(Locally) Differentially Private Combinatorial Semi-Bandits
This solves the problem of privacy-preserving learning in combinatorial bandits for scenarios like online recommendations, with incremental improvements in regret bounds.
The paper tackles combinatorial semi-bandits under differential privacy and local differential privacy, showing that under common smoothness assumptions, optimal regret bounds can be achieved without additional dimension dependence, nearly matching non-private rates.
In this paper, we study Combinatorial Semi-Bandits (CSB) that is an extension of classic Multi-Armed Bandits (MAB) under Differential Privacy (DP) and stronger Local Differential Privacy (LDP) setting. Since the server receives more information from users in CSB, it usually causes additional dependence on the dimension of data, which is a notorious side-effect for privacy preserving learning. However for CSB under two common smoothness assumptions \cite{kveton2015tight,chen2016combinatorial}, we show it is possible to remove this side-effect. In detail, for $B_{\infty}$-bounded smooth CSB under either $\varepsilon$-LDP or $\varepsilon$-DP, we prove the optimal regret bound is $Θ(\frac{mB^2_{\infty}\ln T } {Δε^2})$ or $\tildeΘ(\frac{mB^2_{\infty}\ln T} { Δε})$ respectively, where $T$ is time period, $Δ$ is the gap of rewards and $m$ is the number of base arms, by proposing novel algorithms and matching lower bounds. For $B_1$-bounded smooth CSB under $\varepsilon$-DP, we also prove the optimal regret bound is $\tildeΘ(\frac{mKB^2_1\ln T} {Δε})$ with both upper bound and lower bound, where $K$ is the maximum number of feedback in each round. All above results nearly match corresponding non-private optimal rates, which imply there is no additional price for (locally) differentially private CSB in above common settings.