其他分享
首页 > 其他分享> > 论文阅读:Factorized Recurrent Neural Architectures for Longer Range Dependence

论文阅读:Factorized Recurrent Neural Architectures for Longer Range Dependence

作者:互联网

该论文主要针对LSTM并无法很好处理Long Range Dependence, 提出分块处理技术,能够保证不影响运算速度的情况下,提供对长范围依赖特性的建模。

 

Belletti, Francois, Alex Beutel, Sagar Jain, and Ed Chi. "Factorized recurrent neural architectures for longer range dependence." In International Conference on Artificial Intelligence and Statistics, pp. 1522-1530. 2018.

标签:Conference,Recurrent,Factorized,Neural,论文,Dependence,Range,longer
来源: https://www.cnblogs.com/hugh2006/p/11556067.html