Negotiating Ontological Boundaries in User-Authored Personal Sensing Systems
The paper presents two open-ended probes using a Wizard of Oz approach to let participants train personalized machine-learning sensing systems on phenomena they define. In a week-long in-the-wild study, participants' interactions revealed four sites of ontological boundary negotiation—what counts as a phenomenon, the subject as part of relations, signal versus noise, and data objectivity—and the authors propose design directions to support such negotiations.