"More Efficient Security for Cloud-Based Machine Learning"

MIT researchers have created a new system to strengthen the security of sensitive data used in online neural networks without compromising the speed at which these networks operate. The system, called GAZELLE, combines two encryption techniques known as homomorphic encryption and garbled circuits. This article continues to discuss GAZELLE in relation to its testing, application, techniques, and process.  

MIT News report "More Efficient Security for Cloud-Based Machine Learning"