METHOD AND APPARATUS FOR GENERATING QUESTION AND ANSWER DATASET BASED ON INPUT PARAGRAPH

A method for generating a question-answering dataset based on input of a paragraph including a plurality of sentences according to an embodiment of the present invention may include the steps of: generating a machine-readable data representation from an input paragraph through a neural network and s...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	SOHN AH RIM, DO GANG HO, KIM DONG HWAN, JEONG WOO TAE
Format:	Patent
Sprache:	eng ; kor
Schlagworte:	CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING ELECTRIC DIGITAL DATA PROCESSING PHYSICS
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	A method for generating a question-answering dataset based on input of a paragraph including a plurality of sentences according to an embodiment of the present invention may include the steps of: generating a machine-readable data representation from an input paragraph through a neural network and storing the data representation in memory; performing sampling on the data representation; calculating a similarity of the data representation associated with a query in the data representation of the memory by using the sampled data as the query; generating a response based on the similarity; and generating a query based on the response and the similarity. 본 발명의 실시예를 따르는 복수의 문장이 포함된 문단의 입력에 기초한 질의응답 데이터셋 생성방법은, 상기 입력된 문단으로부터 신경망 네트워크를 통해 기계가 판독가능한 데이터 표현을 생성하여 메모리에 저장하는 단계; 상기 데이터 표현에서 샘플링을 수행하는 단계; 상기 샘플링된 데이터를 쿼리로 하여, 상기 메모리의 상기 데이터 표현에서 상기 쿼리와 연관이 있는 데이터 표현의 유사도를 계산하는 단계; 상기 유사도를 기초로 응답을 생성하는 단계; 및 상기 응답과 상기 유사도를 기초로 질의를 생성하는 단계를 포함하는 것을 특징으로 한다.