A feature learning method fusing genetic algorithm and meta-learning

By integrating genetic algorithms and meta-learning feature learning methods, this approach addresses the issues of overfitting and slow training speed in small-sample text classification, thereby improving both the accuracy and efficiency of text classification. It is applicable to multi-objective optimization and natural language processing of Chinese and English texts.

CN115841116BActive Publication Date: 2026-06-09GUANGDONG UNIV OF TECH +1
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGDONG UNIV OF TECH
Filing Date
2022-07-15
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing deep learning models suffer from overfitting, slow training speed, and inefficient feature representation in small sample text classification, especially in domains with limited data, where existing methods have failed to effectively improve computational efficiency and classification accuracy.

Method used

A feature learning method that integrates genetic algorithms and meta-learning is adopted, including word segmentation, pre-trained word vector encoding, genetic algorithm feature selection, sentence-level sequence encoding, and ridge regression classifier. The genetic algorithm reduces text dimensionality and improves feature selection efficiency, while the meta-learning framework enhances few-shot learning capabilities.

Benefits of technology

It effectively reduces the data dimensionality of the text classification model, improves the accuracy and training speed of small sample text classification, and solves the problems of overfitting and long training time of the model under small sample data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115841116B_ABST
    Figure CN115841116B_ABST
Patent Text Reader

Abstract

The application provides a feature learning method integrating genetic algorithm and meta learning, comprising the following steps: (1) adopting a pre-training word segmentation model to perform word segmentation on input text; (2) using a pre-training word vector to encode the word segmentation result; (3) then adopting a genetic algorithm to perform feature screening; (4) adopting an Att-Bilstm model for sentence-level sequence encoding; (5) next, adopting an optimal feature combination as the input of a meta learning model; (6) adopting a ridge regression classifier. The application provides a method integrating genetic algorithm feature extraction, a small sample learning framework and a ridge regression classifier; the genetic algorithm is adopted to reduce the dimension of input text, the small sample learning framework improves the rapid learning ability between multiple tasks, and the ridge regression classifier avoids the disadvantage of insufficient generalization ability.
Need to check novelty before this filing date? Find Prior Art