US Patent Issued to International Business Machines on Aug. 12 for "Multi-speaker data augmentation for improved end-to-end automatic speech recognition" (New York, Connecticut Inventors)

ALEXANDRIA, Va., Aug. 12 -- United States Patent no. 12,387,717, issued on Aug. 12, was assigned to International Business Machines Corp. (Armonk, N.Y.).

"Multi-speaker data augmentation for improved end-to-end automatic speech recognition" was invented by Samuel Thomas (White Plains, N.Y.), Hong-Kwang Kuo (Pleasantville, N.Y.), George Andrei Saon (Stamford, Conn.) and Brian E. D. Kingsbury (Cortlandt Manor, N.Y.).

According to the abstract* released by the U.S. Patent & Trademark Office: "Features of two or more single speaker utterances are concatenated together and corresponding labels of the two or more single speaker utterances are concatenated together. Single speaker acoustic embeddings for each of the single speaker utterances of ...