一种基于CNNFeed网络模型的文本分类方法
By combining CNN unit modules and Transformer feedforward neural networks into a single layer, a scalable layered structure is constructed, which solves the problems of low computational efficiency of Transformer and decreased fitting ability of TextCNN, and achieves efficient text classification capabilities.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GANSU WANWEI INFORMATION TECH CO LTD
- Filing Date
- 2022-12-15
- Publication Date
- 2026-07-17
AI Technical Summary
Existing Transformer models are computationally inefficient and difficult to scale in text classification tasks, while TextCNN models have reduced ability to fit data when the task is complex and cannot efficiently handle large-scale data.
Design a CNNFeed network model that combines CNN unit modules and Transformer feedforward neural networks into a single layer. Employ same convolution and layer normalization operations to construct a scalable layered structure that can adapt to complex tasks by increasing the number of layers.
It improves the computational efficiency and data fitting ability of the model, enabling it to efficiently handle complex text classification tasks and adapt to text classification needs with different amounts and complexities of data.
Smart Images
Figure CN115905539B_ABST