Data sharing, key to effective research, raises an interesting challenge in the digital era: how does one present data in a way that preserves the privacy or anonymity of the data, yet allows sufficient access to information needed for research? This tradeoff—between accessibility and over-exposure—has numerous implications, notably in such critical sectors as energy and healthcare, where research needs may conflict with privacy concerns.
I3P researchers are exploring this tradeoff with the aid of a model that relates analytical capabilities to privacy parameters. Drawing on a set of ontologies that show real-world inference, such as the link between diagnosis and gender in medical records, the model includes sanitizing agents that adapt over time to changing data.
