Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Fairness under Competition
Authors: Ronen Gradwohl, Eilam Shapira, Moshe Tennenholtz
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Next, in order to explore the empirical prevalence of our theoretical results, in Section 5 we describe several experiments we ran on Lending Club loan data. |
| Researcher Affiliation | Academia | Ronen Gradwohl Department of Economics University of Haifa EMAIL Eilam Shapira Faculty of Data and Decision Sciences Technion Israel Institute of Technology EMAIL Moshe Tenneholtz Faculty of Data and Decision Sciences Technion Israel Institute of Technology EMAIL |
| Pseudocode | No | The paper describes algorithms and methods but does not present any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code is available at https://github.com/eilamshapira/Fairness Under Competition. |
| Open Datasets | Yes | For the first three we used Lending Club loan data for the years 2007-2015, a dataset that includes roughly 890,000 peer-to-peer loans through Lending Club (2021). ... we also ran experiments in which classifiers training sets were independently chosen, experiments using a different dataset (specifically, Becker and Kohavi, 1996) |
| Dataset Splits | No | The paper describes how training data was selected for individual classifiers (e.g., 'random examples', 'disjoint', 'consisting only of individuals with a mortgage') but does not specify explicit train/test/validation splits from the overall datasets in percentages or counts needed to reproduce a comprehensive data partitioning strategy. |
| Hardware Specification | No | Running all experiments took a few hours on a home computer. |
| Software Dependencies | No | The fairness adjustment was implemented using the open-source Fairlearn (2025) package. |
| Experiment Setup | No | The paper describes the types of classifiers used (logistic regression, decision tree, random forest) and how their training data was prepared. However, it does not provide specific hyperparameters such as learning rates, batch sizes, number of epochs, or optimizer settings for these models. |