US Patent Issued to Amazon Technologies on Jan. 20 for "Random token segmentation for training next token prediction models" (California, New York, New Jersey, Washington Inventors)

ALEXANDRIA, Va., Jan. 20 -- United States Patent no. 12,530,527, issued on Jan. 20, was assigned to Amazon Technologies Inc. (Seattle).

"Random token segmentation for training next token prediction models" was invented by Zijian Wang (San Jose, Calif.), Yuchen Tian (Santa Clara, Calif.), Mingyue Shang (Jersey City, N.J.), Praphruetpong Athiwaratkun (Jersey City, N.J.), Ming Tan (Jersey City, N.J.), Parminder Bhatia (Kearny, N.J.), Andrew Oliver Arnold (New York), Ramesh M Nallapati (Frfemont, Calif.), Sudipta Sengupta (Sammamish, Wash.), Bing Xiang (Mount Kisco, N.Y.), Atul Deo (Kirkland, Wash.) and Ankur Deepak Desai (Redmond, Wash.).

According to the abstract* released by the U.S. Patent & Trademark Office: "Random token segmentation ma...