Mathematics of Generative AI: Theory, Algorithms & Applications

Generative Artificial Intelligence (GenAI) is profoundly transforming the scientific and technological landscape, raising fundamental questions about the mathematical principles that underlie generative models, along with their potentialities and weaknesses. As a matter of fact, while deep learning architectures such as diffusion models, transformers, and variational autoencoders are achieving large success in generating complex data, a comprehensive understanding of their theoretical foundations remains an open and rapidly evolving field of inquiry

The aim of the MathGen Summit is to bring together international researchers (including early career) from different fields to foster a dialogue on the theoretical, algorithmic, and application-related aspects of GenAI techniques. The focus will be on the analysis and development of mathematical models and computational practices that explain, predict, or enhance generative behavior, encompassing areas such as:

  • Geometry, topology, and measure theory in latent representations and manifold learning;
  •  Partial differential equations and dynamical systems approaches to generative flows;
  •  Computational methods underlying generative training and inference;
  •  Applications of Generative AI in science, society, industry.  

Under the support of

  • Università degli studi di Milano. Website

  • Società Italiana di Matematica Applicata ed Industriale
    Website

  • Gruppo Nazionale Calcolo Scientifico, INDAM. Website

  • Decision Science Research Center (DESIRE). Website

Scientific Committee

Paola Causin, Department of Mathematics, University of Milan

Giovanni Naldi, Environmental and Science Department, University of Milan

Alessandro Benfenati, Environmental and Science Department, University of Milan

Alessio Marta, , Department of Mathematics, University of Milan

Elena Morotti, Department of Political and Social Science, University of Bologna

Andrea Sebastiani, Department of Informatics, Bioengineering, Robotics and Systems Engineering, University of Genova

Giovanni Alberti, Department of Mathematics, University of Genova