Lavinia Dunagan
ləvɪnjə dʌnəgɪn
Hi there! I'm a PhD student at the University of Michigan School of Information interested in NLP, computational social science, and digital humanities. Right now, I'm particularly interested in the intersection of NLP with 1) archives and 2) religious speech. I'm advised by Dallas Card. My research is supported by an NSF Graduate Research Fellowship. I am in a lab with the estimable Ben Litterer and Meera Desai.
I have a degree in Computer Science from the University of Washington. My last year of undergrad I worked with Jungo Kasai on improving the use of text generation metrics; the summer after graduating I worked with Sofia Serrano at the University of Washington on contrastive explanation regularization.
Nowadays, I think about how (and which) texts are framed as authoritative, how we can computationally model text reuse, and how LLMs fit (or don't) into historical inquiry.
Please reach out to me with any questions!
