Summary of Fairness-accuracy Trade-offs: a Causal Perspective, by Drago Plecko et al.
Fairness-Accuracy Trade-Offs: A Causal Perspective
by Drago Plecko, Elias Bareinboim
First submitted to arxiv on: 24 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
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 paper presents a novel approach to analyzing the tension between fairness and accuracy in machine learning systems from a causal lens. It introduces the notion of path-specific excess loss (PSEL) that measures how much the predictor’s loss increases when enforcing causal fairness constraints. The authors show that the total excess loss (TEL) can be decomposed into local PSELs, allowing for a detailed understanding of the fairness-utility trade-off. They also propose a new neural approach for causally-constrained fair learning and introduce the causal fairness/utility ratio to summarize the trade-off across different pathways. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how machine learning systems can be biased against certain groups. It’s like when you’re trying to make a decision, but some rules might not be fair. The authors want to find a way to balance fairness with being useful. They introduce new ideas called path-specific excess loss and total excess loss that help us understand this trade-off better. They also suggest a new way to learn without biasing certain groups. |
Keywords
» Artificial intelligence » Machine learning
