In this directory, we give a simple example of how to use the C++ Differential Privacy library.
There are around 200 animals at Farmer Fred's zoo. Every day, Farmer Fred feeds the animals as many carrots as they desire. The animals record how many carrots they have eaten per day. For this particular day, the number of carrots eaten can be seen in animals_and_carrots.csv.
At the end of each day, Farmer Fred often asks aggregate question about how many carrots everyone ate. For example, he wants to know how many carrots are eaten each day, so he knows how many to order the next day. The animals are fearful that Fred will use the data against their best interest. For example, Fred could get rid of the animals who eat the most carrots!
To protect themselves, the animals decide to use the C++ Differential Privacy library to aggregate their data before reporting it to Fred. This way, the animals can control the risk that Fred will identify individuals' data while maintaining an adequate level of accuracy so that Fred can continue to run the zoo effectively.
The animals have implemented a CarrotReporter tool in animals_and_carrots.h to obtain DP aggregate data to report to Fred. We document one of these reports in report_the_carrots.cc.
Each row in animals_and_carrots.csv is composed of the name of an animal, and the number of carrots it has eaten, comma-separated.
Notice that each animal owns at most one row in the data. This means that we provide per-animal privacy. Suppose that some animal appears multiple times in the csv file. That animal would own more than one row in the data. In this case, using this DP library would not guarantee per-animal privacy! The animals would first have to pre-process their data in a way such that each animal doesn't own more than one row.
If Farmer Fred continues to query the carrot data many times in a given day, then he can draw conclusions about the data that breaks privacy. For example, asking about the mean number of carrots eaten repeatedly would yield a distribution centered about the true mean. To prevent this, the animals have implemented a privacy budget tracking feature in CarrotReporter. The budget is a fraction; at the beginning of the day, the budget is 1. Each time the animals answer one of Fred‘s questions, the privacy budget decreases. When the privacy budget is no longer positive, the animals refuse to answer any more of Fred’s queries.
$ cd examples/cc $ bazel run report_the_carrots