Research on generating text summaries based on improved transformer

Weijun Gao; Wenjing Mao; Jian Liu; Kai Wang

doi:10.1117/12.2685548

15 August 2023 Research on generating text summaries based on improved transformer

Weijun Gao, Wenjing Mao, Jian Liu, Kai Wang

Proceedings Volume 12719, Second International Conference on Electronic Information Technology (EIT 2023); 127192G (2023) https://doi.org/10.1117/12.2685548
Event: Second International Conference on Electronic Information Technology (EIT 2023), 2023, Wuhan, China

Abstract

With the advent of the information age, a large number of information has emerged, and people need to deal with more and more information. In this case, text summarization technology came into being. As a Text summarization model, Transformer has the problem of obtaining key text information and out of vocabulary, which is difficult to effectively solve and resulted in the accuracy of the generated summary not achieving the desired effect, an improved transformer model was proposed, The model constructs an Aggregation module between the encoder and decoder to make the generated abstract more appropriate to the original content ,and the pointer network is introduced at the decoding end of the model to complete the decoding of the text summary, The function of pointer network that allows the unlisted words in the vocabulary to be copied directly from the source text is used to generate a text summary, which improves the flexibility of summary generation, Using the News2016zh dataset to test the performance of the model. In the experimental results, the improved Transformer model performs better than the benchmark model on ROUGE evaluation indicators.

Citation Download Citation

Weijun Gao, Wenjing Mao, Jian Liu, and Kai Wang "Research on generating text summaries based on improved transformer", Proc. SPIE 12719, Second International Conference on Electronic Information Technology (EIT 2023), 127192G (15 August 2023); https://doi.org/10.1117/12.2685548

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available