On the Vulnerability of Data Points under Multiple Membership Inference Attacks and Target Models
This work addresses privacy risks in machine learning by analyzing data point vulnerability to membership inference, though it is incremental as it builds on existing MIA research.
The paper tackles the problem of identifying vulnerable data points under multiple membership inference attacks (MIAs) and target models, defining new metrics that capture this vulnerability and showing that MIA has an inference tendency to some data points despite low overall performance, with attack accuracies ranging from 0.5 to 0.9 across 54 implemented MIAs.
Membership Inference Attacks (MIAs) infer whether a data point is in the training data of a machine learning model. It is a threat while being in the training data is private information of a data point. MIA correctly infers some data points as members or non-members of the training data. Intuitively, data points that MIA accurately detects are vulnerable. Considering those data points may exist in different target models susceptible to multiple MIAs, the vulnerability of data points under multiple MIAs and target models is worth exploring. This paper defines new metrics that can reflect the actual situation of data points' vulnerability and capture vulnerable data points under multiple MIAs and target models. From the analysis, MIA has an inference tendency to some data points despite a low overall inference performance. Additionally, we implement 54 MIAs, whose average attack accuracy ranges from 0.5 to 0.9, to support our analysis with our scalable and flexible platform, Membership Inference Attacks Platform (VMIAP). Furthermore, previous methods are unsuitable for finding vulnerable data points under multiple MIAs and different target models. Finally, we observe that the vulnerability is not characteristic of the data point but related to the MIA and target model.