Thèse préparée par Kenneth‐​Ifeanyi Ezukwoke

Title : Aritificial Intelligence for Failure Analysis decisions‐​flow in Semiconductor Industry 4.0

Beginning of the­sis : 05/​10/​2020
End of the­sis : 30/​09/​2023

Abstract : The Failure Analysis (FA) 4.0 pro­ject will address a fun­da­men­tal chal­lenge for the digi­tal world to ensure that increa­sin­gly com­plex elec­tro­nic sys­tems ope­rate with com­plete relia­bi­li­ty and safe­ty. Intelligent auto­ma­tion of this ana­ly­sis deci­sion pro­cess using arti­fi­cial intel­li­gence and machine lear­ning is the objec­tive of the Industry 4.0 consor­tium ; crea­ting a reliable and effi­cient semi­con­duc­tor industry.

This research focuses on deve­lo­ping a com­pre­hen­sive, cen­tral recor­ding ana­ly­sis of metro­lo­gy para­me­ters and process‐​related inho­mo­ge­nei­ties, tra­cked along the pro­duc­tion pro­cess in cor­re­la­tion with exis­ting fai­lure cata­logues. Newly dis­co­ve­red fai­lure mecha­nisms iden­ti­fied in final com­po­nents will be explo­red as part of a novel holis­tic approach for qua­li­ty assu­rance and pro­cess impro­ve­ment. The pro­ject is cha­rac­te­ri­sed by dif­ferent research chal­langes lin­ked with avai­lable data and the unex­haus­tive aspect of human decision‐​making must be consi­de­red with adap­tive sta­tis­ti­cal lear­ning tools.

Keywords : Failure ana­ly­sis ; Industry 4.0 ; Natural lan­guage pro­ces­sing ; Artificial Intelligence ; Machine lear­ning ; Decision theory

Thesis defense :  Possibly 06/​12/​2023

Supervisor :

Partners or/​and fun­ders : EURIPIDES² PENTA

Project : Penta Euripides project Failure Analysis 4.0