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..
Making Decisions that Reduce Discriminatory Impacts
Authors: Matt Kusner, Chris Russell, Joshua Loftus, Ricardo Silva
ICML 2019 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate our approach with an example: how to increase students taking college entrance exams in New York City public schools. |
| Researcher Affiliation | Academia | 1The Alan Turing Institute 2University of Oxford 3University of Surrey 4New York University 5University College London. |
| Pseudocode | No | The paper describes an optimization framework (MILP) but does not provide structured pseudocode or an algorithm block. |
| Open Source Code | Yes | We use the Python interface to the Gurobi optimization package to solve the MILP8. 8https://github.com/mkusner/reducing_discriminatory_impact |
| Open Datasets | Yes | We compiled a dataset on 345 high schools from the New York City Public School District, largely from the Civil Rights Data Collection (CRDC)6. 6https://ocrdata.ed.gov/ ... We construct both N(i) and s(i, j) using GIS coordinates for each school in our dataset7: 7https://data.cityofnewyork.us/Education/School-Point-Locations/jfju-ynrr |
| Dataset Splits | No | The paper does not specify explicit training, validation, or test dataset splits. It mentions fitting parameters via maximum likelihood using an 'observed dataset'. |
| Hardware Specification | No | The paper does not specify any hardware details used for running experiments. |
| Software Dependencies | No | The paper mentions using 'the Python interface to the Gurobi optimization package' but does not specify version numbers for either Python or Gurobi. |
| Experiment Setup | Yes | We then solve the optimization problem in eq. (5) (using the MILP framework in Section 3.3) with the structural equation for Y in eq. (7), and a budget b of 25 schools. We construct both N(i) and s(i, j) using GIS coordinates for each school in our dataset7: N(i) is the nearest K = 5 schools to school i and s(i, j) is the inverse distance in GIS coordinate space. |