Summary of Long-term Fairness in Sequential Multi-agent Selection with Positive Reinforcement, by Bhagyashree Puranik et al.
Long-Term Fairness in Sequential Multi-Agent Selection with Positive Reinforcement
by Bhagyashree Puranik, Ozgur Guldogan, Upamanyu Madhow, Ramtin Pedarsani
First submitted to arxiv on: 10 Jul 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Computers and Society (cs.CY); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed Multi-agent Fair-Greedy policy balances score maximization and fairness in selection processes like college admissions or hiring. This paper investigates the possibility of designing sequential decision-making to positively impact long-term social fairness. The authors prove that under identical score distributions across groups, the resource pool and admissions converge to a long-term fairness target set by the agents. They also provide empirical evidence of equilibria under non-identical score distributions using synthetic and real-world datasets. However, the paper cautions against uncoordinated behavior by the agents, which can lead to a reduction in the fraction of under-represented applicants. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how decision-making processes can be designed to promote long-term fairness. It proposes a new policy called Multi-agent Fair-Greedy that balances two goals: maximizing scores and being fair. The authors test this policy on different datasets and find that it works well when the score distributions are the same for all groups. However, they also warn that if agents don’t work together or follow rules, fairness can actually decrease over time. |
