A deep learning model learning method and a computing device for performing the same are disclosed. A deep learning model learning method according to an embodiment disclosed herein is a method performed in a computing device having one or more processors, and a memory for storing one or more programs executed by one or more processors, the first training data obtaining and extracting a learning feature vector therefrom, dividing the first learning data into a plurality of groups based on the extracted learning feature vector, and a correct answer value feature vector from the correct values labeled in the first learning data extracting , dividing the correct answer value feature vectors into a plurality of groups so as to correspond to the group of learning feature vectors. Calculating group reference information for each group of correct value feature vectors, and setting quality weights for second learning data using group reference information for each group.
展开▼
