NeurIPS 2020

Training Stronger Baselines for Learning to Optimize

Meta Review

This is a strong paper with sound fundamentals, and while the techniques used are not necessarily novel, the authors compensated throuugh a broad set of experiments, which will be helpful to a reader. The effectiveness of the proposed approach has been sufficiently demonstrated and code is made available for better reproducibility. One point of clarification that would help the reader is how much improvement comes from curriculum learning and how much comes from imitation learning.