PRivacy fOr smart cities – PROPOLIS

Cities are experiencing unparalleled growth, facing key challenges as they endeavor to remain healthy, sustainable and safe places for citizens to live and work. Through a variety of devices equipped with sensors, meters and cameras, the Internet of Things (IoT) is bringing new opportunities for cities to leverage data for artificial-intelligence (AI)-based analytics, in contexts as varied as traffic management, use of infrastructure, environmental control and safety. While the collection and processing of massive amounts of citizens’ data is what enables those new opportunities, it is the source of serious privacy concerns.
The overall objective of PROPOLIS is to address those privacy concerns during the entire AI lifecycle, and develop a comprehensive understanding of privacy in a context of smart city analytics. The project will design and implement privacy-preserving solutions for (i) the training phase, where an AI model is trained from the collected data, and for (ii) the posterior inference phase, which leverages the trained models for future queries and inferences. The proposed project will contemplate a variety of potential adversaries and ensure privacy for the different data involved, namely, privacy of the training data (and thus, the citizens who volunteer it), privacy of the query, and finally even privacy of the model. The developed solutions will advance the state of the art on and rely on differential privacy, homomorphic encryption and secure multi-party computation.
The PROPOLIS consortium consists of two partners (one academic, one industrial) from France and two partners (one academic, one industrial) from Germany. Each partner will bring its expertise to address the different analytics and dedicated privacy problems. The academic partners will collaborate with industrial partners to design and develop the proposed privacy preserving data analytics modules. While KIT and Urban Institute will focus on the AI training phase and investigate differentially-private mechanisms for citizens’ privacy protection, EURECOM and SAP will tackle the inference phase (addressing query and model privacy) and explore privacy-preserving inference solutions based on homomorphic encryption and/or secure multi-party computation.
The major synergistic effect of this international collaboration is an end-to-end privacy approach to smart-city applications, where neither intellectual property nor the privacy of citizens will be at risk. The ultimate goal of PROPOLIS is to allow reaching the full potential of smart-cities applications through the IoT, bringing countless benefits to society, economy and in general to citizens’ quality of life.