US Patent Issued to OpenAI OpCo on June 30 for "Utilizing a multi-layer security architecture for agent reinforcement fine-tuning of a machine learning model" (California Inventors)

ALEXANDRIA, Va., July 16 -- United States Patent no. 12,671,713, issued on June 30, was assigned to OpenAI OpCo LLC (San Francisco).

"Utilizing a multi-layer security architecture for agent reinforcement fine-tuning of a machine learning model" was invented by John Allard (San Francisco), James Blomo (San Carlos, Calif.), Filipe de Avila Belbute Peres (San Francisco), William Hang (San Francisco), Joseph Palermo (San Francisco), Chaitanya Ravuri (Saratoga, Calif.), Theophile Sautory (San Francisco), Karan Sharma (San Francisco), Beining Zhou (Stanford, Calif.) and Wenjie Zi (San Francisco).

According to the abstract* released by the U.S. Patent & Trademark Office: "The present disclosure relates to systems, non-transitory computer-readabl...