PhD defence : Kaifeng Zou
Team: IMAGeS
Date & time : Thursday, September 28th, at 14:00 at ICube, Illkirch (Room A207)
Title : "Advancements in Generative Models: Enhancing Interpretability and Control of Complex Data through Disentanglement and Conditional Generation."
Abstract : Generative models are a class of machine learning models that aim to learn the underlying distribution of a given dataset and generate new data points that resemble the original data. These models have gained significant attention in recent years due to their ability to produce realistic and diverse samples of data. Generative models, such as VAEs ( Variational Autoencoders) , GANs (Generative Adversarial Networks), EBMs (Energy-Based Models), diffusion models, have shown significant promise in many fields, including image generation, speech synthesis, and natural language processing, and continue to be an active area of research, with new models and techniques being developed to improve their performance and broaden their applications. One of the most important application of generative model is disentangled representation, which refers to a type of feature learning in which the underlying factors or attributes of data are learned and represented independently. In our research, we utilize disentangled representations to tackle the challenge of sex determination and provide insights into the classification results. This is achieved by generating hip bones for the same individual from both sexes and subsequently conducting a comparison to identify sex-related distinctions. Additionally, we aim to acquire knowledge about the high-level factor and its attributes by learning the associated representation, allowing us to effectively control label-related characteristics. To achieve this, we introduce two innovative VAE frameworks aimed at learning the label-associated representation and enhancing VAE's generation quality simultaneously. Additionally, our research also makes a contribution to conditional generation. We apply a diffusion model to sequential data, showcasing its ability to generate 3D facial expressions, which involve time series data. This reverse process provides remarkable flexibility, enabling various types of conditioning and generation through a single training procedure.
The jury will be composed of:
Le dépôt des candidatures pour les postes d’enseignants-chercheur est ouvert. Les offres sont...
L'équipe de l'Université de Strasbourg et la délégation Alsace du CNRS se sont brillamment...
Le vendredi 20 septembre a eu lieu la réunion de lancement du projet INTERREG 2PhaseEx, au...
Paris 27 aout 2024 – ARCHOS annonce que POLADERME, filiale du Startup studio Medtech du groupe...
La 11e journée du département de mécanique s'est tenue le 18 juin 2024. Lors de cette...
A l'occasion de la soirée de gala du 103ème congrès de l’association française des professionnels...
Le 32ème Congrès Français de Thermique de la Société française de thermique (SFT) organisé par le...
L'un des 3 Prix du meilleur poster de la 11èmes journées de la Fédération de Médecine...
La neurostimulation guidée par l’imagerie cérébrale pour traiter les patients atteints d’épilepsie...
L'un des 3 Prix du meilleur poster de la 11èmes journées de la Fédération de Médecine...
L'AABC Europe (Advanced Automotive Battery Conference) est un événement annuel rassemblant des...