher                  Professor                                                Ecole des mines de Saint-Etienne       Co-director of thesis
    • Pascal Gounet                    Physical Failure Analyst Engr.               STMicroelectronics                              Guest