Summary of Axiomatic Characterisations Of Sample-based Explainers, by Leila Amgoud et al.
Axiomatic Characterisations of Sample-based Explainers
by Leila Amgoud, Martin C. Cooper, Salim Debbaoui
First submitted to arxiv on: 9 Aug 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: 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 paper delves into explaining decisions made by black-box classifiers, focusing on feature-based explainers that generate explanations from samples or datasets. The authors identify desirable properties for these explainers, highlighting relationships and incompatibilities between them. They then introduce a family of explainers that satisfy two key properties, providing sufficient reasons (weak abductive explanations) to unravel subfamilies that satisfy subsets of compatible properties. This leads to a full characterization of all explainers satisfying any subset of compatible properties. Notably, the paper introduces the first broad family of explainers guaranteeing the existence of explanations and their global consistency. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making complex computer models more understandable. Right now, these “black-box” models make decisions without explaining why they made those choices. The authors are trying to figure out how to create explanations for these models using certain types of data or samples. They identify some important characteristics that an explanation should have and then show how different approaches can satisfy these characteristics. This research helps us better understand how complex computer models work and could lead to more reliable and trustworthy decisions. |
