AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias

Fairness is an increasingly important concern as machine learning models are to support decision making in high-stakes applications such as mortgage, hiring, and prison sentencing. This paper introduces a new open source toolkit for algorithmic fairness, AI Fairness 360 (AIF360), released an Apache v2. 0 license {https: //github. com/ibm/aif360). The main of this toolkit are to help facilitate the transition of fairness algorithms to use in an industrial setting and to provide a common for fairness researchers to share and evaluate algorithms. The package includes a comprehensive set of fairness metrics for datasets and, explanations for these metrics, and algorithms to mitigate bias in and models. It also includes an interactive Web experience (https: //aif360. mybluemix. net) that provides a gentle introduction to the and capabilities for line-of-business users, as well as extensive, usage guidance, and industry-specific tutorials to enable data and practitioners to incorporate the most appropriate tool for their into their work products. The architecture of the package has been to conform to a standard paradigm used in data science, thereby improving usability for practitioners. Such architectural design and enable researchers and developers to extend the toolkit with their algorithms and improvements, and to use it for performance benchmarking. A-in testing infrastructure maintains code quality.