On Verbal Humour from Semantic and Rhetorical Perspectives in American Sitcom Friends

Verbal humor in the province of semantics and rhetorics was focused on and the excerpts in American sitcom Friends was taken as its research object.The famous classic theory: the semantic script theory,was given a sketch outline accompanying examples to state how SSTH works in analysing the verbal humor.Moreover,the employment of different rhetoric skills which elicit humourous effects in a powerful way were appreciated in detail.

Contrastive Rhetoric and Inter-language Theory

A learner’s stages of L2 development are connected by his or her L1 and culture.It is,accordingly,of paramount importance to understand the second language learners’culture and learning process and better assist them through this process in the way of teaching them English.Similarly,inter-language theory(IL)and contrastive rhetoric are affected by factors,such as learner’s L1,learning experiences,and culture.This paper talks about these two theory’s characteristics,constructs,and importance,so language instructors may better understand the L2 learning phenomena and think out better methods to help language learners improve their language skills.

Teaching Writing Contrastively： Contrastive Rhetoric in the Teaching of Writing

Writing fluently and expressively in English is a necessary skill for communication, as well as a required skill for Chinese students at the tertiary level. Since college students have already been trained on their Chinese writing skills for a considerable length of time and are able to write in Chinese quite successfully, the application of contrastive rhetoric in the teaching of English writing is very effective. This paper explores the possibility of teaching English writing to Chinese students at the tertiary level and proposes an approach to this teaching model.

Contrastive Rhetoric: Chinese Students and Academic Writin

ＣｏｎｔｒａｓｔｉｖｅＲｈｅｔｏｒｉｃ：ＣｈｉｎｅｓｅＳｔｕｄｅｎｔｓａｎｄＡｃａｄｅｍｉｃＷｒｉｔｉｎｇＬｉＨｕｉｌｉｎ（ＤｅｐａｒｔｍｅｎｔｏｆＦｏｒｅｉｇｎＬａｎｇｕａｇｅ．ＫｕｎｍｉｎｇＩｎｓｔｉｔｕｔｅｏｆＴｅｃｈｎｏｌｏｇｙ，Ｋｕｎｍｉｎ...

A Brief Review of Lexico-semantic Contrasts between English and Chinese

Contrastive linguistics belongs to a branch of linguistics.It studies two or more languages synchronically.However,lexicology is the study of lexical items of a language.An English sentence or a Chinese sentence formed based on words and clauses.Because of differences in Cultures,the lexical meanings between English and Chinese differ in many aspects,such as motivation of words and sense relationships.This paper is a brief review that just likes a book report,mainly based on the book Contrastive Linguistics,of lexico-semantic contrasts between English and Chinese,illustrated with several examples in both English and Chinese.

Study of Second Language Writing:Contrastive Rhetoric——Cultural Thought P atterns in Inter-cultural Education by Robert B.Kaplan

Contrastive rhetoric is an area of research in second language acquisition that identifies problems in composition en?countered by second language writers and, by referring to the rhetorical strategies of the first language, which attempts to explain them.The article'Cultural Thought Patterns in Inter-cultural Education by Robert B.Kaplan' indicates that different ways of under?standing language and writing styles.The English style is linear while the oriental is spiral.The author gives suggestions how to help those from other cultures learn how to write in acceptable academic English.However,he points out this is not because it is the cor?rect way but because it helps the students attain their goal of academic success.A student can write an adequate essay in his native language,can he write an adequate essay in a second language? In the following paragraphs,I will go to the details to present the lit?erature review,some case analysis,discussion and so on. It's about the contractive rhetoric based on the article of Robert B.Kaplan, then give conclusion.

Semantic Features and Applications in Translation of English Words in Pairs

English words in pairs are a special form of English idioms, which have different kinds and are used widely. For English learners, words in pairs are one of the difficult points. This paper discusses their form patterns, semantic relations, grammatical functions, rhetoric features and their application in translation. Its purpose is to help learners understand and use them accurately and correctly so as to improve language expressing ability.

