ENGILBERGE Martin

Tiến sĩ
Nhóm nghiên cứu : MLIA
Ngày đi : 06/12/2020
https://lip6.fr/Martin.Engilberge

Ban lãnh đạo nghiên cứu : Matthieu CORD

Deep multimodal embeddings and grounding

Nowadays Artificial Intelligence (AI) is omnipresent in our society. The recent development of learning methods based on deep neural networks also called "Deep Learning" has led to a significant improvement in visual and textual representation models. In this thesis, we aim to further advance image representation and understanding. Revolving around Visual Semantic Embedding (VSE) approaches, we explore different directions: We present relevant background covering images and textual representation and existing multimodal approaches. We propose novel architectures further improving retrieval capability of VSE and we extend VSE models to novel applications and leverage embedding models to visually ground semantic concept. Finally, we delve into the learning process and in particular the loss function by learning differentiable approximation of ranking based metric.

Bảo vệ luận án : 06/12/2020

Hội đồng giám khảo :

M. AVRITHIS Yannis, Senior Researcher, INRIA Rennes [Rapporteur]
M. THOME Nicolas, Professeur, CNAM [Rapporteur]
Mme LARLUS Diane, Senior Research Scientist, NAVER Labs
M. PONCE Jean, Directeur de Recherche, INRIA - ENS
M. GALLINARI Patrick, Professeur, Sorbonne Université
M. PEREZ Patrick, Directeur de Recherche, Valeo.ai
M. CORD Matthieu, Professeur, Sorbonne Université

Ngày đi : 06/12/2020

Bài báo khoa học 2018-2020

Mentions légales
Sơ đồ site