US Patent Issued to Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung on Sept. 15 for "Methods and apparatuses for compressing parameters of neural networks" (German Inventors)

ALEXANDRIA, Va., Sept. 15 -- United States Patent no. 12,737,586, issued on Sept. 15, was assigned to Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V. (Munich).

"Methods and apparatuses for compressing parameters of neural networks" was invented by Paul Haase (Berlin), Arturo Marban Gonzalez (Berlin), Heiner Kirchhoffer (Berlin), Talmaj Marinc (Berlin), Detlev Marpe (Berlin), Stefan Matlage (Berlin), David Neumann (Berlin), Hoang Tung Nguyen (Berlin), Wojciech Samek (Berlin), Thomas Schierl (Berlin), Heiko Schwarz (Berlin), Simon Wiedemann (Berlin) and Thomas Wiegand (Berlin).

According to the abstract* released by the U.S. Patent & Trademark Office: "An encoder for encoding weight parameters of a neural network is co...