Patent · US Expired

Method for training neural networks

US7406451B2 · kind B2 · utility

32Cited by

7References

21Claims

0Family size

Assignee

BAYER TECHNOLOGY SERVICES GMBH · DE

Inventors

Thomas Mrziglod · Wittlich, DE
Georg Mogk · Lingen (Ems), DE

Key dates

Filing date	Mar 23, 2004
Grant date	Jul 29, 2008
Priority date	—
Expiry date	Sep 28, 2025

Classification

Technology area (CPC G)Physics
CPC primaryG06N3/09
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

The invention relates to a system and a method for training a number of neural networks, by determining a first training data record, wherein the training data have a particular accuracy, generating a number of second data training records by perturbing the first training data record with a random variable, and training each of the neural networks with one of the training data records. A prognosis and an estimation of the prognosis error can be carried out by means of such a system.

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