Balancing behavioral privacy and information utility in pervasive sensing

Sensors of diverse capabilities and modalities, carried by us or deeply embedded in the physical world, have invaded our personal, social, work, and urban spaces. A big enabler have been mobile phones that we always carry and which have diverse sensors embedded in them or wirelessly connected to them. Our relationship with these sensors is a complicated one. On the one hand, these sensors collect rich data that are shared and disseminated, often initiated by us, with a broad array of service providers, interest groups, friends, and family. Embedded in this data is information that can be used to algorithmically construct a virtual biography of our activities, revealing intimate behaviors and lifestyle patterns. On the other hand, we increasingly depend directly and indirectly on information originating from these sensors for making a variety of decisions, both routine and critical, in our lives. The quality of these decisions and our confidence in them depend directly on the quality of the sensory information and our trust in the sources. Drawing upon examples of this two-faceted relationship with sensors from our work in applications such as mobile health and sustainable buildings, this talk will discuss the challenges inherent in designing a sensor information flow and processing architecture that is sensitive to the concerns of both producers and consumers. For the pervasive sensing infrastructure to be trusted by both, it must support a privacy-utility trade-off that balances the tension between risk of information sharing to the producer and the value of information sharing to the consumer. The talk will describe an approach to enabling this tradeoff based on interpreting privacy and utility in terms of inferences that are allowed or prohibited, and supported by mechanisms for controlled transformation of the quality of information via selective obfuscation driven by statistical models of individual behaviors, and robust fusion of information of uncertain quality and obfuscation from sensors of varied trustworthiness. The design of a contextaware privacy-preserving sensing stack for Android-based mobile phones will also be presented.