My name is Daniel Edmiston. I am a Ph.D. candidate in the Department of Linguistics at the University of Chicago. Previous to coming to Chicago, I completed an M.A. in the Department of Linguistics at Seoul National University.
My primary areas of interest are computational linguistics and natural language processing, with a peripheral interest in the Korean language. I was originally trained as a generative linguist, with a focus in syntax and semantics.
Most of my work focuses on the intrinsic evaluation of distributed representations of linguistic entities such as words or morphemes, as learnt by neural networks used to solve NLP tasks. This involves using methods from unsupervised learning and topological data analysis to investigate the representations for hints of what linguists consider to be linguistic structure.
In my short career, I’ve been fortunate to collaborate with many excellent mentors and friends. I have worked on extraposition in the Austronesian language Malagasy with Eric Potsdam, have worked to formalize and adapt the Distributed Morphology framework to operate over strings with Marina Ermolaeva, and have worked on various linguistically-inspired NLP projects with Karl Stratos and Taeuk Kim.