A Corpus-based Contrastive Study on the Acquisition of Synonyms of Chinese EFL Learners

This paper investigates Chinese EFL learners＇ acquisition of completely, totally and absolutely through a contrastive analysis of their collocation, colligation, semantic preference and semantic prosody based on the Ten-thousand English Compositions of Chinese Learners （TECCL Corpus） and Corpus of Contemporary American English （COCA）. The results show that adjective and verb are two major words frequently co-occurring with the group of synonyms, but only a few of them meet the default collocation criterion （frequency〉3, MI〉3）, and many co-occurring words used by Chinese learners don＇t appear in COCA. Besides, the co-occurring adjectives and verbs in TECCL are so diverse that they don＇t establish fixed semantic relations as in COCA, therefore, there is a big difference in semantic preference and semantic prosody between Chinese learners and native speakers. With the advance of English proficiency, more collocates are used by college learners than middle school learners, but there is no clear and significant improvement in semantic preference and semantic prosody.

面向视频数据的多模态情感分析

多模态情感分析旨在从文本、图像和音频数据中提取和整合语义信息,从而识别在线视频中说话者的情感状态。尽管多模态融合方案在此研究领域已取得一定成果,但是已有方法在处理模态间分布差异和关系知识的融合方面仍有欠缺,为此,提出一种多模态情感分析方法。设计一种多模态提示门(MPG)模块,其能够将非语言信息转换为融合文本上下文的提示,利用文本信息对非语言信号的噪声进行过滤,得到包含丰富语义信息的提示,以增强模态间的信息整合。此外,提出一种实例到标签的对比学习框架,在语义层面上区分隐空间中的不同标签以进一步优化模型输出。在3个大规模情感分析数据集上的实验结果表明,该方法的二分类精度相对次优模型提高了约0.7%,三分类精度提高了超过2.5%,达到0.671。该方法能够为将多模态情感分析引入用户画像、视频理解、AI面试等领域提供参考。

语义增强图像-文本预训练模型的零样本三维模型分类

目前,基于对比学习的图像-文本预训练模型(CLIP)在零样本3维模型分类任务上表现出了巨大潜力,然而3维模型和文本之间存在巨大的模态鸿沟,影响了分类准确率的进一步提高。针对以上问题,该文提出一种语义增强CLIP的零样本3维模型分类方法。该方法首先将3维模型表示成多视图;然后为了增强零样本学习对未知类别的识别能力,通过视觉语言生成模型获得每张视图及其类别的语义描述性文本,并将其作为视图和类别提示文本之间的语义桥梁,语义描述性文本采用图像字幕和视觉问答两种方式获取;最后微调语义编码器将语义描述性文本具化为类别的语义描述,其拥有丰富的语义信息和较好的可解释性,有效减小了视图和类别提示文本的语义鸿沟。实验表明,该文方法在ModelNet10和ModelNet40数据集上的分类性能优于现有的零样本分类方法。

融合自监督对比学习的雾天街景语义分割算法

针对雾天环境能见度低导致街道物体识别困难、分割速度慢等问题,提出了一种融合自监督对比学习的雾天街景语义分割算法。该算法选用轻量级网络MobileNetV2作为主干网络,设计深度聚合空洞空间金字塔池化模块,并使用带有扩张系数的深度可分离卷积替换普通卷积丰富特征多样性。然后通过融合对比学习框架,增加语义相似像素的相似度,保持不同语义像素之间的距离,从而提高模型对小目标物体细节边缘的表达能力和辨别能力。最后提出一种新的融合损失函数,采用监督学习和自监督学习共同指导网络训练,学习深度特征表示。实验结果表明,该模型在Foggy Cityscapes数据集上的平均交并比可达到74.35%,类别平均像素准确率为83.59%,像素准确率可达到95.85%,相比语义分割网络DeepLabV3+模型分别提高了3.82%、3.99%和1.02%,同时模型参数量为2.88M,比DeepLabV3+模型的参数量缩减近55%,优化了网络计算消耗。该算法在雾天语义分割中拥有良好的性能,在降低模型参数量的同时保持了高分割精度,具有良好的鲁棒性。

