Summary of Learning Fair Ranking Policies Via Differentiable Optimization Of Ordered Weighted Averages, by My H. Dinh et al.
Learning Fair Ranking Policies via Differentiable Optimization of Ordered Weighted Averages
by My H. Dinh, James Kotary, Ferdinando Fioretto
First submitted to arxiv on: 7 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
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 approach develops efficiently-solvable fair ranking models for Learning to Rank (LTR) applications, which are critical components in various platforms impacting society. Conventional LTR models have been shown to produce biased results, prompting a discussion on how to address the disparities introduced by ranking systems prioritizing user relevance. While previous fair LTR models have limitations in accuracy or efficiency, this paper integrates efficiently-solvable fair ranking models based on Ordered Weighted Average (OWA) functions into the training loop of an LTR model to achieve favorable balances between fairness, user utility, and runtime efficiency. The approach also enables backpropagation through constrained optimizations of OWA objectives, allowing for integrated prediction and decision models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Fair Learning to Rank (LTR) is important because it helps platforms like job search, healthcare information retrieval, and social media content feeds be fair and unbiased. Current LTR models are biased, so people need a way to make them fair. This paper shows how to do that using something called Ordered Weighted Average (OWA) functions. It’s efficient and works well. |
Keywords
* Artificial intelligence * Backpropagation * Prompting
