PhD defence, Yevhenii Prokopalo

Date : 20/10/2022
Time : 14h00
Location : IC2, Boardroom and online
 

Title : Human assisted correction for speaker diarization of an incremental collection of documents
 

Jury members :

• Claude BARRAS, Chercheur, HDR, Vocapia, Reviewer
• Corinne FREDOUILLE, Professeur, LIA, Avignon, Reviewer
• Sophie ROSSSET, Professeur, Université Paris Saclay, Examiner
• Hervé BREDIN, Chargé de Recherche, IRIT, Toulouse, Examiner
• Gaël LE LAN, Chercheur, META FASTAI, Examinateur
• Anthony LARCHER, Professeur, LIUM, Université du Mans, PhD Director
• Loïc BARRAULT , Chercheur, META AI, Co-supervisor

 

Abstract:

This research aims at designing an autonomous speaker diarization system able to adapt and evaluate itself. The goal of this research was to apply the concept of human-assisted lifelong learning to the speaker diarization task, also known as speaker segmentation and clustering. More specifically, this work aims at designing an efficient way of interaction between the automatic diarization system and a human domain expert to improve the quality of diarization generated by an automatic system while limiting the workload for the human domain expert. This manuscript proposes an alternative point of view on the definition of the lifelong learning intelligent systems, a dataset designed for evaluation of the lifelong learning diarization systems and the metric for evaluation of human-assisted systems. The main contribution of this work lies in the development of the human-assisted within-show and cross-show diarization methods.

 

Keywords:

Lifelong learning, Active learning, Speaker diarization, Diarization of an incremental collection of documents