Learning algorithms are now important tools for parsing neuroscientific data and, increasingly, sources of inspiration for formalizing biological learning. My thesis work explored both dimensions of learning algorithms. In my early PhD, I explored the uses of machine learning employed to predict neural recordings. In visual physiology, understanding often progresses by constructing hand-built models that explain responses – but what does it mean if a machine learning model does better, both for the individual research project and the field more generally? The theme of this earlier work is nonlinearity of responses, and the slippery nature of understanding nonlinear systems with simplified models. For these projects, machine learning acted as a foil and a tool for the typical approaches of neurophysiology. As my interests developed, I grew increasingly interested in attempting to understand the brain itself using the mathematical language of machine learning. What learning algorithms work well for massively parallel, distributed systems? Which of these map onto what we know about how the brain learns, if any? What predictions can be made about biological phenomena if we assume certain learning algorithms? These questions remain very interesting to me. However, I currently believe that such a normative approach to understanding biological learning requires better information about biological learning, particularly about how brains coordinate learning among constituent neurons. After my PhD I hope to use emerging recording technologies to ask how neural systems coordinate learning, to develop machine learning tools to assist these technologies, and to understand the results in the context of our mathematical understanding of what effective learning requires.
After graduating from Williams College with a degreee in physics, teaching high school chemistry for a year in Mexico, and obtaining a master’s in nanoscale simulation and biomaterials at Northwestern, I joined this awesome lab for computational neuroscience! I’m all about nonlinear life trajectories, which is fair, since the brain is nonlinear too.