Auto-grading,as an instruction tool,could reduce teachers’workload,provide students with instant feedback and support highly personalized learning.Therefore,this topic attracts considerable attentions from researcher...Auto-grading,as an instruction tool,could reduce teachers’workload,provide students with instant feedback and support highly personalized learning.Therefore,this topic attracts considerable attentions from researchers recently.To realize the automatic grading of handwritten chemistry assignments,the problem of chemical notations recognition should be solved first.The recent handwritten chemical notations recognition solutions belonging to the end-to-end trainable category suffered fromthe problem of lacking the accurate alignment information between the input and output.They serve the aim of reading notations into electrical devices to better prepare relevant edocuments instead of auto-grading handwritten assignments.To tackle this limitation to enable the auto-grading of handwritten chemistry assignments at a fine-grained level.In this work,we propose a component-detectionbased approach for recognizing off-line handwritten Organic Cyclic Compound Structure Formulas(OCCSFs).Specifically,we define different components of OCCSFs as objects(including graphical objects and text objects),and adopt the deep learning detector to detect them.Then,regarding the detected text objects,we introduce an improved attention-based encoder-decoder model for text recognition.Finally,with these detection results and the geometric relationships of detected objects,this article designs a holistic algorithm for interpreting the spatial structure of handwritten OCCSFs.The proposedmethod is evaluated on a self-collected data set consisting of 3000 samples and achieves promising results.展开更多
In recent years,the Mars Text,a kind of newly born Internet language,has gained its surprising popularity among the younger generation,which arouses the intense opposition from the adults. With the purpose of having a...In recent years,the Mars Text,a kind of newly born Internet language,has gained its surprising popularity among the younger generation,which arouses the intense opposition from the adults. With the purpose of having a profound and proper understanding of it,a comparative research of the characteristics of the Mars Text and Chinese has been conducted to identify the nature of the former,followed by discussions of the reasons behind its development and the current attitudes held by different social groups towards it. In summary,the Mars Text,as a creative symbol system,deserves respect from social members,but needs to be utilized in a culturally appropriate way.展开更多
Genre is a prevailing term in various disciplines and field, carrying different denotations and connotations in each branch of knowledge. By comparing how genre is defined, categorized and interpreted in different fie...Genre is a prevailing term in various disciplines and field, carrying different denotations and connotations in each branch of knowledge. By comparing how genre is defined, categorized and interpreted in different fields and traditions, this essay tends to provide a better understanding of genre as a term, a concept, a perspective and an approach.展开更多
Court interpreting as a type of dialogue or liaison interpreting has recently drawn an increasing level of attention in a variety of disciplines, including legal science, applied linguistics, and translation studies. ...Court interpreting as a type of dialogue or liaison interpreting has recently drawn an increasing level of attention in a variety of disciplines, including legal science, applied linguistics, and translation studies. Amongst discussions relevant to court interpreting, the area of sight translation has yet to be emphasized, despite its importance to the outcome of a trial. In this study I first had observations of court interpreting activities on several occasions. Based on these observations, I then formulated a questionnaire and distributed it to 86 interpreters in two training workshops organized by the Taiwan High Court in 2013. The key findings are as follows： most interpreters consider sight translation to be more difficult than other forms of interpretation in the courtroom; while conducting sight translation many interpreters pause in order to seek clarification or are interrupted; most interpreters would prefer the judge or the prosecutor to orally summarize the text to be sight translated; and, after ＂negotiation,＂ most practitioners tend to lower the level of formality when dealing with a text written in a formal style. These findings imply that interpreters conducting sight translation render a translation through ＂negotiation＂ with other participants of the activity; this negotiation can be seen as an attempt to build a ＂context＂ from the perspective of liaison interpreting. This study intends to raise the awareness of court interpreters regarding a number of key issues in sight translation, and the results are hoped to be conducive to the future study of other types of liaison interpreting and interpreting as a whole.展开更多
Many of the literary works of Matthew Arnold reflect the contradictory psychology in the 19th century,of which Dover Beach ranks high..It touches the major concerns of Mid-Victorian writing.Since the readers have good...Many of the literary works of Matthew Arnold reflect the contradictory psychology in the 19th century,of which Dover Beach ranks high..It touches the major concerns of Mid-Victorian writing.Since the readers have good acknowledgement of the extrinsic approach,to analyze this poem from a New Criticism point of view might better and objectively probe the central idea of it.展开更多
结合胃镜超声和白光内镜可以更准确地识别胃肠道间质瘤.但是现有的多模态方法往往仅关注于图像特征,忽略了诊断文本信息中所包含的语义信息对于精确理解和诊断医学图像的重要性.为此,本文提出一种新的基于文本引导下的多模态医学图像分...结合胃镜超声和白光内镜可以更准确地识别胃肠道间质瘤.但是现有的多模态方法往往仅关注于图像特征,忽略了诊断文本信息中所包含的语义信息对于精确理解和诊断医学图像的重要性.为此,本文提出一种新的基于文本引导下的多模态医学图像分析算法框架(Text-guided Multi-modal Medical image analysis framework,TMM-Net).TMM-Net使用多阶段的诊断文本来引导模型学习,以提取图像中的关键诊断信息特征,然后通过交叉模态注意力机制促进多模态特征之间的交互.值得注意的是,TMM-Net通过预测病变属性来模拟临床诊断过程,从而增强了可解释性.验证实验在两个中心包含10 025个模态数据对的数据集上进行.结果表明,该方法相比目前最优的GISTs诊断方法精度提升7.7%,同时获得了最高的(Area Under the Curve,AUC)值:0.927,其可解释性可以更好地适合临床需求.展开更多
基金supported by National Natural Science Foundation of China (Nos.62007014 and 62177024)the Humanities and Social Sciences Youth Fund of the Ministry of Education (No.20YJC880024)+1 种基金China Post Doctoral Science Foundation (No.2019M652678)the Fundamental Research Funds for the Central Universities (No.CCNU20ZT019).
