Search for: All records

Social media applications have benefited users in several ways, including ease of communication and quick access to information. However, they have also introduced several privacy and safety risks. These risks are particularly concerning in the context of interpersonal attacks, which are carried out by abusive friends, family members, intimate partners, co-workers, or even strangers. Evidence shows interpersonal attackers regularly exploit social media platforms to harass and spy on their targets. To help protect targets from such attacks, social media platforms have introduced several privacy and safety features. However, it is unclear how effective they are against interpersonal threats. In this work, we analyzed ten popular social media applications, identifying 100 unique privacy and safety features that provide controls across eight categories: discoverability, visibility, saving and sharing, interaction, self-censorship, content moderation, transparency, and reporting. We simulated 59 different attack actions by a persistent attacker — aimed at account discovery, information gathering, non-consensual sharing, and harassment — and found many were successful. Based on our findings, we proposed improvements to mitigate these risks. 

Secure multi-party computation (MPC) techniques can be used to provide data privacy when users query deep neural network (DNN) models hosted on a public cloud. State-of-the-art MPC techniques can be directly leveraged for DNN models that use simple activation functions (AFs) such as ReLU. However, these techniques are ineffective and/or inefficient for the complex and highly non-linear AFs used in cutting-edge DNN models. We present Compact, which produces piece-wise polynomial approximations of complex AFs to enable their efficient use with state-of-the-art MPC techniques. Compact neither requires nor imposes any restriction on model training and results in near-identical model accuracy. To achieve this, we design Compact with input density awareness, and use an application specific simulated annealing type optimization to generate computationally more efficient approximations of complex AFs. We extensively evaluate Compact on four different machine-learning tasks with DNN architectures that use popular complex AFs silu, gelu, and mish. Our experimental results show that Compact incurs negligible accuracy loss while being 2x-5x computationally more efficient than state-of-the-art approaches for DNN models with large number of hidden layers. Our work accelerates easy adoption of MPC techniques to provide user data privacy even when the queried DNN models consist of a number of hidden layers, and trained over complex AFs.