I’m a PhD candidate in Computer Science at the University of Waterloo working in the Computational Neuroscience Research group (CNRG) headed by Chris Eliasmith. My interests are in cognitively constrained representations of language. In particular, I investigate distributed representations of the linguistic content such as words and sentences that can be efficiently implemented in large-scale neural networks while capturing different aspects of linguistic behaviours as observed with users of natural language. To do so, I use a compressed, vector-based representations of words and sentences called semantic pointers that can be manipulated to mimic aspects of syntactic and semantic processing. In my research, I am also investigating how vectors learned from big corpora of text compare to representations derived from much smaller, empirically derived datasets. In 2018 I did a four month internship as a Software Reliability Engineering (SRE) intern at Google, where I designed and implemented a system for detection of automation errors from log data.
I received a master’s degree in Computational Neuroscience at the Bernstein Centre for Computational Neuroscience in Berlin and my Bachelor in Computing from FER at the University of Zagreb. I also worked as a Research Fellow in the Cognovo group at Plymouth University in England. My (occasionally updated) CV is available here.
I come from Croatia, which is a small beautiful country in the Mediterranean whose contours on a map, some people say, look a bit like those of a bird in flight (imagining Istria as the bird’s head). To pronounce my name as a Croatian would, try saying E-va [sounds like “va” in lava]-na [“na” like in retina] Ka-[sounds like “ca” in cartoon]-yeech. Croatians speak Croatian, which is a Slavic language and contains lovely vowelless words such as trg (town square), vrh (peak) and Krk (name of an island). Instead of getting lost while trying to find a trg on a vrh at Krk, one can also visit Plitvička Jezera, Dubrovnik or Paklenica.