Bypassing the Ambient Dimension: Private SGD with Gradient Subspace Identification

Authors: Yingxue Zhou, Steven Wu, Arindam Banerjee

ICLR 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We empirically evaluate PDP-SGD on training neural networks with two datasets: the MNIST (Le Cun et al., 1998) and Fashion MNIST (Xiao et al., 2017). We compare the performance of PDP-SGD with the baseline DP-SGD for various privacy levels ϵ. In addition, we also explore a heuristic method, i.e., DP-SGD with random projection by replacing the projector with a Rk p Gaussian random projector (Bingham and Mannila, 2001; Blocki et al., 2012). We present the experimental results after discussing the experimental setup. More details and additional results are in Appendix D.
Researcher Affiliation Academia Yingxue Zhou , Zhiwei Steven Wu , Arindam Banerjee Department of Computer Science & Engineering, University of Minnesota School of Computer Science, Carnegie Mellon University. Department of Computer Science, University of Illinois Urbana-Champaign
Pseudocode Yes Algorithm 1 Projected DP-SGD (PDP-SGD)
Open Source Code No The paper does not explicitly provide access to the source code for the methodology described in the paper.
Open Datasets Yes We empirically evaluate PDP-SGD on training neural networks with two datasets: the MNIST (Le Cun et al., 1998) and Fashion MNIST (Xiao et al., 2017).
Dataset Splits No The paper describes training and test sets and a small public dataset but does not explicitly provide details about a separate validation dataset split.
Hardware Specification Yes All experiments have been run on NVIDIA Tesla K40 GPUs.
Software Dependencies No The paper mentions 'Py Torch' but does not specify version numbers for any software dependencies.
Experiment Setup Yes The mini-batch size is set to be 250 for both MNIST and Fashion MNIST. For the step size, we follow the grid search method with search space {0.01, 0.05, 0.1, 0.2} to tune the step size for MNIST and the search space is {0.01, 0.02, 0.05, 0.1, 0.2} for Fashion MNIST. We choose the step size based on the training accuracy at the last epoch. [...] For training, a fixed budget on the number of epochs i.e., 30 is assigned for the each task. [...] We choose gradient clip size to be 1.0 for both datasets.