METHODS AND APPARATUS FOR EXTRACTING DATA IN DEEP NEURAL NETWORKS

Disclosed are a method and device for extracting data in a deep learning model. According to one embodiment of the present invention, in a deep learning model, the method for extracting data comprises the steps of: receiving an input query; determining a first decision boundary set, which is a subse...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: JEONG HAE DONG, CHOI JAE SIK, JEON GI YOUNG
Format: Patent
Sprache:eng ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Disclosed are a method and device for extracting data in a deep learning model. According to one embodiment of the present invention, in a deep learning model, the method for extracting data comprises the steps of: receiving an input query; determining a first decision boundary set, which is a subset of a decision boundary set corresponding to a target layer of the deep learning model; extracting a decision region containing the input query based on the first decision boundary set; and extracting data included in the decision region. 딥러닝 모델에 있어서, 데이터를 추출하는 방법 및 장치가 개시된다. 일 실시예에 따른 딥러닝 모델에 있어서, 데이터 추출 방법은 입력 쿼리(input query)를 수신하는 단계, 딥러닝 모델의 타겟 레이어(target layer)에 대응하는 결정 경계(decision boundary) 집합의 부분 집합인 제1 결정 경계 집합을 결정하는 단계, 제1 결정 경계 집합에 기초하여, 입력 쿼리를 포함하는 결정 영역(decision region)을 추출하는 단계 및 결정 영역에 포함된 데이터를 추출하는 단계를 포함한다.