Against Membership Inference Attack: Pruning is All You Need

Authors: Yijue Wang, Chenghong Wang, Zigeng Wang, Shanglin Zhou, Hang Liu, Jinbo Bi, Caiwen Ding, Sanguthevar Rajasekaran

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

Reproducibility Variable Result LLM Response
Research Type Experimental We also verify our theoretical insights with experiments. Our experimental results illustrate that the attack accuracy using model compression is up to 13.6% and 10% lower than that of the baseline and Min-Max game, accordingly.
Researcher Affiliation Academia Yijue Wang1 , Chenghong Wang2 , Zigeng Wang1 , Shanglin Zhou1 , Hang Liu 3 , Jinbo Bi 1 , Caiwen Ding 1 , Sanguthevar Rajasekaran 1 1University of Connecticut 2Duke University 3Stevens Institute of Technology
Pseudocode Yes Algorithm 1 The Process of MIA-Pruning
Open Source Code No The paper does not explicitly state that the source code for the described methodology is publicly available, nor does it provide a link to a code repository.
Open Datasets Yes We use Le Net-5 on MNIST dataset, and VGG-16, Mobile Net V2 and Res Net18 to classify CIFAR-10 and CIFAR-100 dataset. We also use Mobile Net V2 and Res Net-18 models on the Image Net dataset to show the scalability of our proposed method.
Dataset Splits Yes The empirical gain can be calculated by simply sampling data from the training set and validation set.
Hardware Specification No The paper does not provide specific details about the hardware used to run the experiments (e.g., GPU models, CPU types, or memory specifications).
Software Dependencies No The paper does not specify version numbers for any software dependencies or libraries used in the experiments.
Experiment Setup No The detailed setting of training can be found in Appendix Section 4. The detail of pruning rate settings is in Appendix Section 6.