基于半监督学习的路面裂缝检测

针对裂缝自动检测任务中难以获取大量精确标注样本数据的问题,提出LGS-Net(Local Global Similarity-Network)模型。LGS-Net的核心在于利用裂缝图像区域的语义相似性,有效结合少量已标注数据和大量未标注图像数据,通过半监督学习实现裂缝自动检测。为全面评估LGS-Net的性能,实验在GAPs384和Crack500数据集上进行验证。结果表明,在标注资源有限的情况下,LGS-Net能够实现高精度的裂缝检测。通过对检测结果的可视化分析,证明LGS-Net具有在复杂环境下有效识别裂缝的能力。LGS-Net利用路面裂缝图像的语义相似性特征进行检测,能为路面裂缝检测的工程应用提供技术支持。

基于融合编码器与双解码器的半监督疲劳裂纹分割

针对现有基于深度学习的疲劳裂纹分割算法依赖于大量像素级标记的问题,提出了一种半监督疲劳裂纹分割网络SFD-Net.SFD-Net利用对比学习方法进行半监督训练,以减少对大量像素级标记的依赖.同时,它采用融合编码器和双解码器的设计,旨在更好地捕捉裂纹区域的特征并提高分割准确性.通过引入改进注意力模块和边界优化模块,增强了对裂纹特征的表示,提高了裂纹边界的分割质量.在公开的疲劳裂纹数据集上对SFD-Net进行性能验证.结果表明:相较于使用相同标记比例的全监督算法,SFD-Net的分割性能有明显提升;仅使用25%的标记数据时,SFD-Net的交并比(IoU)达到70.6%,超过了使用100%标记数据的其他全监督算法的平均IoU(69.1%);同时,与其他先进的半监督方法相比,SFD-Net在所有标记数据比例下均取得了最高的预测精度.

用于语义分割的自监督对比式表征学习

为了提升语义分割模型的精度,并减少逐像素标注大规模语义分割数据集的人力和时间成本,研究了自监督对比式表征学习的预训练方法,并结合语义分割任务的特点,设计了全局-局部交叉对比学习(GLCCL)方法。该方法将全局图像和局部分块后的一系列图像块输入到网络中分别编码全局和局部视觉表征,并通过构建包含全局对比、局部对比和全局-局部交叉对比的损失函数来指导模型训练,使得模型能够同时学习全局和局部区域的视觉表征以及跨区域语义相关性。使用该方法预训练BiSeNet再迁移到语义分割任务时,对比现有的自监督对比式表征学习和有监督预训练方法分别具有0.24%和0.9%平均交并比(MIoU)的性能提升。实验结果表明,该方法能够采用无标注的数据训练语义分割模型而实现分割效果的提升,具有一定的实用价值。

基于对比学习的大型语言模型反向词典任务提示生成方法

反向词典任务是一种新兴的任务,目的是根据给定的定义来查找对应的单词。大规模语言模型为这一任务提供了新的可能性,但是提示语句的质量会影响大模型的性能。为此,提出了一种基于对比学习的提示生成方法。该方法在从多个语义层面上理解定义语义的同时,还利用对比学习的原理在训练过程中引入了负例,提升了模型的泛化能力。通过这种方法,可以将目标单词缩小到一个小范围内,然后用大模型从这个范围内选择最符合定义语义的单词。实验结果表明,该方法可以有效地提升大规模语言模型在反向词典任务上的表现。提示生成模型有94.7%的概率生成包含目标词的范围,大规模语言模型有58.03%的概率直接选出目标单词,有74.55%的概率在给出5个候选单词时包含目标单词。

