Jan – May 2021

• I learnt about concepts of fairness and bias in Machine Learning and Artificial Intelligence, in context of US Law and regulations, while completing the projects for this class.
• Specifically, I learnt how to :
  • identify different types of biases,
  • use fairness metrics to evaluate fairness for a given dataset, and
  • apply bias mitigation techniques.

• Tools used: pandas, sklearn, Python, IBM’s Fairness360, Google’s What if tool

• Knowledge gained:
  • Laws, terminology: Protected classes, regulated domains,
  • Fairness metrics: Disparate impact, statistical parity difference,
  • Bias mitigation techniques: disparate impact remover, reweighing.