Dynamical Structure-Preserving Manifolds (dSPM)
A framework for analytically programming reservoir computers using physics-based representations, eliminating the need for traditional training while maintaining interpretability.
My research sits at the confluence of biology, computational neuroscience, dynamical systems theory, and pure mathematics. I'm particularly interested in how mathematical structures—especially quaternions and geometric algebra—can illuminate both neural computation and fundamental physics.
How can we understand neural circuits as dynamical systems? How do continuous dynamics give rise to discrete computations?
The deep connections between four-dimensional rotation mathematics, group theory, number theory, and fundamental physics.
Building machine learning systems whose internal representations we can actually understand, analyze, and control.
The unreasonable effectiveness of mathematics in describing physical reality—and what that tells us about both.