US Patent Issued to GDM Holding on Feb. 3 for "Reduced computation real time recurrent learning" (British, American Inventors)

ALEXANDRIA, Va., Feb. 3 -- United States Patent no. 12,541,680, issued on Feb. 3, was assigned to GDM Holding LLC (Mountain View, Calif.).

"Reduced computation real time recurrent learning" was invented by Jacob Lee Menick (London), Erich Konrad Elsen (San Francisco) and Karen Simonyan (London).

According to the abstract* released by the U.S. Patent & Trademark Office: "A computer-implemented method for training a recurrent neural network using forward propagation rather than back propagation through time. The method is particularly suited to training sparse recurrent neural networks, and may be implemented on specialized hardware."

The patent was filed on Feb. 5, 2021, under Application No. 17/169,083.

*For further information, includi...