Research and Implementation of Network Comment Automatic Generation Algorithm Based on Pre-trained Language Model

XJTU 2024 CSUC Submission17 Authors

02 Apr 2024 (modified: 03 Apr 2024)XJTU 2024 CSUC SubmissionEveryoneRevisionsBibTeXCC BY 4.0
Keywords: Automatic news commentary, deep learning, pre trained language model, BART
Abstract: With the explosive growth of information, how to quickly generate high-quality comments on various news content has become a hot research topic. Unlike traditional machine learning models with automatic comment generation algorithms, emerging deep learning models with automatic comment generation algorithms have their own unique advantages. Starting from the pre trained language model, we trained a large number of corpora and obtained a model that performs well in the task of automatic news comment generation. This proves the feasibility and unique advantages of the pre trained language model based automatic comment generation algorithm, and provides a foundation for further exploration of more efficient pre trained language model based automatic comment generation algorithms in the future.
Submission Number: 17
Loading