Private Sequential Learning

Abstract

<jats:p> Can we learn privately and efficiently through sequential interactions? A private learning model is formulated to study an intrinsic tradeoff between privacy and query complexity in sequential learning. The formulation involves a learner who aims to learn a scalar value by sequentially querying an external database and receiving binary responses. In the meantime, an adversary observes the learner’s queries, although not the responses, and tries to infer from them the scalar value of interest. The objective of the learner is to obtain an accurate estimate of the scalar value using only a small number of queries while simultaneously protecting his or her privacy by making the scalar value provably difficult to learn for the adversary. The main results provide tight upper and lower bounds on the learner’s query complexity as a function of desired levels of privacy and estimation accuracy. The authors also construct explicit query strategies whose complexity is optimal up to an additive constant. </jats:p>