新書推薦：
《
现代欧洲史：1500—1815
》
售價：HK$
105.6
《
高颜值创意饮品：咖啡 茶饮 鸡尾酒 气泡水
》
售價：HK$
71.8
《
慢慢来，好戏都在烟火里
》
售價：HK$
59.8
《
一间自己的房间
》
售價：HK$
47.8
《
波段交易的高级技术：股票和期权交易者的资金管理、规则、策略和程序指南
》
售價：HK$
94.8
《
人，为什么需要存在感：罗洛·梅谈死亡焦虑
》
售價：HK$
81.6
《
锁国：日本的悲剧
》
售價：HK$
93.6
《
AI智能写作： 巧用AI大模型 让新媒体变现插上翅膀
》
售價：HK$
70.8
|
| 編輯推薦： |
|
本书以新方法、新思路为导线，介绍短文本表示建模方法与理论研究成果，兼顾内容的基础性和前沿性。
|
| 內容簡介： |
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。
|
| 關於作者： |
|
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。
|
| 目錄：
|
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
|
| 內容試閱：
|
人类的语言，是人类独有的进化千万年后形成的信息表达方式。相较于具有
原始信号输入的图像（像素）和语音（声谱），符号化的自然语言属于更高层次的
抽象实体。因此，众多自然语言处理应用的步（也是至关重要的一步）就是
将符号化的自然语言表示成计算机能理解的形式，即文本表示建模。在当前网络
空间短文本信息超载的形势下，“短文本表示建模”研究应运而生。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。短文本表示建模是一项对于互联网信息时代机器智能至关重要且充满
挑战的研究任务，有益于众多应用场景，如信息检索、文本分类、自动问答、情
感计算、主题检测和信息推荐等。通过合理的表示建模来“理解”短文本并衍生
相关应用，不仅可以为人们的生活提供诸多便利，而且符合企业营销和国家战略
的需要，也是基于关键字匹配策略的传统网络空间信息处理技术达到一定瓶颈之
后的必然选择。
由于设计思想和具体应用的多样性，短文本表示建模方法众多，研究成果非
常分散，这不利于初学者在短时间内系统地掌握这方面的方法和技术。因此，本
书着重挑选具有代表性的短文本概念化表示建模方法和短文本向量化表示建模方
法进行详细介绍，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
笔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对具体
应用问题的求解。本书既尝试通过理论分析，探究以概念化和向量化为代表的各
类短文本表示建模方法的原理本质，又借助理论算法对实际应用进行实验与验证，
以便读者对短文本表示建模有更“立体”的把握。
本书内容（除第1 章外）可分为四部分。部分，简要介绍了短文本表示
建模研究相关的自然语言处理基础理论（第2 章）和支撑短文本表示建模研究的
知识库资源（第3 章）。第二部分，介绍了短文本表示建模研究的方法理论基础，
从显式语义建模（第4 章）、半显式语义建模（第5 章）、隐式语义建模（第6 章）
等三个维度切分短文本表示建模理论模型体系，分别简要介绍相关典型理论及现
2 短文本表示建模及应用
有研究不足。第三部分，针对短文本概念化表示建模（第7 章）和短文本向量化
表示建模（第8 章），分别阐释了问题描述、方法综述、总结分析。第四部分，探
讨和验证了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在短文本检
索中的应用（第9 章），总结全书并展望未来研究趋势（第10 章）。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。
本书在撰写过程中得到了北京理工大学计算机学院、北京市海量语言信息处
理与云计算应用工程技术研究中心、社会安全风险感知与防控大数据应用国家工
程实验室的老师和学生的支持和帮助；本书的出版得到了北京理工大学计算机学
院的大力支持。笔者在此对给予支持和资助的单位与个人表示衷心感谢！
由于笔者水平有限，书中疏漏之处在所难免，敬请读者批评指正。
王亚珅 黄河燕
|
|
購物車/收銀台(
0
) | 在線留言板
| 付款方式
| 運費計算
| 聯絡我們
| 幫助中心 |
加入書簽
HOME
新書上架
