Scaling up the Banded Matrix Factorization Mechanism for Differentially Private ML
This work addresses scalability issues for practitioners using differentially private machine learning in large-scale settings, representing an incremental improvement.
The paper tackled the scalability limitations of DP-BandMF, a state-of-the-art differentially private mechanism, by developing techniques to handle large-scale training with over 10^4 iterations and 10^7 parameters, achieving negligible utility degradation.
Correlated noise mechanisms such as DP Matrix Factorization (DP-MF) have proven to be effective alternatives to DP-SGD in large-epsilon few-epoch training regimes. Significant work has been done to find the best correlated noise strategies, and the current state-of-the-art approach is DP-BandMF, which optimally balances the benefits of privacy amplification and noise correlation. Despite it's utility advantages, severe scalability limitations prevent this mechanism from handling large-scale training scenarios where the number of training iterations may exceed $10^4$ and the number of model parameters may exceed $10^7$. In this work, we present techniques to scale up DP-BandMF along these two dimensions, significantly extending it's reach and enabling it to handle settings with virtually any number of model parameters and training iterations, with negligible utility degradation.