摘要
语言学理论对自然语言处理的影响不仅可以从基于规则的理性主义方法看到,也可以从基于机器学习的经验主义方法中感到。由于缺乏对社会语言学理论和成果的了解,目前的自然语言处理仍然局限于语言的符号学理论的框架,没有从社会结构和社会互动的角度来理解语言和言语交际,因此难以突破语义处理的困境。突破困境的出路在于语境语义模型的建立,根据社区语境分类的大数据语料将有助于深度学习,发现有用的语用参数并构建语境语义处理的模型。
The influences of linguistic theories on natural language processing (NLP) find expression in the empiricist method of machine learning as well as in the rule-based rationalist method. Due to the lack of understanding of sociolinguistie theories, the NLP today is still limited by the theory of semiotics. The failure of understanding language from the social structural and interactional perspectives makes a barrier for NLP in processing messages in language use. For the approach of deep learning with big data to solve the problem, it is suggested that the training corpus is classified by community-defined situations of speech interaction.
出处
《云南师范大学学报(哲学社会科学版)》
CSSCI
北大核心
2017年第3期1-9,共9页
Journal of Yunnan Normal University:Humanities and Social Sciences Edition
关键词
语言学理论
自然语言处理
社区语境
大数据
语境语义
linguistic theory
natural language processing
community-defined situation
big data
situational meaning