Pattern recognition with fuzzy neural network

This paper explains the character code recognition with the Boolean classifier. The binary values are used both for inputs and outputs, while the learning of the circuit with a set of patterns is done by modified algorithms used in some Boolean neural networks. The use of the fuzzy logic approach offers the possibility of creating a character recognition theory which is fault-tolerant and applicable to all sorts of typefaces and fonts. It provides several examples of patterns scanned with different resolutions and learned with a part of the same set of samples which demonstrates the quality of the fuzzy Boolea classifier.     

基于无监督图神经网络的学术文献表示学习研究

学术文献特征表示,是学术文献搜索、分类组织、个性化推荐等学术大数据服务的关键步骤。研究表明,图神经网络能够有效学习文献的特征表示,然而当前研究主要集中在有监督学习方法上,不仅对数据集的大小和质量的要求较高,且学习到的文献特征表示与具体任务高度耦合。基于此,本文将四种无监督图神经网络方法引入学术文献表示学习,从Cora、CiteSeer和DBLP (database systems and logic programming)数据集的引文网络、共被引网络和文献耦合网络中学习文献的表示向量,并应用于文献分类和论文推荐两大下游任务。研究结果表明,(1)深度互信息图神经网络适合于文献分类任务,对抗正则化变分图自编码器则在论文推荐任务上性能更佳;(2)Cora数据集上的结果表明,相较于共被引和文献耦合网络,引文网络更适合于学习通用的文献表示向量。     

Hierarchical Text Categorization Using Neural Networks

This paper presents the design and evaluation of a text categorization method based on the Hierarchical Mixture of Experts model. This model uses a divide and conquer principle to define smaller categorization problems based on a predefined hierarchical structure. The final classifier is a hierarchical array of neural networks. The method is evaluated using the UMLS Metathesaurus as the underlying hierarchical structure, and the OHSUMED test set of MEDLINE records. Comparisons with an optimized version of the traditional Rocchio's algorithm adapted for text categorization, as well as flat neural network classifiers are provided. The results show that the use of the hierarchical structure improves text categorization performance with respect to an equivalent flat model. The optimized Rocchio algorithm achieves a performance comparable with that of the hierarchical neural networks.     

新世纪国际人工智能研究领域可视化分析

目的：全面了解新世纪国际人工智能领域的研究现状与研究热点。方法：运用TDA软件,利用文献统计分析、关键词共现分析的方法揭示研究热点。结果：国际人工智能领域文献量呈上升趋势,其中美国发文量排名第一,中国位居第六。人工智能的主研究领域包括计算机科学、工程学、自动化与控制系统3个学科。研究热点为遗传算法、神经网络和机器学习。结论：新世纪国际人工智能研究涉及多个学科,研究热点集中在知识获取、知识表示和问题求解3个宏观层面,神经网络、遗传算法和机器学习是研究者们关注的核心。     

Enabling new interactions with library digital collections: automatic gender recognition in historical postcards via deep learning

The Walter Havighurst Special Collections from University Archives & Preservation at Miami University's King Library has a growing collection of over 600,000 historical postcards, with approximately 30,000 digitized, primarily from the Midwest during 1890–1919. This collection supports various lines of inquiry from users, such as analyzing the evolution of gender portrayal in popular media in the United States. However, manually separating the collection into postcards of males and females would take thousands of hours, which prevents the library from supporting sociological analyses at scale. After assembling an open postcard dataset, we trained deep neural networks (i.e., YOLOv5x object detection models) to automatically detect people and classify them as male or female. Our approach limited biases in favor of one outcome by balancing the number of males and females via multi-label stratified 10-fold cross-validation. We showed that this approach can accurately detect and classify females and confidently detect and label males for the library's collection of historical postcards. Our precision of 94.9 % and recall of 33.0 % from 1890 to 1919 on male gender detection exceed the performances of 94.7 % and 31 % respectively for recognition on World War I postcards in past studies. By employing our trained deep neural networks, the library can enhance its metadata within hours and support new research inquiries at scale.     

动态约束性概念网络与知识检索研究

本文在神经网络与语义网络的理论基础上 ,构建了动态约束性概念网络 ,介绍了概念网络的控制约束性与时变动态性等特征 ,并提出了基于概念网络的知识检索模型和启发式动态搜索算法     

情感分析研究的知识结构及热点前沿探析

为了解国内外情感分析领域的研究状况,揭示该领域的知识结构、研究热点与发展动态,本文采用共被引分析、聚类分析、共词分析、战略坐标分析等方法,借助CiteSpace、UCINET、BICOMB、SPSS等软件,对Web of Science数据库收录的以情感分析为主题的相关文献进行计量分析与知识图谱绘制。分析结果表明,情感分析的应用、深度学习与神经网络、电子商务下的产品评论、事物情感特征评分、社交网络下用户生成内容、语义定向广告技术以及文本语言属性分析构建了情感分析的知识结构,产品评论与口碑、数据挖掘与人工智能、无监督学习、HadoopMapReduce与支持向量机以及神经网络与深度学习为该领域的研究热点,而顾客评论、推荐系统、极性分类、主题模型、电影评论、推特数据将是未来该领域主要研究方向。     

