短文本表示建模及应用

　　短文本表示建模，通常是指将短文本转化成机器可以诠释的形式，旨在帮助机器“理解”短文本的含义。本书详细介绍了短文本表示建模研究体系中具有代表性的短文本概念化表示建模研究分支和短文本向量化表示建模研究分支的相关研究方法，既涵盖了大量经典算法，又特别引入了近年来在该领域研究中涌现出的新方法、新思路，力求兼顾内容的基础性和前沿性。同时，本书融入了作者多年来从事以概念化和向量化为核心的短文本表示建模方法与理论研究的经验和成果，并以短文本检索这一典型应用问题为例，详细介绍了如何把短文本概念化表示建模方法和短文本向量化表示建模方法以及先进的设计思想融入具体应用问题的求解。本书可供计算机、信息处理、自动化、系统工程、应用数学等专业的教师以及相关领域的研究人员和技术开发人员参考。

　　王亚珅，博士，高级工程师，2012年毕业于北京理工大学计算机学院获学士学位，2018年毕业于北京理工大学计算机学院获博士学位，目前任社会安全风险感知与防控大数据应用国家工程实验室知识智能室主任，研究方向包括自然语言处理、知识工程、社交网络分析等。获2018年中国博士后科学基金会第64批面上资助等，主持中国电科集团新一代人工智能专项行动计划项目“基于大数据智能的立体化社会治安防控”等。获2018年人工智能学会优秀博士学位论文奖等。任中国人工智能学会青年工作委员会成员、会员，《无人系统技术》期刊青年编委。近五年，以作者身份发表TKDE、TKDD、ACL、WWW等会议/期刊论文20余篇，以完成人身份受理发明专利20余项。

第1 章 绪论···················································································· 1
1.1 研究背景及意义 ······································································· 1
1.2 基本定义及问题描述 ································································· 2
1.3 研究问题图解 ·········································································· 6
1.4 本书内容组织结构 ···································································· 7
第2 章 理论与技术基础 ····································································· 9
2.1 分布假说 ················································································ 9
2.2 向量空间模型 ········································································ 10
2.3 词频 − 逆文档频率 ·································································· 10
2.4 链接分析 ··············································································· 11
2.5 马尔可夫随机场 ····································································· 15
2.6 参数分布估计 ········································································ 17
2.7 词语向量化 ··········································································· 20
2.8 语言模型 ·············································································· 24
2.9 数据平滑算法 ········································································ 26
2.10 模型求解算法 ······································································ 28
2.11 向量语义相似度计算 ······························································ 32
2.12 查询扩展 ············································································· 34
第3 章 面向短文本表示建模的知识库资源 ··········································· 37
3.1 引言 ···················································································· 37
3.2 百科类知识库资源 ·································································· 37
3.3 词汇语义知识库资源 ······························································· 41
3.4 知识库资源对比分析 ······························································· 46
第4 章 显式语义建模 ······································································ 48
4.1 引言 ···················································································· 48
4.2 显式语义分析模型 ·································································· 48
4.3 概念化模型 ··········································································· 49
4.4 显式语义建模总结分析 ···························································· 51
第5 章 半显式语义建模 ··································································· 52
5.1 引言 ···················································································· 52
5.2 概率化潜在语义分析模型 ························································· 52
5.3 潜在狄利克雷分布模型 ···························································· 53
5.4 层次化狄利克雷过程模型 ························································· 54
5.5 半显式语义建模总结分析 ························································· 58
第6 章 隐式语义建模 ······································································ 59
6.1 引言 ···················································································· 59
6.2 潜在语义分析模型 ·································································· 59
6.3 神经网络语言模型 ·································································· 61
6.4 CBOW 模型和Skip-Gram 模型 ··················································· 65
6.5 隐式语义建模总结分析 ···························································· 67
第7 章 短文本概念化表示建模 ·························································· 68
7.1 引言 ···················································································· 68
7.2 问题描述 ·············································································· 68
7.3 短文本概念化方法 ·································································· 69
7.4 短文本概念化方法总结分析 ······················································ 95
7.5 本章小结 ············································································· 105
第8 章 短文本向量化表示建模 ························································· 107
第9 章 概念化和向量化在短文本检索问题中的应用 ······························ 149
第10 章 总结与展望 ······································································ 200
参考文献 ······················································································· 204