Google AI Introduces an Open Source Machine Learning Library for Auditing Differential Privacy Guarantees with only Black-Box Access to a Mechanism

Google researchers address the challenge of maintaining the correctness of differentially private (DP) mechanisms by introducing a large-scale library for auditing differential privacy, DP-Auditorium. Differential privacy is essential for protecting data privacy with upcoming regulations and increased awareness of data privacy issues. Verifying a mechanism for its ability to uphold differential privacy in a complex […]

The post Google AI Introduces an Open Source Machine Learning Library for Auditing Differential Privacy Guarantees with only Black-Box Access to a Mechanism appeared …

ai shorts applications artificial intelligence box challenge data data privacy differential differential privacy editors pick google library machine machine learning open source open source machine learning privacy regulations researchers scale staff tech news technology