Allocating Variance to Maximize Expectation

Leme, Renato Purita Paes, Stein, Cliff, Teng, Yifeng, Worah, Pratik

arXiv.org Machine Learning 

We design efficient approximation algorithms for maximizing the expectation of the supremum of families of Gaussian random variables. In particular, let $\mathrm{OPT}:=\max_{\sigma_1,\cdots,\sigma_n}\mathbb{E}\left[\sum_{j=1}^{m}\max_{i\in S_j} X_i\right]$, where $X_i$ are Gaussian, $S_j\subset[n]$ and $\sum_i\sigma_i^2=1$, then our theoretical results include: - We characterize the optimal variance allocation -- it concentrates on a small subset of variables as $|S_j|$ increases, - A polynomial time approximation scheme (PTAS) for computing $\mathrm{OPT}$ when $m=1$, and - An $O(\log n)$ approximation algorithm for computing $\mathrm{OPT}$ for general $m>1$. Such expectation maximization problems occur in diverse applications, ranging from utility maximization in auctions markets to learning mixture models in quantitative genetics.