
The Price of Differential Privacy For Online Learning
We design differentially private algorithms for the problem of online li...
read it

Mitigating Bias in Adaptive Data Gathering via Differential Privacy
Data that is gathered adaptively  via bandit algorithms, for example ...
read it

DifferentiallyPrivate Federated Linear Bandits
The rapid proliferation of decentralized learning systems mandates the n...
read it

Algorithms for Differentially Private MultiArmed Bandits
We present differentially private algorithms for the stochastic MultiAr...
read it

Contextual Bandits for adapting to changing User preferences over time
Contextual bandits provide an effective way to model the dynamic data pr...
read it

Differentially Private Online Submodular Optimization
In this paper we develop the first algorithms for online submodular mini...
read it

NoRegret Algorithms for Private Gaussian Process Bandit Optimization
The widespread proliferation of datadriven decisionmaking has ushered ...
read it
Differentially Private Contextual Linear Bandits
We study the contextual linear bandit problem, a version of the standard stochastic multiarmed bandit (MAB) problem where a learner sequentially selects actions to maximize a reward which depends also on a user provided perround context. Though the context is chosen arbitrarily or adversarially, the reward is assumed to be a stochastic function of a feature vector that encodes the context and selected action. Our goal is to devise private learners for the contextual linear bandit problem. We first show that using the standard definition of differential privacy results in linear regret. So instead, we adopt the notion of joint differential privacy, where we assume that the action chosen on day t is only revealed to user t and thus needn't be kept private that day, only on following days. We give a general scheme converting the classic linearUCB algorithm into a joint differentially private algorithm using the treebased algorithm. We then apply either Gaussian noise or Wishart noise to achieve jointdifferentially private algorithms and bound the resulting algorithms' regrets. In addition, we give the first lower bound on the additional regret any private algorithms for the MAB problem must incur.
READ FULL TEXT
Comments
There are no comments yet.