This paper studies weighted low-rank for matrix approximation, providing theoretical improvements over the existing results by [Razenshteyn et al. ’16]. The authors should include more experimental results in the final version of the paper as suggested in the rebuttal.