Understanding the Role of Adaptivity in Machine Teaching: The Case of Version Space Learners

In real-world applications of education, an effective teacher adaptively chooses the next example to teach based on the learner’s current state. However, most existing work in algorithmic machine teaching focuses on the batch setting, where adaptivity plays no role. In this paper, we study the case of teaching consistent, version space learners in an interactive setting. At any time step, the teacher provides an example, the learner performs an update, and the teacher observes the learner’s new state. We highlight that adaptivity does not speed upthe teaching process when considering existing models of version space learners,such as the “worst-case” model (the learner picks the next hypothesis randomlyfrom the version space) and the “preference-based” model (the learner pickshypothesis according to some global preference). Inspired by human teaching, wepropose a new model where the learner picks hypotheses according to some localpreference defined by the current hypothesis. We show that our model exhibitsseveral desirable properties, e.g., adaptivity plays a key role, and the learner’stransitions over hypotheses are smooth/interpretable. We develop adaptive teachingalgorithms, and demonstrate our results via simulation and user studies.