About me
I am an Assistant Professor in the Data Science department of EURECOM. I build probabilistic methods that let machines perceive and reason about the world under uncertainty, at the meeting point of generative modelling and information theory. My work spans continuous-time, functional and infinite-dimensional diffusion processes; multimodal and discrete generative models; and a family of diffusion-based estimators — MINDE, SΩI, INFO-SEDD, TENDE, DIPHINE — that make classically intractable information-theoretic quantities (mutual information, O-information, transfer entropy, Φ-ID) measurable at scale. A substantial current effort goes into trajectory inference — recovering latent dynamics from snapshot data, where path-space laws are not identifiable from finitely many marginals — and into building foundation models for it.
I regularly publish in the main tracks of ICML, ICLR, NeurIPS and AISTATS, and at their workshops. I am P.I. of the ANR project Gen-T and co-P.I. of P-AI-LOT, an industrial chair on agentic LLMs with Amadeus, and of industrial and European projects on machine perception; industrial patents filed; ELLIS member.
I received my M.Sc. (2016) and Ph.D. (2021) from Politecnico di Torino in Electronic and Telecommunications Engineering. Before EURECOM I worked on statistical signal processing for satellite navigation systems. Here is my full CV.
Contact
Campus SophiaTech: 450, route des Chappes – 06410 BIOT (CS 50193 – 06904 Sophia Antipolis cedex)
Tel: +33 (0)4 93 00 81 45
giulio.franzese[AT]eurecom.fr
