Method and device for training text auditing model

The invention provides a method and device for training a text auditing model, and relates to the field of artificial intelligence, in particular to the field of natural language processing. According to the specific implementation scheme, a domain language model, a student model, unlabeled data and...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: CHEN YONGFENG, WANG ZANBO, HUANG SHUO, CAO YUHUI
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention provides a method and device for training a text auditing model, and relates to the field of artificial intelligence, in particular to the field of natural language processing. According to the specific implementation scheme, a domain language model, a student model, unlabeled data and labeled data are obtained, and the labeled data comprise text information and auditing labels; taking the text information and the auditing label in the annotation data as input and expected output respectively, and performing fine tuning training on the domain language model to obtain a teacher model; inputting the unlabeled data into the teacher model, and outputting a pseudo-audit label to obtain pseudo-labeled data; and training the student model through the pseudo-annotation data to obtain a text auditing model. According to the embodiment, training can be carried out on small-scale manual annotation data and large-scale non-annotation data, and the text auditing model with a good effect and a high speed is o