ALEXANDRIA, Va., July 21 -- United States Patent no. 12,687,965, issued on July 21, was assigned to Marvell Asia Pte Ltd (Singapore).

"Method and system for approximating activation function with floating point format type using a table" was invented by Ulf Hanebutte (Gig Harbor, Wash.).

According to the abstract* released by the U.S. Patent & Trademark Office: "A method includes storing a first/second set of biased exponents associated with a first/second plurality of input floating point data respectively in a memory. The first set of biased exponents defines a positive piecewise linear range associated with a positive piecewise linear approximation region of a function. The second set of biased exponents defines a negative piecewise li...