Masters Thesis
A Mathematical Model of Morphogenesis
How do cells know when to stop dividing? How does differentiation govern structure? How does symmetry breaking occur? These are the questions at the heart of my MASc thesis.
Overview
My thesis explores the fundamental questions of biological development through the lens of mathematics and machine learning. I am building a Python engine powered by an equivariant neural network to model morphogenesis, the biological process that causes an organism to develop its shape.
Inspirations
- Conway's Game of Life (emergent complexity from simple rules)
- Neural Cellular Automata (differentiable, learnable cellular automata)
- Alphafold (deep learning applied to biological structure prediction)
Key Questions
“How do cells know when to stop dividing?”
“How does differentiation govern structure?”
“How does symmetry breaking occur?”
Approach
The engine uses an equivariant neural network architecture that respects the rotational and translational symmetries inherent in biological systems. Training data is derived from experimental observations of cellular development across multiple model organisms.
Status
This project is currently on hold. Midway through my Master's degree, I made the decision to step away from academia and pursue the frontier of AI directly, co-founding a start-up called Bread Technologies. A high-level overview and code for the project can be found on Github.