Your browser is unsupported

We recommend using the latest version of IE11, Edge, Chrome, Firefox or Safari.

Enhancing privacy-preserving computation with advanced cryptography

Peihan Miao

Just about every day, we share large amounts of personally identifiable data to access a variety of services—with our doctors, on social media, or via an online retailer for our latest purchase. The onus is on the recipients of that information to keep it safe while using it for the computations they need to perform.

Therefore, the need for privacy-preserving computation on big data is greater than ever. Peihan Miao, an assistant professor of computer science, is working to improve functionality and efficiency of these systems by developing new advanced cryptographic tools.

Miao and Mohammad Hajiabadi from Pennsylvania State University are collaborating on two sides of the privacy conundrum. They seek to improve the efficiency, functionality, and security of the underlying cryptographic tools from a theoretical perspective and a practical one.

“We are trying to provide a way to compute on sensitive data while protecting the privacy of the underlying data,” Miao said.

She is working to improve a fundamental cryptography tool called oblivious transfer, a protocol in which one party sends several pieces of information to a recipient but is oblivious about which, if any, are received. Improving its functioning and efficiency would help users to keep up with the demands for privacy created by machine learning applications and imposed by governments, such as the European Union’s General Data Protection Regulation and the California Privacy Rights Act, both of which protect consumers from the transfer of their personal data.

Using these advanced cryptographic tools will open up a lot of applications, as what companies can do now is quite limited. For example, using new privacy-preserving tools, multiple hospitals would have the ability to run analytics or share data with one another through secure multi-party computation.

“Secure multi-party computation enables multiple entities to share their data for more advanced analytics, such as running machine learning or deep learning algorithms, to perform more comprehensive study on the data. This is all done while preserving their data privacy,” Miao said.

One practical problem Miao hopes to enhance through her research is private set intersection, a function that allows two parties each holding a set of private elements to compare encrypted versions of these sets of data, in order to compute the intersection of their private elements.

For example, an advertiser might compare its encrypted data on its campaigns against encrypted data from a website to determine the effectiveness of an approach. Private set intersection also can be used for breached password monitoring, a service offered by Google Chrome, Apple iCloud Keychain, and Microsoft Edge. These platforms keep track of a user’s passwords and would be able to detect when the user’s password was breached—but without learning what the password is.

Miao, who joined UIC in 2020, received a $250,000 National Science Foundation grant for this work, titled Expanding the Realm of Oblivious Transfer: New Tools for Cryptography. The grant period begins June 1. Hajiabadi received a similar grant for his portion of the project. Miao’s research focuses on cryptography and security, including secure multiparty computation, applied cryptography, secure authentication, and blockchain. She is seeking new graduate students to participate in her research.