US Patent Issued to Microsoft Technology Licensing on Feb. 10 for "Training a machine learning model using an acceleration pipeline with popular and non-popular micro-batches" (Washington Inventor)

ALEXANDRIA, Va., Feb. 11 -- United States Patent no. 12,547,469, issued on Feb. 10, was assigned to Microsoft Technology Licensing LLC (Redmond, Wash.).

"Training a machine learning model using an acceleration pipeline with popular and non-popular micro-batches" was invented by Divya Mahajan (Seattle).

According to the abstract* released by the U.S. Patent & Trademark Office: "Innovations in training a machine learning model using an accelerator pipeline with deferred updates to model parameters are described. An accelerator identifies one or more first-class micro-batches ("MBs") and a second-class MB of a working set. A first-class MB contains, as inputs, frequently accessed embeddings stored in graphics processing unit ("GPU") memory. ...