图引导的特征融合和分组对比学习的域自适应语义分割

在无监督域自适应语义分割任务中,有效地融合源域和目标域的特征以及解决不同类别像素数量分布不均衡的问题是提升跨域语义分割网络性能的关键。为了充分融合源域和目标域的特征,建立源域和目标域之间的长距离上下文关系,本文构建了双跨域图卷积网络,利用图卷积来引导源域和目标域的特征进行融合。本文分别构造了跨域位置相似矩阵和通道相似矩阵,提出了跨域位置图卷积和跨域通道图卷积。为了解决数据集中存在的类不平衡问题,同时提取到更多域不变特征,本文提出了分组对比学习策略,通过在组内构造正负样本,拉近2个域相同类之间的距离并拉远2个域不同类之间的距离。实验证明,本文提出的方法在数据集GTA5到Cityscapes和SYNTHIA到Cityscapes上的跨域语义分割均取得了良好的效果。

基于无监督语义哈希的高效相似题检索模型

相似题检索旨在从数据库中找到与给定查询试题考查目标相似的试题。随着在线教育的不断发展,试题数据库日益庞大,且由于试题数据的专业属性使标注相关性非常困难,因此需要一种高效且无需标注的相似题检索模型。无监督语义哈希能在无监督信号的前提下将高维数据映射为低维且高效的二值表征。但不能简单地将语义哈希模型应用在相似题检索模型中,因为试题数据具有丰富的语义信息,而二值向量的表征空间有限。为此,提出一个能获取、保留关键信息的相似题检索模型。首先,设计了一个关键信息获取模块获取试题数据的关键信息,并引入去冗余目标损失去除冗余信息;其次,在编码过程中引入随时间变化的激活函数,减少编码信息损失;再次,为了最大化利用汉明空间,在优化过程中引入比特平衡目标和比特无关目标以优化二值表征的分布。在MATH和HISTORY数据集上的实验结果表明,相较于表现最好的文本语义哈希模型DHIM(Deep Hash InfoMax),所提模型在2个数据集的3个召回率设置上分别平均提升约54%和23%;在检索效率方面,所提模型比最优的相似题检索模型QuesCo具有明显的优势。

面向跨模态检索的查询感知双重对比学习网络

近期,跨模态视频语料库时刻检索(VCMR)这一新任务被提出,它的目标是从未分段的视频语料库中检索出与查询语句相对应的一小段视频片段.现有的跨模态视频文本检索工作的关键点在于不同模态特征的对齐和融合,然而,简单地执行跨模态对齐和融合不能确保来自相同模态且语义相似的数据在联合特征空间下保持接近,也未考虑查询语句的语义.为了解决上述问题,提出一种面向多模态视频片段检索的查询感知跨模态双重对比学习网络(QACLN),该网络通过结合模态间和模态内的双重对比学习来获取不同模态数据的统一语义表示.具体地,提出一种查询感知的跨模态语义融合策略,根据感知到的查询语义自适应地融合视频的视觉模态特征和字幕模态特征等多模态特征,获得视频的查询感知多模态联合表示.此外,提出一种面向视频和查询语句的模态间及模态内双重对比学习机制,以增强不同模态的语义对齐和融合,从而提高不同模态数据表示的可分辨性和语义一致性.最后,采用一维卷积边界回归和跨模态语义相似度计算来完成时刻定位和视频检索.大量实验验证表明,所提出的QACLN优于基准方法.

一种基于语义引导和对比学习的战场图像去烟算法

烟雾作为现代战争中最常见的作战产物,不可避免地会降低作战场景的可视性,进而影响下游军事智能系统的性能,因此对含烟图像进行复原处理非常重要,现有算法通常忽略图像中高阶的语义信息和降质图像本身都可以为提高网络去烟能力提供有价值的监督信息。为此,提出一种基于语义引导和对比学习的生成对抗网络来去除战场图像中的烟雾。通过在低阶视觉任务中融入高阶语义特征,将语义信息作为引导帮助网络更好地恢复图像的结构和色彩信息;利用对比学习范式将清晰和含烟图像构建为正、负样本,并采用对比约束使去烟后的图像与清晰图像接近,并远离含烟图像。此外,为模拟真实的战场含烟场景,首次构建一套含烟战场数据集,推进了相关研究的发展。实验结果表明,与现有图像去烟算法相比,所提方法在定量和定性指标上均达到了先进水平。