Optimal Efficiency-Envy Trade-Off via Optimal Transport

Authors: Steven Yin, Christian Kroer

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We test our solution with both artificial data, and simulated data from a realistic simulator for blood donor matching developed by [21]. Finally, in Figure 4 we investigate the quality of the empirical solutions as the sample size increases. It can been seen that the approximation gap decreases faster than the theoretical rate, confirming our sample complexity bound in Theorem 2.
Researcher Affiliation Academia Steven Yin Department of Industrial Engineering and Operations Research Columbia University New York, NY 10027 sy2737@columbia.edu Christian Kroer Department of Industrial Engineering and Operations Research Columbia University New York, NY 10027 christian.kroer@columbia.edu
Pseudocode Yes Algorithm 1: Projected SGD for Envy Constrained Optimal Transport
Open Source Code Yes Our code is provided as supplemental material.
Open Datasets Yes We test our solution with both artificial data, and simulated data from a realistic simulator for blood donor matching developed by [21].
Dataset Splits No The paper states in the checklist 'Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)? [Yes] See Appendix A.3'. However, Appendix A.3 itself (which is not provided in the current text) or the main text does not explicitly provide percentages or counts for training/validation/test splits.
Hardware Specification No The paper's checklist indicates that hardware specifications are in Appendix A.3, which is not included in the provided text. No hardware details are mentioned in the main body of the paper.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., 'PyTorch 1.9', 'Python 3.8').
Experiment Setup No The paper's checklist indicates that training details and hyperparameters are in Appendix A.3, which is not included in the provided text. No specific experimental setup details like hyperparameter values or training configurations are provided in the main body.