CS547 Human-Computer Interaction Seminar (Seminar on People, Computers, and Design)
Fridays 11:30am-12:30pm PT · Gates B3 · Open to the public|
Mark Algee-Hewitt
Stanford
Microgenres: (Mis)Classifying Disciplinary Style
October 26, 2018
This project takes a classification-based approach to the grammar and syntax of disciplinary writing. While the use of machine learning to differentiate among, for example, authors or literary genres, has a long history of use in the computational Humanities, these models tend to focus predominately on semantics, which can only resolve high-level differences if they are based on subject or topic. In this talk, I explore the ways in which style can serve as a set of features that can differentiate among academic writing from different fields. With such a model, misclassifications can identify deviations in style and indicate interdisciplinary writing. When authors incorporate history, anthropology, or psychology into their texts, do they signal these shifts with stylistic markers that exist independently from semantics? Are there identifiable disciplinary signals in academic writing and can these help us explain the incompatibilities between how different fields communicate their information? And, can these micro-generic shifts help us understand the complex history of disciplinary writing in novelistic contexts? |
|