Patent · US Active

Floating-point unit stochastic rounding for accelerated deep learning

US11449574B2 · kind B2 · utility

4Cited by
7References
37Claims
0Family size

Assignee

Inventors

Key dates

Filing dateApr 13, 2018
Grant dateSep 20, 2022
Priority date
Expiry dateNov 29, 2038

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06N3/084
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Techniques in advanced deep learning provide improvements in one or more of accuracy, performance, and energy efficiency. An array of processing elements comprising a portion of a neural network accelerator performs flow-based computations on wavelets of data. Each processing element has a respective compute element and a respective routing element. Each compute element has a respective floating-point unit enabled to perform stochastic rounding, thus in some circumstances enabling reducing systematic bias in long dependency chains of floating-point computations. The long dependency chains of floating-point computations are performed, e.g., to train a neural network or to perform inference with respect to a trained neural network.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.