Processing

Please wait...

Settings

Settings

Goto Application

1. WO2020141882 - EXPLAINABLE ARTIFICIAL INTELLIGENCE MODELING AND SIMULATION SYSTEM AND METHOD

Publication Number WO/2020/141882
Publication Date 09.07.2020
International Application No. PCT/KR2020/000022
International Filing Date 02.01.2020
IPC
G06F 11/34 2006.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
11Error detection; Error correction; Monitoring
30Monitoring
34Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation
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
G06T 7/00 2006.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
CPC
G06F 11/34
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
11Error detection; Error correction; Monitoring
30Monitoring
34Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; ; Recording or statistical evaluation of user activity, e.g. usability assessment
G06F 11/3457
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
11Error detection; Error correction; Monitoring
30Monitoring
34Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; ; Recording or statistical evaluation of user activity, e.g. usability assessment
3457Performance evaluation by simulation
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
G06T 2207/20084
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
2207Indexing scheme for image analysis or image enhancement
20Special algorithmic details
20084Artificial neural networks [ANN]
G06T 7/00
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
G06T 7/0004
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
0002Inspection of images, e.g. flaw detection
0004Industrial image inspection
Applicants
  • 에스케이 주식회사 SK HOLDINGS CO., LTD. [KR]/[KR]
Inventors
  • 이병민 LEE, Byung Min
  • 김영희 KIM, Young Hee
  • 김종문 KIM, Jong Moon
  • 천기쁨 CHUN, Ki Peum
Agents
  • 양성환 YANG, Seong Hwan
Priority Data
10-2019-000099804.01.2019KR
Publication Language Korean (KO)
Filing Language Korean (KO)
Designated States
Title
(EN) EXPLAINABLE ARTIFICIAL INTELLIGENCE MODELING AND SIMULATION SYSTEM AND METHOD
(FR) SYSTÈME ET PROCÉDÉ DE MODÉLISATION ET DE SIMULATION EXPLICABLES D'INTELLIGENCE ARTIFICIELLE
(KO) 설명 가능한 인공지능 모델링 및 시뮬레이션 시스템 및 방법
Abstract
(EN)
Provided is a system and method that generates an artificial intelligence workflow model in which image conversion, measurement, and image searches can be performed, by selecting/combining algorithms suitable for a workflow having been created/modified according to a subject area such as a display manufacturing process, and that conducts a simulation of the generated model. An explainable artificial intelligence modelling and simulation method according to an embodiment of the present invention comprises the steps of: designing an artificial intelligence workflow model once algorithms suitable for a workflow having been created/modified according to a subject area are selected from algorithms stored in advance; and conducting, when input information is input, a simulation of the artificial intelligence workflow model for the input information. As described above, according to embodiments of the present invention, it is possible to visualize the connections between algorithms on a workflow basis and automatically assess the performance thereof through an integrated simulation, and it is also possible to perform performance comparison by simultaneously conducting simulations of multiple artificial intelligence algorithms on a workflow basis. In addition, an improvable part for the performance of the algorithms can be suggested by enabling the causes/basis of classification results to be described with images and data using a standard model.
(FR)
L'invention concerne un système et un procédé qui génèrent un modèle de flux de travail d'intelligence artificielle dans lequel une conversion d'images, une mesure et des recherches d'images peuvent être effectuées, par sélection/combinaison d'algorithmes convenant à un flux de travail qu'on a créé/modifié selon un domaine tel qu'un processus de fabrication d'affichages, et qui effectuent une simulation du modèle généré. Un procédé de modélisation et de simulation explicables d'intelligence artificielle selon un mode de réalisation de la présente invention comprend les étapes consistant : à concevoir un modèle de flux de travail d'intelligence artificielle une fois que des algorithmes convenant à un flux de travail qu'on a créés/modifiés selon un certain domaine sont sélectionnés parmi des algorithmes stockés à l'avance ; et à réaliser, lorsque des informations d'entrée sont entrées, une simulation du modèle de flux de travail d'intelligence artificielle pour les informations d'entrée. Comme décrit ci-dessus, selon des modes de réalisation de la présente invention, il est possible de visualiser les connexions entre des algorithmes sur une base de flux de travail et d'évaluer automatiquement ses performances par l'intermédiaire d'une simulation intégrée, et il est également possible d'effectuer une comparaison de performances par réalisation simultanée de simulations de multiples algorithmes d'intelligence artificielle sur une base de flux de travail. De plus, une partie d'amélioration concernant les performances des algorithmes peut être suggérée par autorisation de la description des causes/bases de résultats de classification par des images et par des données, selon un modèle standard.
(KO)
디스플레이의 제조공정 등의 주제 영역에 맞게 작성/편집된 워크플로우에 적합한 알고리즘을 선택/조립하여 이미지 변환, 계측 및 이미지 검색이 가능한 인공지능 워크플로우 모델을 생성하고, 시뮬레이션을 수행하는 시스템 및 방법이 제공된다. 본 발명의 실시예에 따른 설명 가능한 인공지능 모델링 및 시뮬레이션 방법은 기저장된 알고리즘 중 주제 영역에 맞게 작성/편집된 워크플로우에 적합한 알고리즘이 선택되면, 인공지능 워크플로우 모델을 설계하는 단계; 및 입력정보가 입력되면, 입력정보를 대상으로 인공지능 워크플로우 모델에 대한 시뮬레이션을 수행하는 단계;를 포함한다. 이에 의해, 이상 설명한 바와 같이, 본 발명의 실시예들에 따르면, 워크플로우 기반으로 알고리즘의 연결을 가시화하고 통합된 시뮬레이션으로 자동 성능 검증할 수 있으며, 워크플로우 기반의 복수의 인공지능 알고리즘을 동시에 시뮬레이션하여 상호 성능 비교를 할 수 있다. 또한, 분류 분석의 결과가 이미지와 표준화된 형상 데이터로 원인/근거를 설명할 수 있게 하여 알고리즘의 성능 개선방향을 제시할 수 있다.
Also published as
Latest bibliographic data on file with the International Bureau