Giovanni Luca Marchetti

Senior Research Scientist @ AI4S, Stockholm



CV,  Google Scholar,  GitHub,  X/Twitter,  LinkedIn

I intersect pure mathematics with machine learning. I am interested in algebraic and geometric aspects of neural networks, manifold/representation learning, high-dimensional statistics, and topological data analysis.


Recently I have been developing, with collaborators, Neuroalgebraic Geometry: a general theory of deep learning through the lens of algebraic geometry.

Neuroalgebraic Geometry = Algebraic Geometry ⊗ Deep Learning

Publications

Below is a selection of my academic works, clustered by topic. For a complete list, visit my Google Scholar profile. Click to expand.

In Rubik's Abstract Polytopes, I generalize the Rubik's cube to arbitrary regular polytopes. Below, you can play with the Rubik's torus.