Rawlsian Fairness in Online Bipartite Matching: Two-Sided, Group, and Individual

Authors: Seyed Esmaeili, Sharmila Duppala, Davidson Cheng, Vedant Nanda, Aravind Srinivasan, John P. Dickerson

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we empirically test our algorithms on a real-world dataset. In this section, we verify the performance of our algorithm and our theoretical lower bounds for the KIID and group fairness setting using algorithm TSGFKIID (Section 5.1).
Researcher Affiliation Academia 1 University of Maryland, College Park 2 Colorado College
Pseudocode Yes Algorithm 1: PPDR( xv) Algorithm 2: TSGFKIID(α, β, γ) Algorithm 3: TSGFKAD(α, β, γ)
Open Source Code Yes Code to reproduce our experiments is available in the blinded format ; we will release that code in deblinded form upon acceptance. https://github.com/anonymousUser634534/TSGF
Open Datasets Yes We run our experiments over the widely used New York City (NYC) yellow cabs dataset (Sekuli c, Long, and Demˇsar 2021; Nanda et al. 2020; Xu and Xu 2020; Alonso-Mora, Wallar, and Rus 2017) which contains records of taxi trips in the NYC area from 2013.
Dataset Splits No The paper describes using the NYC dataset and running trials but does not specify explicit train/validation/test splits or their percentages/counts for reproducibility.
Hardware Specification No The paper does not specify any hardware details such as GPU/CPU models, memory, or cloud resources used for the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., programming language versions, library versions, or solver versions).
Experiment Setup Yes We pick out the trips between 7pm and 8pm on January 31, 2013, which is a rush hour with 10,814 drivers and 35,109 trips. We set driver patience u to 3. Following (Xu and Xu 2020), we uniformly sample rider patience v from {1, 2}. Specifically, we randomly assign 70% of the riders and drivers to be advantaged and the rest to be disadvantaged. The value of pe for e = (u, v) depends on whether the vertices belong to the advantaged or disadvantaged group. Specifically, pe = 0.6 if both vertices are advantaged, pe = 0.3 if both are disadvantaged, and pe = 0.1 for other cases. We run TSGFKIID at the scale of |U| = 49, |V | = 172 for 100 trials.