Review for NeurIPS paper: How many samples is a good initial point worth in Low-rank Matrix Recovery?

Weaknesses: (1) More information could be added to the introduction of the matrix sensing problem. For example, are A_1…A_m known? Are they fixed or random? What are applications of matrix sensing and maybe what is its relationship to machine learning? Such information perhaps makes this topic clearer and more motivating, especially to the people who touch this topic for the first time.