Summary of Exagree: Towards Explanation Agreement in Explainable Machine Learning, by Sichao Li et al.
EXAGREE: Towards Explanation Agreement in Explainable Machine Learning
by Sichao Li, Quanling Deng, Amanda S. Barnard
First submitted to arxiv on: 4 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computers and Society (cs.CY); 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 A novel framework, EXplanation AGREEment (EXAGREE), is introduced to address fundamental ranking-based explanation disagreement problems in machine learning, particularly from stakeholder-centered perspectives. The approach leverages a Rashomon set for attribution predictions and optimizes within this set to identify Stakeholder-Aligned Explanation Models (SAEMs) that minimize disagreement with diverse stakeholder needs while maintaining predictive performance. Rigorous empirical analysis on synthetic and real-world datasets demonstrates that EXAGREE reduces explanation disagreement and improves fairness across subgroups in various domains. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine learning models need explanations to be trusted, transparent, and fair. But disagreements among these explanations make it hard to rely on them, especially in important situations. This paper fixes four main problems with ranking-based explanation disagreements and creates a new way to bridge different interpretations in explainable machine learning. The approach uses a special set for attribution predictions and optimizes within that set to find models that agree with different people’s needs while still being good at predicting things. |
Keywords
* Artificial intelligence * Machine learning
