In a blog post for the U.S. National Institute of Standards and Technology, Google Brain Researchers Nicolas Papernot and Abhradeep Thakurta explore how differential privacy can be applied to machine-learning technologies. Papernot and Thakurta make a case for differentially private machine learning algorithms, which they say can be used to "quantify and bound leakage of private information from the learner’s training data." Proper deployment of these algorithms with responsible training can help prevent memorization.
Full Story