WEIGHT SETTING APPARATUS FOR NEURAL NETWORK LEARNING

The present invention relates to a weight setting device for neural network learning, and more specifically, to a weight setting device for neural network learning, which performs stabilized neural network learning and processes low-frequency data so that the low-frequency data does not affect the w...

Full description

Saved in:
Bibliographic Details
Main Authors LEE JUSANG, SHIN JOONCHOUL, OCK CHEOLYOUNG, KIM WANSU
Format Patent
LanguageEnglish
Korean
Published 31.05.2023
Subjects
CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The present invention relates to a weight setting device for neural network learning, and more specifically, to a weight setting device for neural network learning, which performs stabilized neural network learning and processes low-frequency data so that the low-frequency data does not affect the weight settings of neural network learning. The weight setting device for neural network learning is supported by an operating system and artificial intelligence open platform which operates hardware resources of a processor, a memory, and input/output devices. AI model is operated on the artificial intelligence open platform. A learning unit trains the AI model to perform basic probability testing, frequency limit testing, and dropout processing. An application program, which is an application executed by the artificial neural network of the AI model, is executed. 본 발명은 신경망 학습의 가중치 설정 장치에 관한 것으로, 보다 상세하게는 안정화된 신경망 학습을 수행하고, 빈도가 낮은 데이터가 신경망 학습의 가중치 설정에 영향을 미치지 않도록 처리하는 신경망 학습의 가중치 설정 장치에 관한 것이다.
Bibliography:Application Number: KR20210161223