Reviews: Accelerated Stochastic Matrix Inversion: General Theory and Speeding up BFGS Rules for Faster Second-Order Optimization

This paper presents an accelerated version of the sketch-and-project algorithm, an accelerated algorithm for matrix inversion and accelerated variants of deterministic and stochastic quasi-Newton updates. I strongly believe that this line of research is of interest for the ML and optimization communities, and that the algorithms and theoretical results presented in this paper are significant and novel. Moreover, the numerical results presented in the paper clearly illustrate the effectiveness of the approaches presented in the paper. For this reason, I strongly recommend this paper for publication. Below I list my minor concerns with the paper.