Xueyan Tang

Instance Specific Approximations for Unconstrained Submodular Maximization with Modular Costs

Subset selection for profit maximization is important to applications like web mining, recommendation, and machine learning, which are commonly modeled as unconstrained submodular maximization with modular costs (USM-MC) \max_S\subseteq V f(S)-łambda c(S) where f is a nonnegative monotone submodular utility function, c...