Processing

Please wait...

Settings

Settings

Goto Application

1. WO2020064209 - MACHINE LEARNING SYSTEM AND A METHOD, A COMPUTER PROGRAM AND A DEVICE FOR CREATING THE MACHINE LEARNING SYSTEM

Publication Number WO/2020/064209
Publication Date 02.04.2020
International Application No. PCT/EP2019/071761
International Filing Date 13.08.2019
IPC
G06N 3/04 2006.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architecture, e.g. interconnection topology
G06N 3/08 2006.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
G06N 3/063 2006.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
06Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
063using electronic means
CPC
G06N 20/00
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
20Machine learning
G06N 3/04
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
G06N 3/0445
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0445Feedback networks, e.g. hopfield nets, associative networks
G06N 3/0481
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0481Non-linear activation functions, e.g. sigmoids, thresholds
G06N 3/063
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
06Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
063using electronic means
G06N 3/08
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
Applicants
  • ROBERT BOSCH GMBH [DE]/[DE]
Inventors
  • PFEIL, Thomas
Priority Data
10 2018 216 471.126.09.2018DE
10 2018 220 608.229.11.2018DE
Publication Language German (DE)
Filing Language German (DE)
Designated States
Title
(DE) MASCHINELLES LERNSYSTEM, SOWIE EIN VERFAHREN, EIN COMPUTERPROGRAMM UND EINE VORRICHTUNG ZUM ERSTELLEN DES MASCHINELLEN LERNSYSTEMS
(EN) MACHINE LEARNING SYSTEM AND A METHOD, A COMPUTER PROGRAM AND A DEVICE FOR CREATING THE MACHINE LEARNING SYSTEM
(FR) SYSTÈME D'APPRENTISSAGE AUTOMATIQUE AINSI QUE PROCÉDÉ, PROGRAMME INFORMATIQUE ET DISPOSITIF POUR CRÉER LE SYSTÈME D'APPRENTISSAGE AUTOMATIQUE
Abstract
(DE)
Die Erfindung betrifft ein Maschinelles Lernsystem (12), insbesondere ein tiefes neuronales Netz (20). Das maschinelle Lernsystem (12) umfasst eine Mehrzahl von Schichten, die miteinander verbunden sind. Die Schichten ermitteln jeweils abhängig von einer Eingangsgröße und zumindest einem Parameter, der in einem Speicher (200) hinterlegt ist, eine Ausgangsgröße. Die Parameter derjenigen Schichten, die mit einer weiteren, insbesondere vorhergehenden, Schicht verbunden sind, sind jeweils mittels einer höheren Auflösung als die Parameter derjenigen Schichten, die mit einer Mehrzahl von weiteren, insbesondere vorhergehenden, Schichtenverbunden sind, in dem Speicher (200) hinterlegt. Die Erfindung betrifft ferner ein Verfahren, ein Computerprogramm, sowie eine Vorrichtung zum Erstellen des maschinellen Lernsystems (12).
(EN)
The invention relates to a machine learning system (12), particularly a deep neural network (20). The machine learning system (12) comprises a plurality of layers that are connected to one another. The layers determine an output variable, always dependent on an input variable and at least one parameter stored in a memory (200). The parameters of those layers which are connected to a further, particularly preceding, layer, are each stored in the memory (200) at a higher resolution than the parameters of those layers that are connected to a plurality of further, particularly preceding, layers. The invention also relates to a method, a computer program, and a device for creating the machine learning system (12).
(FR)
L'invention concerne un système d'apprentissage automatique (12), notamment un réseau neuronal profond (20). Le système d'apprentissage automatique (12) comprend une pluralité de couches qui sont reliées entre elles. Les couches déterminent une grandeur de sortie, respectivement en fonction d'une grandeur d'entrée et d'au moins un paramètre stocké dans une mémoire (200). Les paramètres des couches qui sont reliées à une autre couche, notamment à la couche précédente, sont stockés dans la mémoire (200) respectivement avec une résolution plus élevée que les paramètres des couches qui sont reliées à plusieurs autres couches, notamment aux couches précédentes. L'invention concerne en outre un procédé, un programme informatique et un dispositif pour créer le système d'apprentissage automatique (12).
Also published as
Latest bibliographic data on file with the International Bureau