摘要
自动化的文本生成是自然语言处理中一项极具挑战性的课题,在机器翻译和计算机视觉等人工智能研究领域中都具有重要的应用。近年来,随着对深度学习研究的不断推进,基于深度学习模型的文本生成方法受到人们广泛的关注。对深度学习中的长短时记忆(LSTM)网络模型进行研究,并应用到古诗词生成的场景中。实验结果表明,长短时记忆网络具有良好的文本生成效果。
Automatic text generation is a challenging problem in natural language programming,which plays an important role in the artificial intelli?gence fields such as machine translation and computer vision.With great efforts have been devoted to deep learning for years,the text gen?eration approaches based on deep models have attracted wide attention.Carries out a research on the Long Short Term Memory(LSTM)Net?work,and applies it into the generation of Chinese poetry.Experimental result shows that favorable performance can be achieved.
作者
董虎胜
DONG Hu-sheng(Suzhou Institute of Trade and Commerce,Suzhou 215009)
基金
苏州经贸职业学院校级科研项目(No.KY-ZR1705)
