BOUNOUA Mustapha

Mustapha Bounoua is a postdoctoral fellow at EURECOM's Data Science department. His research focuses on advancing generative modeling techniques, particularly in multi-modal and diffusion-based frameworks. He specializes in information theory applications, unpaired data alignment, and neural estimation methods. His work bridges machine learning, computer vision, and statistical learning.

Research interests include diffusion models for multi-modal data fusion, entropy minimization in unpaired domains, and score-based approaches for information estimation. His contributions address challenges in generative adversarial networks, latent space representations, and mutual information estimation through neural diffusion processes.

Publications highlight advancements in masked diffusion techniques for multi-modal data, score-based information estimation (SΩI), and latent diffusion architectures. These works reflect a focus on scalable and theoretically grounded solutions for complex generative tasks.

No scientific awards or grants are explicitly listed in the provided information. He currently holds no formal advising roles, with no listed students or postdoctoral trainees.

Collaborations are centered at EURECOM's Data Science group, with active participation in top-tier conferences like ICML, NeurIPS, and ICLR. His research outputs demonstrate a commitment to both foundational theory and applied generative modeling challenges.