Edited by Roland Vogl
Chapter 14: Generalizability: Machine learning and humans-in-the-loop
Automated decision tools, which increasingly rely on machine learning (ML), are used in systems that permeate our lives. Examples range from systems for offering credit and employment, to serving advertising. We explore the relationship between generalizability and the division of labor between humans and machines in decision systems. An automated decision tool is generalizable to the extent that it produces outputs that are as correct as the outputs it produced on the data used to create it. The generalizability of a ML model depends on the training, data availability, and the underlying predictability of the outcome that it models. Ultimately, whether a tool’s generalizability is adequate for a particular decision system depends on how it is deployed, usually in conjunction with human adjudicators. Taking generalizability explicitly into account highlights important aspects of decision system design, as well as important normative trade-offs, that might otherwise be missed.
You are not authenticated to view the full text of this chapter or article.
Elgaronline requires a subscription or purchase to access the full text of books or journals. Please login through your library system or with your personal username and password on the homepage.
Non-subscribers can freely search the site, view abstracts/ extracts and download selected front matter and introductory chapters for personal use.
Your library may not have purchased all subject areas. If you are authenticated and think you should have access to this title, please contact your librarian.