Geometry as a Guide: Building AI That Understands Shapes

Speaker:  Prof.  Simone Melzi

22 Gennaio 2026 | 9:30
DEIB, Sala Conferenze "E. Gatti" (Ed. 20)

Contatti: Prof. Luca Magri

Sommario

On January 22^nd, 2026, at 9:30 am the seminar on "Geometry as a Guide: Building AI That Understands Shapes" will take place at DEIB Conference Room "Emilio Gatti" (Building 20).

Geometric data are central to many scientific and technological challenges, from computer graphics and robotics to medical imaging and 3D vision. While machine learning has achieved extraordinary success with images and text, learning from geometric data remains fundamentally more challenging due to its non-Euclidean nature. In this talk, I will discuss how machine learning and geometry can mutually reinforce each other when processing 3D shapes.
This talk will discuss how geometry information can be fed into learning-based methods to understand and manipulate geometric structures, and how, conversely, even simple models applied to 3D data naturally rediscover strong geometric priors. This perspective points toward AI systems that do not merely process 3D data, but truly understand shape.

Biografia

Simone Melzi is an Associate Professor at the University of Milano-Bicocca, in the Department of Informatics, Systems and Communication (DISCo). He received The Marie-Curie Individual Fellowships H2020-MSCA-IF-EF-ST-2019, the BE-FOR-ERC 2020 grant from the Sapienza University of Rome, the EG-Italy PhD thesis Award 2018 and the Eurographics Young Researcher Award 2023. He is a member of the Junior Fellow of Eurographics and a scholar of the European Lab for Learning and Intelligent Systems (ELLIS). He works on geometry processing, 3D shape analysis and artificial intelligence.