Training sample generation method and device, electronic equipment and storage medium

Abstract

The invention provides a training sample generation method. The training sample generation method comprises the steps of obtaining to-be-processed initial text data; performing word segmentation processing on the initial text data to form keywords matched with the initial text data; screening the initial text according to a keyword matched with the initial text data to form a target text for a specified service; training a corresponding text processing model through the target text; and performing domain data augmentation processing on the target text according to a training result of the textprocessing model to form a training sample for a specified service. The invention further provides a training sample generation device, electronic equipment and a storage medium. According to the method, the pertinence of the training sample can be improved, so that the training sample is more suitable for a machine reading understanding task. Meanwhile, the recognition accuracy of the neural network model in the specific service field is improved, and the robustness of the neural network model is enhanced.

Description

technical field [0001] The present invention relates to machine learning technology, in particular to a training sample generation method, device, electronic equipment and storage medium. Background technique [0002] In the existing technology, with the development of machine learning technology, machine reading can be realized through the BERT (BidirectionalEncoder Representations from Transformers, bidirectional encoder derived from Transformers) mechanism. In the process of processing text data using the model corresponding to BERT, The text data can be split into characters one by one, and then each character is input into the corresponding model of BERT in turn to obtain the corresponding output result. However, in this process, there are defects that the model trained for open domains does not perform well in specific domains (low accuracy, poor robustness), and the lack of training data resources for machine reading comprehension tasks in specific domains. question....

AI Technical Summary

Problems solved by technology

However, in this process, there are defects that the model trained for open domains does not perform well in specific domains (low accuracy, poor robustness), and the lack of training data resources for machine reading comprehension tasks in specific domains. question

Images