Book Announcement: Critical Digital Humanities
Dear SIGCIS Folks, It was a real pleasure meeting everyone in St. Louis! I thought I'd share an announcement for my latest book, Critical Digital Humanities: The Search for a Methodology. It was just published by the University of Illinois Press as part of the "Topics in the Digital Humanities Series." https://www.press.uillinois.edu/books/catalog/48xfp2zp9780252042270.html If you want to purchase it, you can use the offer code S19UIP on IUP's website to get 30% off the paperback, hardcover, and ebook formats. The final chapter of the book gives what I'm calling the "intellectual history" of an algorithm, k-nearest neighbor. The abstract for this chapter, which would be the one of most interest to this community, reads as such: This chapter turns to a lower level of computation to produce a cultural critique and historicization of one of the most important algorithms used in digital humanities and other big data applications in the present moment, the k-Nearest Neighbor or k-NN algorithm. The chapter reconstructs the partial genealogy, the intellectual history, of this important algorithm that was key to sense making in the mid-twentieth century and has found continued life in the twenty-first century. In both its formalized description, its exposition in the papers introducing and refining the rule and its implementation in algorithmic form, and in its actual use, the k-nearest neighbor algorithm draws on dominant mid-twentieth century ideologies and tropes, including partitioning, segregation, suburbanization, and democratization. In the process of situating the k-NN algorithm within the larger field containing other residual and emergent statistical methods, the author seeks to produce an intervention within the developing critical theory of algorithmic governmentality. My present book project continues this line of inquiry by focusing on computer vision algorithms and has been greatly influenced by the SHOT/SIGCIS community and many discussions had at the last conference. Best, Jed -- James E. Dobson, Ph.D. Dartmouth College 420 Moore Hall http://www.dartmouth.edu/~jed
participants (1)
-
James E. Dobson