A Novel Ehanced Move Recognition Algorithm Based on Pre-trained Models with Positional Embeddings

TL;DR

This paper introduces an enhanced move recognition algorithm for Chinese scientific abstracts that leverages pre-trained models with positional embeddings, significantly improving accuracy by incorporating word position information.

Contribution

It proposes a novel EP-ERNIE_AT-GRU framework that integrates positional embeddings into pre-trained models for better semantic understanding of abstracts.

Findings

13.37% higher accuracy on split dataset

7.55% improvement over basic model

Effective learning of contextual semantics with positional info

Abstract

The recognition of abstracts is crucial for effectively locating the content and clarifying the article. Existing move recognition algorithms lack the ability to learn word position information to obtain contextual semantics. This paper proposes a novel enhanced move recognition algorithm with an improved pre-trained model and a gated network with attention mechanism for unstructured abstracts of Chinese scientific and technological papers. The proposed algorithm first performs summary data segmentation and vocabulary training. The EP-ERNIE$\_$AT-GRU framework is leveraged to incorporate word positional information, facilitating deep semantic learning and targeted feature extraction. Experimental results demonstrate that the proposed algorithm achieves 13.37$\%$ higher accuracy on the split dataset than on the original dataset and a 7.55$\%$ improvement in accuracy over the basic comparison model.