Review for NeurIPS paper: Efficient Online Learning of Optimal Rankings: Dimensionality Reduction via Gradient Descent