* Title : Ginzburg-Landau dynamics for data science

* Speaker : Seunggyu Lee

* Abstract : 

Artificial intelligence is transforming many areas of science and technology, yet the mathematical principles underlying modern data-driven methods remain only partially understood. In this talk, I will discuss data science from the viewpoint of PDEs, with particular emphasis on dynamics associated with the Ginzburg-Landau energy functional. The main idea is to reinterpret several representative processes in AI and data science, including neural network training, data generation, and topological filtration, as evolution problems governed by energy, geometry, and diffusion. These examples illustrate how PDE dynamics can provide a mathematical framework for understanding and designing modern data-driven methods.