文摘Auto-grading,as an instruction tool,could reduce teachers’workload,provide students with instant feedback and support highly personalized learning.Therefore,this topic attracts considerable attentions from researchers recently.To realize the automatic grading of handwritten chemistry assignments,the problem of chemical notations recognition should be solved first.The recent handwritten chemical notations recognition solutions belonging to the end-to-end trainable category suffered fromthe problem of lacking the accurate alignment information between the input and output.They serve the aim of reading notations into electrical devices to better prepare relevant edocuments instead of auto-grading handwritten assignments.To tackle this limitation to enable the auto-grading of handwritten chemistry assignments at a fine-grained level.In this work,we propose a component-detectionbased approach for recognizing off-line handwritten Organic Cyclic Compound Structure Formulas(OCCSFs).Specifically,we define different components of OCCSFs as objects(including graphical objects and text objects),and adopt the deep learning detector to detect them.Then,regarding the detected text objects,we introduce an improved attention-based encoder-decoder model for text recognition.Finally,with these detection results and the geometric relationships of detected objects,this article designs a holistic algorithm for interpreting the spatial structure of handwritten OCCSFs.The proposedmethod is evaluated on a self-collected data set consisting of 3000 samples and achieves promising results.
文摘In recent years,the Mars Text,a kind of newly born Internet language,has gained its surprising popularity among the younger generation,which arouses the intense opposition from the adults. With the purpose of having a profound and proper understanding of it,a comparative research of the characteristics of the Mars Text and Chinese has been conducted to identify the nature of the former,followed by discussions of the reasons behind its development and the current attitudes held by different social groups towards it. In summary,the Mars Text,as a creative symbol system,deserves respect from social members,but needs to be utilized in a culturally appropriate way.
文摘Genre is a prevailing term in various disciplines and field, carrying different denotations and connotations in each branch of knowledge. By comparing how genre is defined, categorized and interpreted in different fields and traditions, this essay tends to provide a better understanding of genre as a term, a concept, a perspective and an approach.
文摘Court interpreting as a type of dialogue or liaison interpreting has recently drawn an increasing level of attention in a variety of disciplines, including legal science, applied linguistics, and translation studies. Amongst discussions relevant to court interpreting, the area of sight translation has yet to be emphasized, despite its importance to the outcome of a trial. In this study I first had observations of court interpreting activities on several occasions. Based on these observations, I then formulated a questionnaire and distributed it to 86 interpreters in two training workshops organized by the Taiwan High Court in 2013. The key findings are as follows： most interpreters consider sight translation to be more difficult than other forms of interpretation in the courtroom; while conducting sight translation many interpreters pause in order to seek clarification or are interrupted; most interpreters would prefer the judge or the prosecutor to orally summarize the text to be sight translated; and, after ＂negotiation,＂ most practitioners tend to lower the level of formality when dealing with a text written in a formal style. These findings imply that interpreters conducting sight translation render a translation through ＂negotiation＂ with other participants of the activity; this negotiation can be seen as an attempt to build a ＂context＂ from the perspective of liaison interpreting. This study intends to raise the awareness of court interpreters regarding a number of key issues in sight translation, and the results are hoped to be conducive to the future study of other types of liaison interpreting and interpreting as a whole.
文摘Many of the literary works of Matthew Arnold reflect the contradictory psychology in the 19th century,of which Dover Beach ranks high..It touches the major concerns of Mid-Victorian writing.Since the readers have good acknowledgement of the extrinsic approach,to analyze this poem from a New Criticism point of view might better and objectively probe the central idea of it.
文摘结合胃镜超声和白光内镜可以更准确地识别胃肠道间质瘤.但是现有的多模态方法往往仅关注于图像特征,忽略了诊断文本信息中所包含的语义信息对于精确理解和诊断医学图像的重要性.为此,本文提出一种新的基于文本引导下的多模态医学图像分析算法框架(Text-guided Multi-modal Medical image analysis framework,TMM-Net).TMM-Net使用多阶段的诊断文本来引导模型学习,以提取图像中的关键诊断信息特征,然后通过交叉模态注意力机制促进多模态特征之间的交互.值得注意的是,TMM-Net通过预测病变属性来模拟临床诊断过程,从而增强了可解释性.验证实验在两个中心包含10 025个模态数据对的数据集上进行.结果表明,该方法相比目前最优的GISTs诊断方法精度提升7.7%,同时获得了最高的(Area Under the Curve,AUC)值:0.927,其可解释性可以更好地适合临床需求.
