Cargando…
News headline generation based on improved decoder from transformer
Most of the news headline generation models that use the sequence-to-sequence model or recurrent network have two shortcomings: the lack of parallel ability of the model and easily repeated generation of words. It is difficult to select the important words in news and reproduce these expressions, re...
Autores principales: | Li, Zhengpeng, Wu, Jiansheng, Miao, Jiawei, Yu, Xinmiao |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Nature Publishing Group UK
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9270459/ https://www.ncbi.nlm.nih.gov/pubmed/35804183 http://dx.doi.org/10.1038/s41598-022-15817-z |
Ejemplares similares
-
New results make headline news
por: Riesselmann, K
Publicado: (2001) -
Profiling the barriers to the spreading of news using news headlines
por: Sittar, Abdul, et al.
Publicado: (2023) -
Fraudulent News Headline Detection with Attention Mechanism
por: Liu, Hankun, et al.
Publicado: (2021) -
Longitudinal analysis of sentiment and emotion in news media headlines using automated labelling with Transformer language models
por: Rozado, David, et al.
Publicado: (2022) -
Kurdish News Dataset Headlines (KNDH) through multiclass classification
por: Badawi, Soran, et al.
Publicado: (2023)