Mathis Gerdes

I am a theoretical physicist, drawn to how structure (the symmetries and geometry a theory carries, the constraints it imposes) shapes what we can compute and understand. Much of my work brings modern machine learning to bear where that structure is rich but hard to exploit, from gauge theories on the lattice, to the geometry of string compactifications, to generative models that stay faithful to the physics, with the aim of making computation a genuine extension of theoretical reasoning.

Before my current position at MIT, I did my PhD at the University of Amsterdam with Miranda Cheng and Christoph Weniger.