Patent · US Active

Training machine learning models by determining update rules using recurrent neural networks

US11615310B2 · kind B2 · utility

0Cited by

0References

17Claims

0Family size

Assignee

DeepMind Technologies Limited · GB

Inventors

Misha Man Ray Denil · London, GB
Tom Schaul · London, GB
Marcin Andrychowicz · London, GB
Joao Ferdinando Gomes de Freitas · London, GB
Sergio Gomez Colmenarejo · London, GB
Matthew William Hoffman · London, GB
David Benjamin Pfau · London, GB

Key dates

Filing date	May 19, 2017
Grant date	Mar 28, 2023
Priority date	—
Expiry date	Oct 20, 2039

Classification

Technology area (CPC G)Physics
CPC primaryG06N3/0985
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

Methods, systems, and apparatus, including computer programs encoded on computer storage media for training machine learning models. One method includes obtaining a machine learning model, wherein the machine learning model comprises one or more model parameters, and the machine learning model is trained using gradient descent techniques to optimize an objective function; determining an update rule for the model parameters using a recurrent neural network (RNN); and applying a determined update rule for a final time step in a sequence of multiple time steps to the model parameters.

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