Processing

Please wait...

Settings

Settings

Goto Application

1. WO2019214309 - MODEL TEST METHOD AND DEVICE

Publication Number WO/2019/214309
Publication Date 14.11.2019
International Application No. PCT/CN2019/075438
International Filing Date 19.02.2019
IPC
G06N 3/02 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
CPC
G06K 9/62
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
G06K 9/6217
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
G06K 9/6256
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
6256Obtaining sets of training patterns; Bootstrap methods, e.g. bagging, boosting
G06N 20/20
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
20Machine learning
20Ensemble learning
G06N 3/02
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
G06N 3/0454
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
0454using a combination of multiple neural nets
Applicants
  • 创新先进技术有限公司 ADVANCED NEW TECHNOLOGIES CO., LTD.
Inventors
  • 周俊 ZHOU, Jun
Agents
  • 北京博思佳知识产权代理有限公司 BEIJING BESTIPR INTELLECTUAL PROPERTY LAW CORPORATION
Priority Data
201810443821.010.05.2018CN
Publication Language Chinese (zh)
Filing Language Chinese (ZH)
Designated States
Title
(EN) MODEL TEST METHOD AND DEVICE
(FR) PROCÉDÉ ET DISPOSITIF DE TEST DE MODÈLE
(ZH) 模型测试的方法及装置
Abstract
(EN) The embodiments of the present description provide a model test method and device. According to said method, a sample is acquired from a test sample set; next, the sample is inputted into a plurality of models to be tested included in a model set, so as to acquire an output result of each of the models to be tested; then a test result is determined according to the output result; furthermore, if the test result does not satisfy a predetermined condition, a new sample is generated on the basis of said sample according to a predetermined rule, and the generated new sample is added to the test sample set. In this way, when the model test method is performed cyclically, evaluation is performed on the accuracy and/or the test adequacy of models to be detected, and a new sample, which is generated on the basis of an original sample and is different from the original sample, is added to the test sample set, thereby improving effectiveness of model test.
(FR) Selon certains modes de réalisation, la présente invention concerne un procédé et un dispositif de test de modèle. Selon ledit procédé, un échantillon est acquis parmi un ensemble d'échantillons de test; ensuite, l'échantillon est entré dans une pluralité de modèles à tester inclus dans un ensemble de modèles afin d'acquérir un résultat de sortie de chacun des modèles à tester; puis un résultat de test est déterminé selon le résultat de sortie; de plus, si le résultat de test ne satisfait pas une condition prédéfinie, un nouvel échantillon est généré sur la base dudit échantillon selon une règle prédéfinie, et le nouvel échantillon généré est ajouté à l'ensemble d'échantillons de test. De cette façon, lorsque le procédé de test de modèle est réalisé cycliquement, une évaluation est réalisée sur la précision et/ou l'adéquation de test de modèles à détecter, et un nouvel échantillon, qui est généré sur la base d'un échantillon original et qui est différent de l'échantillon original, est ajouté à l'ensemble d'échantillons de test, ce qui améliore l'efficacité pour tester des modèles.
(ZH) 一种模型测试的方法和装置,根据该方法,首先从测试样本集中获取样本,接着将样本输入模型集包括的多个待测试模型中,以获得各个待测试模型的输出结果,然后根据输出结果确定测试结果,进一步地,在该测试结果未满足预定条件的情况下,按照预定规则,基于上述样本生成新样本,并将所生成的新样本加入测试样本集。如此,在该模型测试的方法被循环执行的情况下,一方面对待检测模型的准确性和/或测试充分程度进行评估,另一方面将基于原样本生成的、与原样本具有差异化的新样本加入测试样本集,提高模型测试的有效性。
Latest bibliographic data on file with the International Bureau