US Patent Issued to International Business Machines on June 30 for "Reinforced generation: reinforcement learning for text and knowledge graph bi-directional generation using pretrained language models" (New York Inventors)

ALEXANDRIA, Va., July 16 -- United States Patent no. 12,670,340, issued on June 30, was assigned to International Business Machines Corp. (Armonk, N.Y.).

"Reinforced generation: reinforcement learning for text and knowledge graph bi-directional generation using pretrained language models" was invented by Pierre L. Dognin (White Plains, N.Y.), Inkit Padhi (White Plains, N.Y.), Igor Melnyk (White Plains, N.Y.) and Payel Das (Yorktown Heights, N.Y.).

According to the abstract* released by the U.S. Patent & Trademark Office: "Obtain access to a pretrained encoder-decoder language model. Using a dataset including a plurality of text-graph pairs, carry out first fine-tuning training on the pre-trained language model by minimizing cross-entropy ...