一句话：结合符号执行与局部可解释性，自动生成测试输入以检测ML模型的个体歧视。
可信AI系统需避免算法偏见。个体歧视指两个个体仅在年龄、性别、种族等受保护属性上不同，却从机器学习模型得到不同决策结果。本文研究如何检测ML模型中的个体歧视。在黑盒设置下，此类检测往往需要大量测试，对非平凡系统不可行。为此，作者提出一种自动生成测试输入的方法，将符号执行与局部可解释性两项技术结合，以更有效地生成可能暴露个体歧视的测试用例。实验中将该方法与已知的最佳基准系统进行比较，实证结果表明其在生成测试用例方面非常有效，能够提升黑盒公平性测试的效率与效果。
Any given AI system cannot be accepted unless its trustworthiness is proven. An important characteristic of a trustworthy AI system is the absence of algorithmic bias. 'Individual discrimination' exists when a given individual different from another only in 'protected attributes' (e.g., age, gender, race, etc.) receives a different decision outcome from a given machine learning (ML) model as compared to the other individual. The current work addresses the problem of detecting the presence of individual discrimination in given ML models. Detection of individual discrimination is test-intensive in a black-box setting, which is not feasible for non-trivial systems. We propose a methodology for auto-generation of test inputs, for the task of detecting individual discrimination. Our approach combines two well-established techniques - symbolic execution and local explainability for effective test case generation. We empirically show that our approach to generate test cases is very effective as compared to the best-known benchmark systems that we examine.
DOI 原文 ·
@inproceedings{paperbot5107,
title = {Black box fairness testing of machine learning models},
author = {Aniya Aggarwal and Pranay Lohia and Seema Nagar and Kuntal Dey and Diptikalyan Saha},
booktitle = {FSE},
year = {2019},
doi = {10.1145/3338906.3338937}
}