Guidelines for Responsible and Human-Centered Use of Explainable Machine Learning
Guidelines for Responsible and Human-Centered Use of Explainable Machine Learning
复制标题
可解释机器学习的负责任和以人为本的使用指南
DOI：
--
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Patrick Hall
中科院分区：
文献类型：
--
作者：
Patrick Hall
可解释的机器学习（ML）使人类能够从ML中学习、自动模型决策、法规遵从性、白帽黑客攻击和ML模型的取证分析。可解释的ML（即可解释的人工智能或XAI）已经在许多开源和商业软件包中实现，可解释的ML也是金融服务等行业商业预测建模的重要、强制性或嵌入式方面。然而，像许多技术一样，可解释的ML可能会被滥用，特别是作为有害黑盒的错误保护，例如公平清洗，以及其他恶意目的，如模型窃取[1]，[31],[34]。本文在介绍3.1 - 3.4节中负责任和以人为中心的使用指南的细节之前，在第2节中介绍了几个定义、示例和资格。本文在第4节中总结了看似自然的ML整体方法，包括可解释（即白盒）模型以及对任何影响人类的ML系统的解释，调试和不同影响分析技术。
Explainable machine learning (ML) enables human learning from ML, human appeal of automated model decisions, regulatory compliance, and white-hat hacking and forensic analysis of ML models. Explainable ML (i.e. explainable artificial intelligence or XAI) has been implemented in numerous open source and commercial packages and explainable ML is also an important, mandatory, or embedded aspect of commercial predictive modeling in industries like financial services. However, like many technologies, explainable ML can be misused, particularly as a faulty safeguard for harmful black-boxes, e.g. fairwashing, and for other malevolent purposes like model stealing [1], [31], [34]. This text presents several definitions, examples, and qualifications in Section 2 before covering the details of responsible and human-centered use guidelines in Sections 3.1 – 3.4. This text concludes in Section 4 with the seemingly natural argument for a holistic approach to ML that includes interpretable (i.e. white-box ) models along with explanatory, debugging, and disparate impact analysis techniques for any ML system that impacts humans.
DOI：
--
发表时间：
2018-11
期刊：
ArXiv
影响因子：
--
作者：
C. Rudin
通讯作者：
C. Rudin