ECS 189G, Spring 23: Fairness in Machine Learning

Are machine learning (ML) algorithms biased against minorities and women?

• A 2016 article investigated COMPAS, an ML algorithm designed to predict recidivism by those convicted of crimes. COMPAS was being used by judges as an aid in sentencing convicts.
• Actually, many ML applications have been found to have serious biases, and many remedies have been sought and proposed.
• Our course will cover tools to measure and correct ML algorithmic bias.