基于动态常识推理与多维语义特征的幽默识别

随着社交媒体的飞速发展,幽默识别任务受到研究者广泛关注。其目标是判断给定文本是否具有幽默表达。现有方法主要基于幽默产生理论,采用规则或神经网络模型提取多种幽默相关特征,如不一致性、情感和语音等。然而,这些方法未充分捕捉文本内部的情感特征,忽视了隐含在幽默文本中的情感表达,影响了幽默识别的准确性。为解决此问题,该文提出了CMSOR方法,以动态常识和多维语义特征为驱动。首先,利用外部常识信息从文本中动态推理说话者的隐式情感表达;然后,引入WordNet词典计算词级语义距离,捕捉不一致性,并计算模糊性特征;最后,基于这三个特征维度构建幽默语义,实现幽默识别。实验证明,CMSOR模型相对于当前基准模型在三个公开数据集上的识别性能均有显著提升。     

文本幽默识别综述：从数据到方法

幽默计算研究致力于利用计算机方法理解和识别幽默表达,挖掘幽默潜在的语义内涵,构建面向幽默的计算模型,实现幽默的自动识别和生成,提升人机交互智能程度.在开发基于幽默计算的人工智能系统的需求日益凸显的趋势下,通过文献调研方式进行幽默识别综述.首先,重点研究了幽默特征的提取方法;其次,从数据和方法两个维度总结了文本幽默识别的研究进展.归纳常用数据集的收集标注过程及特点,系统地对比了包括基于传统机器学习和基于深度学习的文本幽默识别方法;最后,对幽默识别领域的相关研究进行了总结与展望.     

基于潜在语义特性的语义双关语检测及双关词定位

语义双关语是幽默、笑话和喜剧等作品的来源之一,在人类写作的发展进程中具有重要的历史地位。由于语义双关语存在歧义难懂的特点,因此难以挖掘语义双关语的潜在语义信息,故目前语义双关语的检测和双关词的定位是自然语言处理任务中的一项困难和挑战。该文在语义双关语的理论基础上,挖掘了一系列的潜在语义特性,并构建了对应每个特性的特征集,用以检测语义双关语;同时从潜在语义特性出发,提出了一种基于词向量和同义词融合的语义相似度匹配算法实现语义双关词的定位。在SemEval 2017 Task 7和Pun of the Day数据集上均取得了较好的实验结果,验证了该文所提出的检测算法和定位算法。     

基于BDI的语义搜索及其在数字旅游中的应用

目前的上下文处理方法对语义技术支持不足,导致语义搜索在精准度和智能化上存在局限。提出一种基于BDI(BeliefDesire-Intention)的语义搜索方法,对动态上下文进行语义描述,利用BDI机制优化搜索结果。该方法将用户查询、动态上下文与BDI框架的Belief融合,通过产品特征语义、业务语义推理上下文特征和搜索目标特征作为Desire,通过基于Min Hash的语义相似度计算进行候选Intention排序。结合项目需求,在JADEX开源BDI Agent环境下,开发了基于上述语义搜索的旅游产品搜索系统。该系统通过ID3决策树提取用户当前行程的上下文特征实现了语义搜索功能,实验表明该方法搜索准确率相比仅使用语义相似度搜索方法明显提高,能够满足旅游产品搜索的要求。     

语言知识驱动的词嵌入向量的可解释性研究

神经网络语言模型应用广泛但可解释性较弱,其可解释性的一个重要而直接的方面表现为词嵌入向量的维度取值和语法语义等语言特征的关联状况。先前的可解释性工作集中于对语料库训得的词向量进行知识注入,以及基于训练和任务的算法性能分析,对词嵌入向量和语言特征之间的关联缺乏直接的验证和探讨。该文应用基于语言知识库的伪语料法,通过控制注入语义特征,并对得到的词嵌入向量进行分析后取得了一些存在性的基础性结论: 语义特征可以通过控制注入到词嵌入向量中;注入语义特征的词嵌入向量表现出很强的语义合成性,即上层概念可以由下层概念表示;语义特征的注入在词嵌入向量的所有维度上都有体现。     

基于分布的词汇级语义相关度计算综述

在数字化智能信息处理领域,词汇级语言对象在语义上的相关关系可以为多种研究问题提供有效的特征线索。语义相关度计算是语义相关关系的量化手段,而基于分布相似度的计算方法是一类最典型的方法。这类方法将语言对象被转化为语义空间上的一个分布,通过分布的相似性评估对应语言对象的语义相关度。本文详细介绍了基于上下文分布、基于知识资源元素分布两种形式的代表性方法,并从基础资源的规模、质量、可扩展性三个角度,对这些方法进行了总结。     

