作　　者： (）;

机构地区： 中国南方电网有限责任公司

出　　处： 《自动化与仪器仪表》 2020年第6期95-97,102,共4页

摘　　要： 针对传统电网NLP中序列到序列深度学习模型学习误差大这一问题,基于运管智能助理NLP设计了一种新的深度学习模型。研究了NLP中输入序列到输出序列注意力机制,分析了NLP中的神经系统,语言以及程序,建立NLP中输入序列到输出模型,并针对程序细节进行设计,通过得到的注意力机制构建NLP中序列到序列深度学习模型,通过输入层、隐藏层、输出层实现针对序列到序列的深度学习。设定对比实验,结果表明,给出的深度学习模型具有很好的学习性能,学习结果误差较小。 Aiming at the problem of large learning error in sequence-to-sequence deep learning model in traditional NLP,a new deep learning model is designed based on the NLP of operation management assistant.The attention mechanism of input sequence to output sequence in NLP is studied.The nervous system,language and program in NLP are analyzed.The input sequence to output model in NLP is established,and the program details are designed.The attention mechanism is used to construct NLP.The sequence-to-sequence deep learning model implements deep learning for sequence-to-sequence through input layer,hidden layer,and output layer.The contrast experiment is set up.The results show that the given deep learning model has a good learning performance and the learning result error is small.

关 键 词： 序列到序列 深度学习 学习模型 电网领域

领　　域： []