quarantine

A pre-release featuring the new exponential mechanism capable of maintaining (epsilon, 0) differential privacy while maximizing a policy function u. This allows for another way of using differential privacy entirely, however, there is no published documentation and code may have bugs. Please wait for the next release for full documentation and checked code.

#Made with ❤️ at 🏠.

who (last stable release)

✅ Last stable release of who (v1.0).
Fixed:

• Sensitivity measurements of the query
  • Fixed str+int concatenation
  • Guaranteed —preserves sensitivity measurements in their true form

who

The first release, featuring the laplace mechanism and the random response mechanism.
Big thanks to privacybook.pdf and @Quantalabs

[Unstable: Dataset numbers are rounded when calculating the query sensitivity which could result in non (epsilon, 0) differential privacy. Use v1.1 instead]