Private Machine Learning via Randomised Response
This addresses privacy concerns in machine learning for scenarios where data must be protected from all parties, though it appears incremental as it applies an existing privacy technique to a specific model.
The paper tackles the problem of private machine learning under adversarial actors by introducing a framework based on randomized response, which allows for consistent estimation of the true model, specifically demonstrated with logistic regression.
We introduce a general learning framework for private machine learning based on randomised response. Our assumption is that all actors are potentially adversarial and as such we trust only to release a single noisy version of an individual's datapoint. We discuss a general approach that forms a consistent way to estimate the true underlying machine learning model and demonstrate this in the case of logistic regression.