Project Overview
Participatory sensing and data surveillance are gradually integrated into an inseparable part of our society. At the same time,
Dynamic Data Driven Applications Systems (DDDAS) has established itself as a transformative paradigm that offers the promise of augmenting the effectiveness of such surveillance systems. Many of the complex and streaming data are personal and highly sensitive to privacy concerns as well as volume issues, and will benefit greatly from privacy enhanced DDDAS capabilities.
The PREDICT project aims to build a holistic framework for PRivacy and security Enhancing Dynamic Information Collection and moniToring using feedback loops.
PREDICT includes the following major research thrusts:
1) Privacy Preserving Data Aggregation with Feedback Control.
Sensitive data streams are aggregated and perturbed to guarantee the state-of-the-art differential privacy for data subjects. Feedback loops are being designed
to dynamically control the aggregation and perturbation, including sampling,
grouping, and perturbation.
2) Dynamic Data Modeling and Uncertainty Quantification. Aggregated
and perturbed data will be injected into applications to augment and correct
the predictive data models. Data integrity will be investigated to understand
the impact of the data perturbation privacy mechanisms.
3) Secure Data Aggregation and Feedback Control without Trusted Aggregator.
Decentralized mechanisms will be
developed to allow data contributors to securely aggregate their data with perturbations and receive feedback
from applications without disclosing additional information to other data contributors or untrusted aggregators.