Toward End-User Debugging of Machine-Learned Classifiers

摘要: Many machine-learning algorithms learn rules of behavior from individual end users, such as task-oriented desktop organizers and handwriting recognizers. These form a generated “program” tailored specifically to the behaviors that user, telling computer what do when future inputs arrive. Researchers, however, have only recently begun explore how an user can debug these programs they make mistakes. We present our progress toward enabling users test learned so everyone benefit intelligent adapted their specific tasks situations.