I am a 6th (and final) year PhD student in Computer Science at Harvard University, working with Demba Ba. In the past I interned at Amazon AWS where I worked with Karim Helwani and graduated with a diploma in Electrical and Computer Engineering from the National Technical University of Athens, where I completed my thesis under the supervision of Petros Maragos.
These days I'm mainly interested in theoretically deriving and understanding deep learning representations. Many learning tasks have physical constraints (such as compute or memory limitations) or symmetries (such as rotational invariance or invariance to the choice of a reference frame) that require the development of specialized architectures. I've worked on this problem for the majority of my PhD; two works that are most representative of my current interests are this one on learning group representations and this one on gauge-invariant learning on arbitrary topologies.
