基于UNILM框架的对抗训练新闻标题生成方法

2021 International Conference on Computer Engineering and Application (ICCEA) Pub Date : 2021-06-01 DOI:10.1109/ICCEA53728.2021.00014

Yue Heng

{"title":"基于UNILM框架的对抗训练新闻标题生成方法","authors":"Yue Heng","doi":"10.1109/ICCEA53728.2021.00014","DOIUrl":null,"url":null,"abstract":"Text generation is now a very mature task. Many methods have been applied to text generation and achieved good results. This paper uses the pre-trained model Unilm with adversarial training to generate news headlines from the Thucnews dataset. We fine tune the methods and parameters in the model and produced some results for reference or comparison","PeriodicalId":325790,"journal":{"name":"2021 International Conference on Computer Engineering and Application (ICCEA)","volume":"219 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A News Title Generation Method Based on UNILM Framework via Adversarial Training\",\"authors\":\"Yue Heng\",\"doi\":\"10.1109/ICCEA53728.2021.00014\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Text generation is now a very mature task. Many methods have been applied to text generation and achieved good results. This paper uses the pre-trained model Unilm with adversarial training to generate news headlines from the Thucnews dataset. We fine tune the methods and parameters in the model and produced some results for reference or comparison\",\"PeriodicalId\":325790,\"journal\":{\"name\":\"2021 International Conference on Computer Engineering and Application (ICCEA)\",\"volume\":\"219 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-06-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 International Conference on Computer Engineering and Application (ICCEA)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICCEA53728.2021.00014\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 International Conference on Computer Engineering and Application (ICCEA)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICCEA53728.2021.00014","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 0

摘要

文本生成现在是一个非常成熟的任务。许多方法已经应用于文本生成，并取得了良好的效果。本文使用带有对抗训练的预训练模型Unilm从Thucnews数据集生成新闻标题。我们对模型中的方法和参数进行了微调，得出了一些可供参考或比较的结果

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

A News Title Generation Method Based on UNILM Framework via Adversarial Training

Text generation is now a very mature task. Many methods have been applied to text generation and achieved good results. This paper uses the pre-trained model Unilm with adversarial training to generate news headlines from the Thucnews dataset. We fine tune the methods and parameters in the model and produced some results for reference or comparison

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

2021 International Conference on Computer Engineering and Application (ICCEA)

自引率

0.00%

发文量

期刊最新文献

Few-shot Image Classification based on LMRNet Design and Test on Acoustic Device for Actively Measuring Underwater Short Distance with High-Precision KVM PT Based Coverage Feedback Fuzzing for Network Key Devices Acoustic impedance inversion base on dual learning Numerical simulation of aerodynamic force and moored state in airship transport process