Project repo for CS7290
Based on the paper: https://arxiv.org/pdf/1703.06856.pdf
Please follow the tutorial (Counterfactual Fairness Final.ipynb) for a detailed explanation.
Team members: Gowtam Potluri, Oliver Penglase, Kai Wu, Mingfei Zhou
Using the law school success example from the paper, we replicate our version of the simulated example to show the difference in a fair prediction of law school success vs. an unfair one using Pyro.
We will demonstrate fairness evaluation on a full (using all attributes including protected in prediction), unaware (using only non-protected observed attributes), and a fair model (inferring the unobserved variable K to predict).