Method, electronic device and readable storage medium for creating a label marking model

An electronic device, comprising:

• at least one processor;
• and
• a storage communicatively connected with the at least one processor;
• wherein,
• the storage stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform a method for creating a label marking model, wherein the method comprises:
• obtaining text data and determining a word or phrase to be marked in the text data;
• according to the word or phrase to be marked, constructing first training samples of the text data corresponding to a word or phrase replacing task and a second training sample corresponding to a label marking task;
• wherein the first training samples comprise the text data and a replacement text with the word or phrase replaced with a substitute word or phrase, and the substitute word or phrase comprises at least one type of:
• a label word or phrase associated with the word or phrase to be marked, an appositive word or phrase associated with the word or phrase to be marked, and a label word or phrase not associated with the word or phrase to be marked;
• and
• training a neural network model with a plurality of the first training samples and a plurality of the second training samples, respectively, until a loss function of the word or phrase replacing task and a loss function of the label marking task satisfy a preset condition, to obtain the label marking model,
• wherein the word or phrase replacing task is a task for predicting a replacement class of the replacement text corresponding to the type of the substitute word or phrase,
• wherein constructing the second training sample of the text data corresponding to the label marking task according to the word or phrase to be marked comprises:
• obtaining a label word or phrase associated with the word or phrase to be marked, and taking the label word or phrase as a label marking result of the word or phrase to be marked;
• and
• taking the text data, the word or phrase to be marked and the label marking result corresponding to the word or phrase to be marked as the second training sample.

A non-transitory computer-readable storage medium storing computer instructions therein, wherein the computer instructions are used to cause the computer to perform a method for creating a label marking model, wherein the method comprises:

• obtaining text data and determining a word or phrase to be marked in the text data;
• according to the word or phrase to be marked, constructing first training samples of the text data corresponding to a word or phrase replacing task and a second training sample corresponding to a label marking task;
• wherein the first training samples comprise the text data and a replacement text with the word or phrase replaced with a substitute word or phrase, and the substitute word or phrase comprises at least one type of:
• a label word or phrase associated with the word or phrase to be marked, an appositive word or phrase associated with the word or phrase to be marked, and a label word or phrase not associated with the word or phrase to be marked;
• and
• training a neural network model with a plurality of the first training samples and a plurality of the second training samples, respectively, until a loss function of the word or phrase replacing task and a loss function of the label marking task satisfy a preset condition, to obtain the label marking model,
• wherein the word or phrase replacing task is a task for predicting a replacement class of the replacement text corresponding to the type of the substitute word or phrase,
• wherein constructing the second training sample of the text data corresponding to the label marking task according to the word or phrase to be marked comprises:
• obtaining a label word or phrase associated with the word or phrase to be marked, and taking the label word or phrase as a label marking result of the word or phrase to be marked;
• and
• taking the text data, the word or phrase to be marked and the label marking result corresponding to the word or phrase to be marked as the second training sample.