PRIVACY PROTECTION IN LEARNING FAIR REPRESENTATIONS

In this paper, we develop a framework to achieve a desir- able trade-off between fairness, inference accuracy and pri- vacy protection in the inference as service scenario. Instead of sending raw data to the cloud, we conduct a random map- ping of the data, which will increase privacy protection and mitigate bias but reduce inference accuracy. To properly ad- dress the trade-off, we formulate an optimization problem to find the optimal transformation map. As the problem is non- convex in general, we develop an iterative algorithm to find the desired map. Numerical examples show that the proposed method has better performance than gradient ascent in the convergence speed, solution quality and algorithm stability.