US Patent Issued to Microsoft Technology Licensing on June 16 for "Training machine-trained models by directly specifying gradient elements" (Washington Inventors)

ALEXANDRIA, Va., June 16 -- United States Patent no. 12,657,469, issued on June 16, was assigned to Microsoft Technology Licensing LLC (Redmond, Wash.).

"Training machine-trained models by directly specifying gradient elements" was invented by Hong Xuan (Bellevue, Wash.), Xi Chen (Bellevue, Wash.), Saurajit Mukherjee (Kirkland, Wash.), Li Huang (Sammamish, Wash.), Kun Wu (Bellevue, Wash.), Arun Kumar Sacheti (Sammamish, Wash.), Kamal Ginotra (Kirkland, Wash.) and Meenaz Aliraza Merchant (Kirkland, Wash.).

According to the abstract* released by the U.S. Patent & Trademark Office: "A computer-implemented technique performs machine learning that bypasses the traditional design of loss functions. The technique includes receiving plural instan...