Membership Inference Attacks Against Machine Learning Models

TLDR

Membership inference attacks aim to determine whether a specific data record was used to train a black‑box model. The study quantitatively investigates how machine learning models leak individual training data and examines factors affecting leakage and mitigation strategies. They train inference models that detect prediction differences between training and non‑training inputs, and evaluate these techniques on classification models from Google and Amazon. The experiments demonstrate that commercial classification models, including those trained on sensitive hospital discharge data, are vulnerable to membership inference attacks.

Abstract

We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on. We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.