基于多粒度语义交互理解网络的幽默等级识别

幽默在人们日常交流中发挥着重要作用.随着人工智能的快速发展,幽默等级识别成为自然语言处理领域的热点研究问题之一.已有的幽默等级识别研究往往将幽默文本看作一个整体,忽视了幽默文本内部的语义关系.该文将幽默等级识别视为自然语言推理任务,将幽默文本划分为"铺垫"和"笑点"两个部分,分别对其语义和语义关系进行建模,提出了一种多...     

基于语义网的产品配置知识表达及检索

基于语义网的产品配置知识表达,通过产品配置本体为产品配置领域知识建模,有效地描述产品配置知识库,解决信息检索和信息集成在语义模糊、语义异构等难点上的困扰;强调配置知识的共享和重用;提供异构系统间数据共享和数据交换的途径．从知识本体的角度出发,详述构成产品配置本体的部件集、部件属性集、配置关系、评价规则集合．文中介绍了一个基于语义网的产品配置知识的原型系统SWB及其检索部分的体系结构,并给出其查询语言的巴克斯范式(BNF)描述．最后给出一个SWB上的查询实例．     

融合量子干涉信息的双重特征文本表示模型

在信息检索领域,量子干涉理论已应用于文档相关性、次序效应等核心问题的研究中,旨在建模用户认知引起的类量子干涉现象.文中从语言理解的需求出发,利用量子理论的数学工具分析语义组合过程中存在的语义演化现象,提出融合量子干涉信息的双重特征文本表示模型(Quantum Interference Based Duet-Feature Text Representation Model, QDTM).模型以约化密度矩阵为语言表示的核心组件,有效建模维度级别的语义干涉信息.在此基础上,构建捕获全局特征信息与局部特征信息的模型结构,满足语言理解过程中不同粒度的语义特征需求.在文本分类数据集和问答数据集上的实验表明,QDTM的性能优于量子启发的语言模型和神经网络文本匹配模型.     

纯XML数据库语义缓存综述

介绍了语义缓存、XML查询语言等概念,对现有的多种纯XML数据库语义缓存技术及其实现、技术特点等进行了详细的阐述。在讨论了当前语义缓存的研究中遇到的新问题与挑战之后,介绍了一种语义缓存匹配视图的快速查找算法U ViewMatch,并总结了语义缓存的技术路线。     

基于多维潜在语义特征的幽默识别

幽默是人类独有的品质,在日常交际中发挥着重要作用.随着人工智能的快速发展,如何让计算机识别幽默成了自然语言处理研究领域的热门研究内容之一.该文针对幽默的自动识别问题,基于幽默理论和领域知识,系统地分析总结了幽默的五类显著特性,包括不一致特性、模糊特性、情感特性、语音特性和句法结构特性,并针对每一类特性构建了多种幽默特征...     

基于BERT的强化语境与语义信息的对话幽默识别模型

幽默识别是自然语言处理的新兴研究领域之一。对话的特殊结构使得在对话中的幽默识别相较于短文本幽默识别更具有挑战性。在对话中,除了当前话语以外,上下文语境信息对于幽默的识别也至关重要。因此,该文在已有研究的基础上结合对话的结构特征,提出基于BERT的强化语境与语义信息的对话幽默识别模型。模型首先使用BERT对发言人信息和话语信息进行编码,其次分别使用句级别的BiLSTM、CNN和Attention机制强化语境信息,使用词级别的BiLSTM和Attention机制强化语义信息。实验结果表明,该文方法能有效提升机器识别对话中幽默的能力。     

基于产品特征向量的产品配置知识表达

针对产品配置对不同产品的适应能力差、配置知识表达复杂的现状,提出基于产品特征向量的知识表达方法。在产品特征可量化的条件下,建立基于产品特征向量的产品配置模型,给出知识表达的理论基础和实现步骤。产品特征向量的引入,实现了对产品特征的自然语言描述向数学语言描述的转化;捆绑在组件上的约束和产品特征向量相关,使约束表达直观、集中,降低了知识表达的难度,并给出成功应用案例。     

THE ACT OF CREATING HUMOROUS ACRONYMS

Our species cannot survive without humor and future human-machine interaction systems will be required to handle humor. From a practical point of view, humor is an important resource for getting selective attention, help in memorizing names and situations, etc. Even if deep modeling of humor in all of its facets is not something available in the near future, there is something concrete that has been achieved and that can help in providing attention to the field. The paper refers to the results of HAHACRONYM, a project devoted to humorous acronym production, a circumscribed task that nonetheless requires various generic components. The project opens the way to developments for creative language. Electronic commerce, for instance, will include flexible and individual-oriented humorous promotion more or less as it happened in the world of broadcasted advertisement.     

