US Patent Issued to QUALCOMM on March 17 for "Efficient machine learning caching via attention output-based token eviction" (California, New York Inventors)

ALEXANDRIA, Va., March 17 -- United States Patent no. 12,579,063, issued on March 17, was assigned to QUALCOMM Inc. (San Diego).

"Efficient machine learning caching via attention output-based token eviction" was invented by Raghavv Goel (San Diego), Mukul Gagrani (Milpitas, Calif.), Junyoung Park (Palo Alto, Calif.), Dalton James Jones (Cupertino, Calif.), Mingu Lee (San Diego), Wonseok Jeon (San Diego), Matthew James Morse (New York), Matthew Harper Langston (Beacon, N.Y.) and Christopher Lott (San Diego).

According to the abstract* released by the U.S. Patent & Trademark Office: "Certain aspects of the present disclosure provide techniques and apparatus for machine learning. In an example method, an input prompt comprising a set of toke...