Information levels shape privacy filter and odometer effectiveness
An Information Hierarchy of Privacy Filters and Odometers
Cryptography and Security
Summary
When tracking how much privacy is lost while analyzing data step-by-step, tools called privacy filters and odometers help control and measure this loss. The authors show that the type of information these tools use greatly affects how well they work. They define three levels of information and find that having more detailed information can improve privacy management by a large margin. They also find that randomness in these tools can further boost their performance, independent of information level.
What this means in practice
- •For data privacy engineers: Design adaptive privacy controls that use richer information to allow more accurate privacy loss tracking and improved data utility during multiple analyses.
- •For security system designers: Develop privacy monitoring tools that incorporate randomness and various information levels to better manage cumulative privacy risks in interactive data systems.
A theory result. No direct application yet.
Authors
Sophie Taylor, Praneeth Vippathalla, Justin P. Coon
Abstract
In sequential data analysis, privacy loss accumulates over time, and privacy filters and odometers are used to control and track this cumulative cost. We study how the information used by differential privacy filters and odometers fundamentally limits what they can achieve. We formalise three levels of information: spec-level information, fixed prior to the interaction; mechanism-level information, which includes the mechanisms instantiated during the interaction; and realisation-level information, which further includes the realised interaction. We classify privacy filters and odometers by the least informative level sufficient to determine their behaviour. Our main results establish strict and arbitrarily large separations in achievable performance across this information hierarchy. For privacy filters, moving up the hierarchy can yield an unbounded utility advantage, quantified by a clipped measure of the number of allowed releases. For privacy odometers, it can yield an arbitrarily large improvement in precision, quantified by the expected peak excess of the reported bound over the realised privacy loss. Further, although the information classes are distinct, we show that moving to a richer class need incur essentially no loss in performance. Finally, we show that randomisation provides a resource orthogonal to information: at every information level, it can yield an arbitrarily large utility improvement over deterministic filters. Thus, information access and randomisation can be viewed as fundamental resources governing privacy procedures in adaptive interactions.