Data‐driven approaches to information access

This paper summarizes three lines of research that are motivated by the practical problem of helping users find information from external data sources, most notably computers. The application areas include information retrieval, text categorization, and question answering. A common theme in these applications is that practical information access problems can be solved by analyzing the statistical properties of words in large volumes of real world texts. The same statistical properties constrain human performance, thus we believe that solutions to practical information access problems can shed light on human knowledge representation and reasoning.     

“易象”新说——兼论《周易》原有《象经》问题

论文对古代典籍所载"易象"一名作了新的解读。认为所谓"易象",仅是当时《周易》的另一种称谓而已,它并不神秘;"易象"与今本《象传》没有直接联系,尚无渊源,《象传》"演德","易象""尊礼尚施",至汉代仍保存有以古礼诠释《周易》的传统;古代《周易》原有《象经》,或以《卦下易经》当之,惜已不传;但在清人的《易》著中,至今仍可见其传承的轨迹。     

Is it better to be happy or to be proud at school before doing a text comprehension task? First evidence with 10-year-old children

Introduction and objective

An experiment was conducted in a school context to examine whether joy and pride, two positive emotions, can promote children's acquisition of knowledge in a reading comprehension situation.

Robert Singleton's Sermon at Paul's Cross in 1535: The ‘True Church’ and the Royal Supremacy

Abstract

On 2 April 1535, Robert Singleton, chaplain to Queen Anne Boleyn, preached a sermon at Paul's Cross. The unique, surviving printed copy of the sermon is housed in the Wren Library of Lincoln Cathedral. From 1533, Singleton had been a member of the circle of avant-garde evangelicals gathered at the margins of the Tudor court under the patronage of Thomas Cromwell. His sermon at Paul's Cross contributed to an orchestrated campaign in support of Cromwell's legislative programme, then proceding through Parliament, to sever all links of ecclesiastical jurisdiction with Rome and to confirm Henry's headship of the Church of England. The sermon interprets Paul's allegory of the earthly and the heavenly Jerusalem in the Epistle to the Galatians as a model for thinking about the distinction between the invisible and visible Church. Singleton harnesses Pauline ecclesiology in an apology for the royal supremacy. Presented here is a transcribed copy of the original printed version.     

Working memory demands of exact and approximate addition

We compared the working memory requirements of two forms of mental addition: exact calculation (e.g., 63 + 49 = 112) and approximation (e.g., 63 + 49 is about 110). In two experiments, participants solved two-digit addition problems (e.g., 63 + 49) alone and in combination with a working memory task (i.e., remembering four consonants). In Experiment 1, participants chose an answer from two alternatives (e.g., exact: 112 vs. 122; approximate: 110 vs. 140). In Experiment 2, participants responded verbally with exact or approximate answers. In both experiments, the working memory load impaired exact and approximate addition performance, but exact addition was affected more. Load also impaired performance on problems with a carry operation in the units (e.g., 28 + 59 or 76 + 57) more than on problems without a unit carry (e.g., 24 + 53 or 76 + 52). These results identify the carry operation as the source of the working memory demands in multidigit addition.     

文本检索模型综述

文本检索是信息检索一个重要的分支。随着互联网信息的迅速膨胀,如何检索到用户最需要的信息变得越来越关键。文本检索模型是文本检索中的核心技术,其性能直接影响到搜索引擎的检索质量。本文对当前的经典检索模型及其研究进展进行介绍,并分析各个模型之间的优缺点。     

古琴减字谱图像的文本切分与提取

古琴打谱是一项专业性很强又费时费力的工作,如何利用先进的人工智能技术来对古琴谱中的谱字进行自动识别解读,哪怕是辅助性的,对于古琴打谱事业的发展无疑会起到重要的促进作用,从可以间接地为保护与弘扬古琴文化作贡献。本文通过研究古琴减字谱这种特殊文本的特点,提出了不同于普通OCR光学识别软件的文本切分方法,内容涉及纸质古琴谱原始扫描图的图像预处理,古琴混合谱中简字谱的行切分与提取,单行简字谱中的谱字切分与提取等算法及其实现。这样就为实现古琴谱进一步的释读提供一种前期处理方法。由于古琴谱字的特殊性,以及所提出方法的通用性,该方法对于丰富汉字文本的切分方法也有着一定的学术意义。     

An ontology constructing technology oriented on massive social security policy documents

To solve the problem brought from enormous policy documents and complex management in the social security domain, the article uses ontology as the way of representing and storing knowledge. The article first constructs the framework of ontology through manual work so that it can ensure the relative accuracy of the ontology structure. Then it achieves the automatic ontology expansion based on the inclusion relationship of property sets or operational object sets. The article uses a semi-automatic method that extracts hierarchical concepts and non-hierarchical concepts from domain thesaurus by using the method combining statistics with rules to construct the ontology. Besides constructing the ontology, the article proposes the concepts of concept phrase vector model and high frequency characteristics phrase vector model. The experiment result indicates that ontology semi-automatic construction process can help experts to construct the social security ontology effectively oriented on massive policy documents and is a considerable reference for the construction of ontology in other domains.