IBM has recently announced that it is all set to launch 'AI Fairness 360' in order to detect Bias in Artificial Intelligence. It will basically check biases related to algorithms and also try to understand how and why they have crept in. AI Fairness 360 is a comprehensive open-source toolkit of metrics launched by IBM.
It is seen that ML is increasingly being used to make high-stakes decisions related to people. The AIF360 Python package includes nine different algorithms, which are developed by the algorithm fairness research community in order to mitigate the unwanted bias.
AIF360 revolves around:
- Credit Scoring
- Predicting medical expenditures
- Classifying face images by gender
AIF360 seems to be different from the currently available open source efforts.