humancompatible.detect: a Python Toolkit for Detecting Bias in AI Models
This toolkit addresses bias detection challenges for AI practitioners, particularly under regulatory requirements like the AI Act, but is incremental as it builds on existing bias detection concepts.
The authors tackled the problem of detecting bias in AI models by introducing humancompatible.detect, a Python toolkit that addresses scalability and computability issues with traditional methods, resulting in new methods like maximum subgroup discrepancy and subsampled ℓ∞ distances for bias evaluation.
There is a strong recent emphasis on trustworthy AI. In particular, international regulations, such as the AI Act, demand that AI practitioners measure data quality on the input and estimate bias on the output of high-risk AI systems. However, there are many challenges involved, including scalability (MMD) and computability (Wasserstein-1) issues of traditional methods for estimating distances on measure spaces. Here, we present humancompatible.detect, a toolkit for bias detection that addresses these challenges. It incorporates two newly developed methods to detect and evaluate bias: maximum subgroup discrepancy (MSD) and subsampled $\ell_\infty$ distances. It has an easy-to-use API documented with multiple examples. humancompatible.detect is licensed under the Apache License, Version 2.0.