METHOD AND SYSTEM FOR PROVIDING OF NAIVE SEMI-SUPERVISED DEEP LEARNING USING UNLABELED DATA

摘要

The present invention relates to a method for providing naive semi-supervised deep learning for improving the performance of a deep learning model by using unlabeled data and a system therefor. According to the present invention, semi-supervised deep learning with improved performance can be provided by using similar labels even in case of the lack of training data. The method for providing naive semi-supervised deep learning comprises: training a classifier; predicting pseudo-labels; pre-training a deep learning model; and fine-tuning the deep learning model.