WO2014105865 SYSTEM AND METHOD FOR PARALLELIZING CONVOLUTIONAL NEURAL NETWORKS

Settings

Stemming

Single Family Member

Include NPL

Feedback

Goto Application

Publication Number WO/2014/105865

Publication Date 03.07.2014

International Application No. PCT/US2013/077606

International Filing Date 23.12.2013

IPC

CPC

View more classifications

View less classifications

Applicants

GOOGLE INC. [US]/[US]

Inventors

KRIZHEVSKY, Alexander
SUTSKEVER, Ilya
HINTON, Geoffrey E.

Agents

PORTNOV, Michael

Priority Data

14/030,938		18.09.2013		US
61/745,717		24.12.2012		US

Publication Language English (en)

Filing Language English (EN)

Designated States

View all

Title

(EN) SYSTEM AND METHOD FOR PARALLELIZING CONVOLUTIONAL NEURAL NETWORKS

(FR) SYSTÈME ET PROCÉDÉ PERMETTANT DE PARALLÉLISER DES RÉSEAUX NEURONAUX CLASSIQUES

Abstract

(EN) A parallel convolutional neural network is provided. The CNN is implemented by a plurality of convolutional neural networks each on a respective processing node. Each CNN has a plurality of layers. A subset of the layers are interconnected between processing nodes such that activations are fed forward across nodes. The remaining subset is not so interconnected.

(FR) La présente invention a trait à un réseau neuronal classique parallèle. Le CNN est mis en œuvre au moyen d'une pluralité de réseaux neuronaux classiques, chacun se situant sur un nœud de traitement respectif. Chaque CNN est doté d'une pluralité de couches. Un sous-ensemble de couches est interconnecté entre les nœuds de traitement de sorte que des activations sont soumises à une action directe d'un bout à l'autre des nœuds. Le sous-ensemble restant n'est pas interconnecté de la sorte.

G	PHYSICS
06	COMPUTING; CALCULATING OR COUNTING
N	COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3	Computer systems based on biological models
02	using neural network models
06	Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
063	using electronic means

G	PHYSICS
06	COMPUTING; CALCULATING OR COUNTING
N	COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3	Computer systems based on biological models
02	using neural network models
06	Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
063	using electronic means

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
K	RECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9	Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62	Methods or arrangements for recognition using electronic means
6217	Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
6256	Obtaining sets of training patterns; Bootstrap methods, e.g. bagging, boosting

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
N	COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3	Computer systems based on biological models
02	using neural network models
04	Architectures, e.g. interconnection topology

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
N	COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3	Computer systems based on biological models
02	using neural network models
04	Architectures, e.g. interconnection topology
0454	using a combination of multiple neural nets

Processing

National Phase Entries

Authority File

Settings

Feedback

Goto Application

1. WO2014105865 - SYSTEM AND METHOD FOR PARALLELIZING CONVOLUTIONAL NEURAL NETWORKS

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
T	IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
1	General purpose image data processing
20	Processor architectures; Processor configuration, e.g. pipelining