Thesis defence

Multi-Resident Data Association in Smart Homes: Modeling and Learning

High Performance Online Deep Neural Network Training from Synthetic Data with Active Learning for Scientific Computing (or: "How to train neural networks to model physical systems accurately — through smarter data, not more compute?")

CIFRE Ph.D defense : Methodology for Reusing Deep Learning Models (Bootstrapping Deep Learning Journeys)

Quantized DNN learning algorithms with limited hardware overhead for Edge implementation

Reconnaissance des émotions au niveau du groupe en conditions naturelles : vers une approche non individuelle respectueuse de la vie privée

Advanced Early Fusion Approaches for Multimodal Image Fusion for Pedestrian and Vehicle Detection

Beyond Exhaustive Exploration: Heuristic-Guided HTN-SAT Planning

Taenite: a secure transactional library for persistent memory

Linux comme un micro-noyau : le cas de la gestion de mémoire

Towards Simpler Transcripts: Investigating Automatic Simplification of French Spontaneous Speech