"Accelerating AI Tasks While Preserving Data Security"

Researchers at the Massachusetts Institute of Technology (MIT) have developed a search engine called SecureLoop, capable of efficiently identifying optimal designs for deep neural network accelerators that preserve data security while improving performance. As computationally intensive Machine Learning (ML) applications, such as chatbots that perform real-time language translation, rise, device manufacturers often incorporate specialized hardware components to quickly move and process the enormous amounts of data demanded by these systems. Selecting the optimal design for these components, known as deep neural network accelerators, is difficult due to the large variety of possible designs. When a designer adds cryptographic operations to keep data secure from attackers, this problem becomes even more complicated. The SecureLoop tool is designed to consider how adding data encryption and authentication measures will affect the performance and energy consumption of the accelerator chip. This article continues to discuss the SecureLoop search tool developed by MIT researchers. 

MIT News reports "Accelerating AI Tasks While Preserving Data Security"

Submitted by grigby1