Pim Huijnen: Opportunities and Challenges of Data-driven Approaches for the Study of Conceptual Change
Thursday 15/11/2018 at 15:00 in room B1.09
Our next meeting will be on the 15th of November. For the second time this year a cultural historian will present a MSDSlab. This time, professor Pim Huijnen will elaborate on the use of semantic networks of words and present some of the data he is working with.
Abstract: Opportunities and Challenges of Data-driven Approaches for the Study of Conceptual Change
Word embeddings – vector representations of words that embed words in a so-called semantic space where the vectors of semantically similar words lie close together – can be used to trace concepts over time without the dependency of particular keywords. In this paper, I would like to show how this technique can add to existing traditions in conceptual history by providing data-driven insights into semantic change. After all, word vector distances can be used to build semantic networks of words. These closely resemble the semantic fields that humanities scholars are familiar with. I will illustrate this approach by our ongoing work on an implementation of word embeddings that helps to trace concepts through time without the dependency of particular keywords. At the same time, we will address in our paper some important challenges that come with the use of word embeddings to represent concepts and conceptual change for the study of history.