Fairness assessment and improvement in Python
There are few widely used Python libraries to assess fairness in your models (Table 7.3). You can use these libraries to identify if the model satisfies fairness definitions according to the different sensitive attributes in a dataset you want to or have used for modeling:
Library |
Library Name for Importing and Installation |
URL |
IBM AI Fairness 360 |
|
|
Fairlearn |
|
|
Black Box Auditing |
|
|
Aequitas |
|