CAREER: Enabling Collaborative Computation on Confidential Data

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

A recurrent problem society faces today is that organizations with sensitive data wish to conduct useful global studies on their aggregate data, but they cannot share this data with each other. For example, healthcare organizations often cannot share patient data due to privacy laws or concerns, but would greatly benefit from conducting a global study over everyone's data to identify, for example, the most successful treatments for rare forms of cancer. The PI's goal is to enable collaborative computation on confidential data. Her approach is to have systems "process encrypted data'" instead of plaintext data. Her research enables organizations to carry out collaborative computation without giving away their data: each organization encrypts the data, shares encrypted data, the collaborative computation (either at the organizations or in a cloud) processes and computes on encrypted data producing an encrypted result that only the organizations can decrypt. The overarching challenge is to support such computation efficiently and to express rich functionalities, when existing cryptographic methods are too slow. Using the open-source platform the PI is building, she is collaborating with top Canadian banks to detect money laundering in a collaborative way, as well as healthcare organizations to build better medical models. The PI’s plan includes enhancing several courses and developing a new course based on topics and results of this research, as well as educational outreach activities targeted toward undergraduate and underrepresented groups, and industry. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.