CryptoCredit: Securely Training Fair Models
This addresses fairness in regulated decision-making for model developers, though it is incremental as it builds on existing encryption methods.
The paper tackles the problem of training fair models without exposing sensitive features by using fully homomorphic encryption, enabling bias testing on encrypted data and demonstrating practicality on the adult income dataset.
When developing models for regulated decision making, sensitive features like age, race and gender cannot be used and must be obscured from model developers to prevent bias. However, the remaining features still need to be tested for correlation with sensitive features, which can only be done with the knowledge of those features. We resolve this dilemma using a fully homomorphic encryption scheme, allowing model developers to train linear regression and logistic regression models and test them for possible bias without ever revealing the sensitive features in the clear. We demonstrate how it can be applied to leave-one-out regression testing, and show using the adult income data set that our method is practical to run.