Marco Fiorucci

My primary research interests lie at the intersection of representation learning and multimodal alignment across vision, language, and 3D data. Specifically, I leverage Optimal Transport, with a focus on Gromov-Wasserstein formulations, to develop computationally efficient, scalable, and robust algorithms capable of recovering cross-modal correspondences without paired data.

As a member of the Visual Intelligence and Machine Perception (VIMP) group, I apply these geometric representations and optimal transport to video action recognition, visual grounding, and image deepfake detection. My goal is to enhance the generalization capabilities of modern computer vision systems.