Revealing Positive and Negative Role Models to Help People Make Good Decisions

Code to accompany the paper "Revealing Positive and Negative Role Models to Help People Make Good Decisions"

Overview

TL;DR: We present various strategies for revealing positive and negative targets to help people make good decisions.

Contents and layout

This repository includes the original datasets, generated bipartite graphs, and implementation code for different strategies clearly.

.
├── Fairness/
├── coverageModel/
├── data/
├── graphs/
├── graphsGen/
├── interventionModel/
├── learningSetting/
├── scripts/
├── standardModel/
└── README.md

Below are the specifics of the repository files in correspondence to the paper sections.

• Fairness/: Python notebooks for fairness results (Section 5, Appendix G.3).
• coverageModel/: Python notebooks for computing the radius coverage model results (Section 5, Appendix G.5).
• data/: The initial binary classification datasets (csv files) used in the experiments: Adult, Student-math, Student-portuguese, and Productivity (Appendix G.1).
• graphs/: The generated graphs with the kNN and threshold methods (Appendix G.1).
• graphsGen/: Python notebooks for generating the graphs (Appendix G.1).
• interventionModel/: Python notebooks for computing the targeted intervention model results (Section 5, Appendix G.4).
• learningSetting/: Python notebooks for computing the learning setting results (Section 5, Appendix G.6).
• scripts/: Scripts for analyzing the results, the algorithms and generating the graphs
• standardModel/: Python notebooks for computing the standard model results (Section 5, Appendix G.2).
• requirements.txt: A list of some packages or dependencies needed to run the project.

Getting started

Prerequisites

• Python 3.11.5
• Recommended: Use anaconda or miniconda from here.

Basic usage

1. Clone the repository:

   git clone https://github.com/knaggita/InformationDisclosure.git
   cd InformationDisclosure

2. Install the required packages:

   pip install -r requirements.txt

Citation

If you find this work useful in your research, please cite our paper:

@misc{blum2026revealingpositivenegativerole,
	title        = {{Revealing Positive and Negative Role Models to Help People Make Good Decisions}},
	author       = {Avrim Blum and Keziah Naggita and Matthew R. Walter and Jingyan Wang},
	year         = 2026,
	url          = {https://arxiv.org/abs/2603.02495},
}

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For questions or collaborations, please contact us here.