TensorFlow Differential Privacy
James Vincent (via Dan Masters):
Google has announced a new module for its machine learning framework, TensorFlow, that lets developers improve the privacy of their AI models with just a few lines of extra code.
TensorFlow is one of the most popular tools for building machine learning applications, and it’s used by developers around the world to create programs like text, audio, and image recognition algorithms. With the introduction of TensorFlow Privacy, these developers will be able to safeguard users’ data with a statistical technique known as “differential privacy.”
[…]
There are some downsides to using differential privacy, though. “By masking outliers, it can sometimes remove relevant or interesting data, especially in varied datasets, like those involving language,” says Erlingsson. “Differential privacy literally means that it’s impossible for the system to learn about anything that happens just once in the dataset, and so you have this tension. Do you have to go get more data of a certain type? How relevant or useful are those unique properties in the dataset?”
Just found this incredible paper by @korolova and team: they straight-up reverse engineered Apple’s differential privacy system. They share implementation details and findings about privacy loss in a real-world system, which is key for broader DP adoption.
Previously:
- Learning With Privacy at Scale
- How Apple’s Hardline Privacy Policy Limits Key Features
- What is Differential Privacy?
