基于LDM技术的具身语义词汇优化平台研究

LDM-based Vocabulary Learning System Optimized by Embodied Semantic Network

本研究利用语言分布式模型(linguistic distributional modeling,LDM)技术建立了具身语义词汇数据库。本文首先介绍了具身语义关联技术的运用现状,聚焦词汇的具身语义行为指标研究的发展动态,然后针对语言分布模型的定义、逻辑和数据分析特点,提出语言信息分布知识的建模方法,最后针对中英双语的语料库研究了中英平行功能的构建(主要包括同义词选择、文本分类、文本预测和平行翻译四大功能)。本研究基于LDM构建的具身语义优化词汇学习系统,能够为未来人工智能时代语言文本信息加工的人脑和电脑关联计算,提供巨大的理论和应用前景。

Based on the linguistic distributional modeling(LDM),an embodied semantic word-base system has been established in this study.This paper first introduces the application of technology associated with embodied semantics and focuses on the development of index researches on embodied semantical behavior.Then,the modeling method of knowledge of language distributional information is proposed in terms of definition,logic and LDM data-analytical approach.Finally,Chinese-English parallel functioning network(including synonym selection,text classification,text prediction and parallel translation)is constructed based on Chinese-English bilingual corpus.LDM-based vocabulary learning system optimized by embodied semantic network is of great significance for theories and practices of future brain-like intelligence in the context of language learning and brain informatics.