Scientific Foundations

This project aims to bridge the gap between theory and practice in privacy-preserving data sharing and analysis. Data collected by organizations and agencies are a key resource in today's information age. However, the disclosure of those data poses serious threats to individual privacy. While differential privacy provides a solid foundation for developing techniques to balance privacy and utility in data sharing, currently there is a significant gap between theory and practice in research in this area.

Users entrust their mobile devices with sensitive data, including business emails, as well as health and financial information. Thus, mobile devices have become an increasingly popular target for attackers. Mobile devices house powerful browsers that are vulnerable to at least as many attacks as their desktop counterparts. Yet, the security of these mobile browsers is understudied by researchers, leading to a lack of current information about ongoing attacks and possible defenses.

This project develops methods to provide citizens information about technologies that obstruct, restrict, or tamper with their access to information. Internet users need an objective, independent, third-party service that helps them determine whether their Internet service provider or government is restricting access to content, specific protocols, or otherwise degrading service. Towards this goal, we are (1) monitoring attempts to block or manipulate Internet content and communications; and (2) evaluating various censorship circumvention mechanisms in real-world deployments}.

The project explores building support for malware detection in hardware. Malware detection is challenging and resource intensive, as the number and sophistication of malware increases. The resource requirements for malware detection limit its use in practice, leaving malware unchecked on many systems. We use a low level hardware detector to identify malware as a computational anomaly using low level features such as hardware events, instruction mixes and memory address patterns.

Secure multi-party computation (SMC) allows several mutually untrusting parties to perform joint computations while keeping their inputs private. This project develops new techniques for constructing two-party secure computation protocols with low communication overhead. Building on the PIs prior work for constructing special-purpose secure computation protocols for greedy algorithms, this project aims to develop new techniques that exploit the algorithmic structure of a function in order to develop more efficient secure computation protocols.