US Patent Issued to NVIDIA on Nov. 3 for "Concurrent neural network training using communication queues" (American, German Inventors)

ALEXANDRIA, Va., Nov. 3 -- United States Patent no. 12,783,891, issued on Nov. 3, was assigned to NVIDIA Corp. (Santa Clara, Calif.).

"Concurrent neural network training using communication queues" was invented by Sangkug Lym (San Jose, Calif.), Nitin Nitin (Oakland, Calif.) and Michael Andersch (Potsdam, Germany).

According to the abstract* released by the U.S. Patent & Trademark Office: "Apparatuses, systems, and techniques to perform neural networks. In at least one embodiment, a processor comprising two or more processing cores is used to cause two or more neural network portions to be trained concurrently. In at least one embodiment, the two or more neural network portions are trained concurrently based, at least in part, on two or m...