EF Learning-Informed Optimization

Project

EF-LI-Opt-1

Neural Generative Models for Algebraic Curves

Project Heads

Michael Joswig, Sebastian Pokutta

Project Members

Zoe Geiselmann

Project Duration

01.08.2025 – 31.07.2028

Located at

TU Berlin

Description

The goal of this project is: (a) to apply generative methods from AI to produce interesting classes of real plane algebraic curves, and (b) to make progress on the problem of classifying M-curves.
To this end we give explicit combinatorial patchworks for all 121 real schemes of curves in degree 7 and for some infinite families of curves whose members have different degrees. This approach strongly benefits from our development of an algorithm which extracts the isotopy type of a patchwork from its combinatorial data.

External Website

Related Pictures

The nested box curve of degree 6