Jonathan So - 2019 Cohort

Jonathan So  js2488@cam.ac.uk

United Kingdom

Department of Engineering, Darwin College

PhD thesis: Probabilistic Representation Learning

Research interests 

1. Probabilistic machine learning

2. Computational neuroscience

In my doctoral research I will investigate how we can learn useful probabilistic representations from data in a fully unsupervised manner, given minimal prior assumptions about underlying structure.  I will look to to be guided by insights from neuroscience, primarily exploring approaches that appear to be somewhat biologically plausible; partly because such insights may provide useful clues as to which avenues may prove fruitful, but also because ultimately our findings may better our understanding of perception in the brain.

Who or what inspired you to pursue your research interests?

My interest in machine learning began in earnest during my undergraduate study, where I was introduced to the biological and computational foundations of cognition in the course Cognitive Neural Networks taught by Prof. Howard Bowman.  It was several years later however, working as an intern under Prof. Maneesh Sahani in the Gatsby Unit at UCL, that I became fascinated with the idea of perception as probabilistic inference, ultimately shaping my desire to work at the intersection of probabilistic machine learning and computational neuroscience.