AI Fairness 360 is an open-source toolkit for measuring and compensating for data bias in AI.
It looks like its got a lot of potential for us to understand our data/model biases. They have a load of example notebooks on their GitHub demonstrating how we could use it.
There are several different mathematical fairness checking and balancing methods available. Some of these should be added to MLOps.
Previous suggestion of this was made by Laurence Jackson.
"https://ai-fairness-360.org/
AI Fairness 360 is an open-source toolkit for measuring and compensating for data bias in AI.
It looks like its got a lot of potential for us to understand our data/model biases. They have a load of example notebooks on their GitHub demonstrating how we could use it.
e.g. https://github.com/Trusted-AI/AIF360/blob/master/examples/tutorial_credit_scoring.ipynb https://github.com/Trusted-AI/AIF360/blob/master/examples/tutorial_medical_expenditure.ipynb "