ALEXANDRIA, Va., June 10 -- United States Patent no. 12,293,298, issued on May 6, was assigned to Baidu USA LLC (Sunnyvale, Calif.).
"Reducing training times of deep neural networks through efficient hybrid parallelism" was invented by Venmugil Elango (Redmond, Wash.).
According to the abstract* released by the U.S. Patent & Trademark Office: "Presented are systems and methods to automatically find efficient parallelization strategies for deep neural networks (DNNs). A computation graph comprising an efficiently ordered sequence of vertices aids in computing the best parallelizing strategy in a relatively short time. Effectiveness of the parallelization strategies is evaluated on various DNNs, and the performance of the strategies propose...