Holistic AI Library
Welcome to the Holistic AI library! This is an open-source tool to assess and improve the trustworthiness of AI systems. The current version of the library offers a set of techniques to easily measure and mitigate Bias across a variety of tasks.
Our long-term goal is to offer a set of techniques to easily measure and mitigate AI risks across five areas: Bias, Efficacy, Robustness, Privacy and Explainability. This will allow a holistic assessment of AI systems.
You can install the library with pip using the following command:
pip install holisticai # basic installation pip install holisticai[bias] # bias mitigation support pip install holisticai[all] # install all packages for bias and explainability