Renormalization Group Flow

Sometimes I do the ‘tardedest things.

This quarter I needed to take an elective class to fulfill my 10 unit quota. Rather than finding something low stress and easy, I opted to take graduate level equilibrium statistical physics. How did such a class even enter my radar you ask? Well, 1) it came “recommended” by my department as a suggested elective for my emphasis: computational neuroscience, and 2) a fraction of the course content exactly tackles the statistical mechanical basis of An Introduction to the Theory of Neural Computation by Hertz, Krugh and Palmer–basically the bible of neural networks.

For the first couple of weeks I reveled in the success of my class choice. I learned mean field theory and the Ising model of ferromagnetism that closely parallels the Hopfield network model of Hertz and Krugh. Even though the class is intended for second year physics grad students I understood the lectures and did quite well on the homeworks. All seemed to be going as I had hoped… and then came week three.

All of the sudden we delved into the realm of quantum operators and scaling theory and renormalization group flow–all of which are theoretical methods well beyond the scope of my book or my interest for that matter. But I made a commitment to myself that I would stay in this class no matter what happens, so these days, in addition to TAing a human physiology class and spending 4 hours a day in my lab, I spend all of my “free” moments trying to wrap my brain around advanced concepts of theoretical physics.

Once again, here’s to my superior judgement with regards to my academic choices!