NeurIPS 2020

Interior Point Solving for LP-based prediction+optimisation

Meta Review

Four reviewers all rated this paper as above the bar for publication. The reviewers felt the paper contained an interesting, novel, and technically correct contribution in the area of end-to-end training for “predict+optimize” problems. There were a number of specific questions/suggestions, especially adding more baselines and providing more experiments and discussion to determine the benefits of different components of the proposed solution. The authors are encouraged to take these comments into account to improve the paper.