Digital Thursday - Jan-Willem Romeijn, "Data-driven science and undercover theory"

The Centre for Digital Humanities' Digital Thursdays are a series of presentations, activities, and debates that showcases interesting research by a broad range of Digital Humanities scholars. On 13 December, Prof. Dr Jan-Willem Romeijn will give a presentation titled "Data-driven science and undercover theory". You are cordially invited to attend.

Abstract

Machine learning promises new avenues for humanities research. While more traditional methods require the assumption of a model, the application of machine learning methods does not seem to require any theoretical starting points. Or does it? In my talk I discuss earlier attempts in philosophy and statistics to dispense with theory, and show that these attempts have merely forced theory to go undercover. I then argue that this has been detrimental to our understanding of the methods and the science based on it, giving particular attention to applications of machine learning in the humanities. The talk ends with some concrete ideas on how we might bring the theoretical assumptions of data-driven science back into view, and how this might aid digital humanities research.

About the speaker

Prof. Dr Jan-Willem Romeijn is a professor of philosophy of science and head of the Department of Theoretical Philosophy at the Faculty of Philosophy of the University of Groningen in the Netherlands. He studied at Utrecht University, where he graduated cum laude in both physics and philosophy. After that he worked as a financial consultant for two years. From 2000 to 2004 he worked on a doctorate in philosophy at the University of Groningen, for which he graduated cum laude in 2005. The first two years after that, Romeijn lectured in philosophy of science and statistics at the Psychology Department of the University of Amsterdam. Until 2009 he carried out a research project on the intersection of cognitive psychology and philosophy of science in the University of Groningen. Between 2011 and 2016 he worked on a research project concerning single-case chance and statistics. More information can be found in Romeijn’s CV and in his list of publications, presentations, and outreach.