公开(公告)号：US20220261650A1

公开(公告)日：2022-08-18

申请号：US17346100

申请日：2021-06-11

Applicant: NVIDIA Corp.

Inventor： Jiawei Zhao ,  Steve Haihang Dai ,  Rangharajan Venkatesan ,  Ming-Yu Liu ,  William James Dally ,  Anima Anandkumar

IPC: G06N3/08 ,  G06F7/483 ,  G06F7/487 ,  G06F5/01 ,  G06F1/03

Abstract: An end-to-end low-precision training system based on a multi-base logarithmic number system and a multiplicative weight update algorithm. The multi-base logarithmic number system is applied to update weights of the neural network, with different bases of the multi-base logarithmic number system utilized between calculation of weight updates, calculation of feed-forward signals, and calculation of feedback signals. The LNS expresses a high dynamic range and computational energy efficiency, making it advantageous for on-board training in energy-constrained edge devices.