Page 747

Ming Ding | Information Security and

Privacy Group | Data61, CSIROPrivacy risk assessment –Examples

•Re-identification risk: Pr𝑥𝑖belongstoanindividual ℎ|𝑦,𝐵

➢Intuition: The probability of determining the 𝑖-threcord belongs to an individual ℎ

➢𝐵denotes the background information held by an adversary. Could be anything!

➢In our discussed examples, 𝑦could be a table of data values, an image downloaded from Google

street view, cross -correlation heatmaps of time series sequences, etc.

➢This metric can be evaluated using Probability theory or empirical experiments

➢The smaller the metric, the stronger the privacy protection

Original data x

(sensitive info: s

non-sensitive info: u)Output/Observation y

(data, query answer, AI/ML

models/parameters, etc.)