Unleashing the Power of Randomization in Auditing Differentially Private ML
This work addresses the challenge of verifying privacy guarantees in ML for practitioners, though it is incremental by building on existing canary-based auditing methods.
The authors tackled the problem of auditing differentially private machine learning algorithms by introducing a methodology based on randomized canaries and Lifted Differential Privacy, achieving significant improvements in sample complexity both theoretically and empirically.
We present a rigorous methodology for auditing differentially private machine learning algorithms by adding multiple carefully designed examples called canaries. We take a first principles approach based on three key components. First, we introduce Lifted Differential Privacy (LiDP) that expands the definition of differential privacy to handle randomized datasets. This gives us the freedom to design randomized canaries. Second, we audit LiDP by trying to distinguish between the model trained with $K$ canaries versus $K - 1$ canaries in the dataset, leaving one canary out. By drawing the canaries i.i.d., LiDP can leverage the symmetry in the design and reuse each privately trained model to run multiple statistical tests, one for each canary. Third, we introduce novel confidence intervals that take advantage of the multiple test statistics by adapting to the empirical higher-order correlations. Together, this new recipe demonstrates significant improvements in sample complexity, both theoretically and empirically, using synthetic and real data. Further, recent advances in designing stronger canaries can be readily incorporated into the new framework.