Essential Open-Source Tools for Bias Detection and Mitigation

As the use of AI technologies permeates every aspect of our lives, from decision-making processes in industries to personal recommendations on social media, the need to develop and deploy unbiased systems becomes more and more critical. We collected a list of open-source tools and libraries created to detect, understand, and address biases in ML models.

For those specifically interested in strategies for mitigating bias in large language models (LLMs), we have compiled a detailed analysis in a separate article:

Now, to the list:

• AI Fairness 360 (AIF360) by IBM: An extensible toolkit that provides algorithms and metrics to detect, understand, and mitigate unwanted algorithmic biases in machine learning models. → GitHub

• Fairlearn: A library to assess and improve the fairness of machine learning models. → GitHub

• What-If Tool: An interactive visual interface designed by Google for probing the behavior of machine learning models. It's useful for investigating model performances on a dataset and can be used for bias detection → GitHub

• FAT Forensics: A Python toolbox that offers functionalities to evaluate the fairness, accountability, and transparency (FAT) of AI systems, including tools for data and model inspection → page

• Themis-ml: A library for fairness-aware machine learning, providing implementations of algorithmic fairness metrics and mitigation methods. → GitHub

• FairTest: A tool for discovering unwarranted associations between an algorithm's outputs and the inputs it was trained on. FairTest enables developers to test their models for bias and discrimination → GitHub

• TensorFlow Fairness Indicators: An extension of TensorFlow Model Analysis that provides metrics and plots to evaluate model fairness. It helps in evaluating and improving model performance for fairness criteria → page

We post helpful lists and bite-sized explanations daily on our X (Twitter). Let’s connect!