Online Testing of User Profile Resilience Against Inference Attacks in Social Networks

Abstract

International audience ; To increase awareness about privacy threats, we have designed a tool, SONSAI, for Facebook users to audit their own profiles. SONSAI predicts values of sensitive attributes by machine learning and identifies user public attributes that have guided the learning algorithm towards these sensitive attribute values. Here, we present new aspects of the system such as the automatic combination of link disclosure attacks and attribute prediction. We explain how we defined sensitive subjects from a survey. We also show how the extended tool is fully interfaced with Facebook along different scenarios. In each case a dataset was built from real profiles collected in the user neighbourhood network. The whole analysis process is performed online, mostly automatically and with accuracy of 0.79 in AUC when inferring the political orientation.