In this exciting talk, David Mimno outlines the limitations and uses of representational modeling by employing examples from Minecraft, data on U.S. gun ownership, and 19th-century novels.
What can we learn about the limitations of topic modeling from the video game Minecraft? What information is not represented in the linear regression of gun ownership and population density in the United States? How can models illuminate the relationship between gendered language and author gender in 19th-century novels? David Mimno, Professor in the Information Science department at Cornell University, connects the dots between numbers and evidence by examining the challenges and opportunities of representational modeling.
Micromégas: The Very Small, the Very Large, and the Object of Digital Humanities
Introduction
by Franco Moretti
♦
Piketty’s Model: Literary History Without Fixed Objects
by Ted Underwood
♦
Paragraphs: The Forgotten Middle
by Mark Algee-Hewitt, Ryan Heuser, and Franco Moretti
♦
Small, Large, ‘What’ and ‘So What’: Perspectives and Challenges in Corpus Linguistics
by Susan Conrad
♦
From Numbers to Evidence: Computer-Assisted Scholarship in the Million-Book Era
by David Mimno
Visit the Stanford Literary Lab's website or Twitter page for information on upcoming events and projects.