An attention network via pronunciation,lexicon and syntax for humor recognition

Humor is one of the most common and attractive expressions in our daily life. It is usually witty and funny. Humor recognition is an interesting but difficult task in natural language processing. Some recent works have used deep neural networks to recognize humorous text. In a different approach, we start from a new perspective based on humor linguistics, including pronunciation, lexicon, and syntax, for recognizing humor based on neural networks, in order to capture humorous incongruity and ambiguity. Specifically, we propose an attention network via pronunciation, lexicon, and syntax (ANPLS) for humor recognition. The ANPLS model contains four units, namely, the pronunciation understanding unit, the lexicon understanding unit, the syntax analysis unit, and the context understanding unit. The pronunciation understanding unit is used to extract the pronunciation-based humor features. The lexicon understanding unit is used to solve the polysemy in humor. The syntax analysis unit aims to capture the syntax information of humor. The context understanding unit is used to obtain the contextual humor features. These four units may have different levels of importance for humor recognition so that we further apply an attention mechanism to assign different weights to these four units. We conduct experiments on three popular datasets, namely, the SemEval2017 Task7 dataset, the 16000 One-Liners dataset, and the Pun of the Day dataset. The experimental results demonstrate that our model can achieve comparable or state-of-the-art performance compared with the existing models.

     

产品数据模型的本体知识表达

针对数据建模不能很好满足产品知识集成与共享的需求,提出产品数据建模需要引入基于本体的形式化语义信息表达机制;在分析EXPRESS语义的基础上,构建描述逻辑语言ALCNRP(D)表达EXPRESS SCHEMA中本体语义知识,为产品全生命周期知识交换与共享建立描述和推理的基础．     

Computational humor

In an effort to focus on tractable problems, computational natural language understanding systems have typically addressed language phenomena that are amenable to combinatorial approaches using static and stereotypical semantic representations. Although such approaches are adequate for much of language, they're not easily extended to capture humans' more creative language interpretation capacities. This paper proposes a model of humor comprehension based on frame-shifting within a simulation-based natural-language-understanding system.     

Types of humor that robots can play

Although humor is a well-known social lubricant defusing a complicated conflict between two parties, the efficacy of humor in human–robot interaction has barely been tested yet. This study compared the characteristics of humor performed by a robot and human to identify the possible type of jokes that a robot may play.In the experiment, a human actor performed disparaging – racist and sexist jokes, and non-disparaging (human condition and sexual) jokes, and a robot counterpart mimicked the same performance. Fifty-eight university students, 30 male and 28 female with mean age 23.10 (SD = 2.00), watched the randomly assigned jokes performed either by the robot or the human actor. The participants rated perceived humorousness, offensiveness, and willingness to share the joke with others, the perceived social presence and social attractions of the actor. The result showed that participants perceived non-disparaging jokes to be more humorous when performed by the human actor. On the other hand, the participants exhibited less disgust toward disparaging jokes when they were performed by the robot actor. This shows that humor can be used as an effective way to enrich the interaction between human and robot; but the acceptable types of humor should be carefully selected.     

The Construction of Away Messages: A Speech Act Analysis

Previous research suggests that "away messages" in instant messaging express informational and entertainment communicative goals while displaying a users' identity. This study investigated the extent to which these communicative goals are reflected in the language structure of away messages, by examining the speech acts performed through the production of 483 away messages crafted by 44 participants. The messages were also analyzed for the use of non-standard orthography and humor. The results show that the messages were constructed primarily with assertives, followed by expressives and commissives, but rarely with directives, confirming that away messages tend to reflect both informational and entertainment goals. Non-standard orthography and humor were also common, although experienced participants used fewer non-standard forms than less experienced participants. These findings are discussed in terms of computer-mediated discourse and online self-presentation.