主题：Preserving Privacy in the Smart Grid
时间：2018年7月6日（星期五）上午10：40
地点：学术会议中心 301
主办单位：计算机与网络安全学院、无线智能网络实验室
报告人简介：
Dr. Yuan Hong is an Assistant Professor in the Department of Computer Science at Illinois Institute of Technology. Prior to joining Illinois Tech, he was an Assistant Professor in the State University of New York (SUNY) at Albany, and received his Ph.D. degree from Rutgers, the State University of New Jersey. His research interests primarily lie in the fields of privacy, security, optimization and data analytics. As such, he is very interested in resolving the security and privacy issues in both fundamental problems (e.g., optimization models) and data intensive systems (e.g., the smart grid, web search, intelligent transportation systems, and data mining). His work has been supported by the NSF.
报告内容摘要：
The smart grid integrates sensors and communication networks into the existing power grid to ubiquitously collect data from the grid for operational intelligence. However, the collection, storage and analysis of such massive amount of data may compromise the privacy of various entities on the power grid, such as energy consumers and microgrids (which refer to small segments of the grid that can locally generate and consume energy).
In this talk, I will present two categories of privacy preserving schemes to quantitatively measure and bound the privacy risks in two different applications in the smart grid infrastructure: (1) energy sharing among microgrids, and (2) streaming smart meter readings for consumers. Specifically, the former scheme is proposed under the theory of secure multiparty computation (SMC) to ensure secure cooperation among multiple microgrids (i.e., globally optimizing the energy sharing with their local energy) without disclosing sensitive information to each other. In the latter scheme (for streaming smart meter readings), we define a formal privacy notion to bound the risks of inferring electric appliances’ ON/OFF status at specific times from smart meter reading streams. Then, I will present efficient algorithms to convert and stream real-time meter readings with low errors while satisfying the privacy notion.
