WCSE

LT-AES: Automatic Academic Paper Evaluation Model

Abstract— With the rapid development of science technology and the enhancement of innovation awareness, more and more papers have been submitted to various conferences and journals. When reviewing a paper, viewers of these conferences or journals ensure that their view of the literature is fast and effective and the results of the review are fair and reasonable. Many scholars have proposed many scoring methods to solve the above problems, but these methods only pay attention to the local features of the literature and do not consider the relationship between different modules in the literature. To this end, this paper proposes an LT-AES model to achieve an automatic evaluation of the paper, and obtain the score value of the academic paper. It was found by the experiment, the LT-AES method proposed in this paper has much higher accuracy than other models, and increasing the accuracy rate of more than 3.6%

Index Terms—Deep Learning, Automatic Scoring, Bi-LSTM, Self-Attention

Liang Dong, Ankang Chu, Yuan Meng
Future Network Innovation Center, Beijing University of Technology, CHINA
Information Department, Beijing University of Technology, CHINA

[Download]

Cite:Liang Dong, Ankang Chu, Yuan Meng" LT-AES: Automatic Academic Paper Evaluation Model " Proceedings of 2020 the 10th International Workshop on Computer Science and Engineering (WCSE 2020), pp. 341-346, Shanghai, China, 19-21 June, 2020.