计算机图像信息资源管理研究*

首先简要分析计算机图像信息资源管理的研究思路。重点探讨开发图像信息资源管理系统过程中，图像信息管理应用人员在对它们的资源组织、获取与表示、存储与检索、识别与监测、网络化共享、应用创新等过程中的需要注意的问题，以便促进图像信息资源在社会主义现代化建设中发挥更大的作用。     

Is plus size equal? The positive impact of average and plus-sized media fashion models on women’s cognitive resource allocation,social comparisons,and body satisfaction

During individual laboratory sessions, 49 women with an actual-ideal self-discrepancy randomly viewed 12 images of media fashion models varying in body types. Heart rate was recorded during image exposure. Self-report social comparisons and body satisfaction were measured following exposure to each image. A visual recognition test was administered following the last image. The results indicated that women reported the greatest body satisfaction and the least amount of social comparisons when viewing plus size models, but body satisfaction decreased and social comparisons increased when viewing average sized followed by thin size models. Further, as social comparisons increased (e.g., internal processing), external resource allocation and encoding decreased. The theoretical and applied implications from these findings are discussed.     

生成式自动文摘的深度学习方法综述

自动文摘是文本挖掘的主要任务之一。相比于抽取式自动文摘,生成式自动文摘在思想上更接近人工摘要的过程,具有重要研究意义。近几年伴随着深度学习方法的发展,基于深层神经网络模型的生成式自动文摘也有了令人瞩目的发展。为了更全面地理解该类方法的思想和研究现状,本文从生成式自动文摘的任务描述入手,梳理了基于RNN (recurrent neural network,循环神经网络)的模型、基于CNN (convolutional neural network,卷积神经网络)的模型、基于RNN+CNN的模型、融合注意力机制的模型和融合强化学习的模型共五大类生成式自动文摘的深度学习方法。这类方法表明,在深层神经网络的训练下,特别是融合注意力机制和强化学习后,摘要效果得以明显提升。在生成式自动文摘研究的未来发展中,除深度学习方法本身的不断应用和改进外,还需关注如何有效实现篇章级语义理解下的摘要、面向不同文本对象特点的摘要和摘要结果自动评价等问题。此外,如何结合传统摘要研究中的成熟方法进一步提高摘要效果,也是一个很有价值的研究方向。     

A parallel neural network implementation in a distributed fault diagnosis system

A part of a transputer based expert diagnostic system, used for fault diagnosis in a particle physics detector, is composed of a neural network performing a classification of failures through a suitable visual pattern recognition. The neural network is able to give a response with regard to some typical malfunctions in the apparatus by means of the values of a histogram of the particle position during several events.This is a software implementation of a feedforward neural network on a transputer based system. This solution is adopted to speedup the computation time of the network.The neural network outputs are used by the expert system and integrated with other information to perform an efficient diagnosis.     

神经网络技术在汉语歧义切分中的应用

针对目前汉语自动分词系统中切分歧义的难点,本文提出利用神经网络模式识别的方法帮助消歧。介绍了所建立的实验系统,并进行了实验分析。     

The algebra of Fourier-dual operations: Logic with exception

In the development of an approach to the implementation of fuzzy logics on neural networks with interconnections via the scheme of Fourier holography, a model of logic with exception is proposed, which is associated with the basic Generalized Modus Ponens rule. The exception is recalled from the associative memory by an inference formed by the basic rule and it modifies the original inference. The results of numerical simulation are based on the experimental data.     

Integrative model for discovering linked topics in science and technology

Linked topics in science and technology (LTSTs) can provide new avenues for technological innovation and are a key step in the transition from basic to applied research. This paper proposes a science and technology semantic linkage integration model for discovering LTSTs. Particularly, the integrative model fuses the term co-occurrence networks of basic and applied research, which expands the completeness of topic networks by enhancing the semantic characteristics of these networks. It is found that link prediction can further reinforce the semantic association of topic terms in networks between basic and applied topics. Simple fusion explicitly linked the topic terms, which can be used as automatic seed marking for subsequent link prediction to identify implicit linking of topic terms. Furthermore, an application to the gene-engineered vaccines field depicted that newly predicted implicit relations can effectively identify LTSTs. The results also show that implicit semantic recognition of LTSTs can be enhanced through simple fusion, while the recognition of LTST can be improved through link prediction. Therefore, the proposed model can assist experts to identify LTSTs that cannot be recognized through simple fusion.     

Predicting and recommending collaborations: An author-, institution-, and country-level analysis

This study examines collaboration dynamics with the goal to predict and recommend collaborations starting from the current topology. Author-, institution-, and country-level collaboration networks are constructed using a ten-year data set on library and information science publications. Different statistical approaches are applied to these collaboration networks. The study shows that, for the employed data set in particular, higher-level collaboration networks (i.e., country-level collaboration networks) tend to yield more accurate prediction outcomes than lower-level ones (i.e., institution- and author-level collaboration networks). Based on the recommended collaborations of the data set, this study finds that neighbor-information-based approaches are more clustered on a 2-D multidimensional scaling map than topology-based ones. Limitations of the applied approaches on sparse collaboration networks are also discussed.     

纸质档案数字化成果原始分辨率篡改的检测方法

纸质档案数字化成果的原始分辨率是光学字符识别（OCR）精度的重要影响因素,为保证档案图像的原始分辨率大于等于300dpi,在档案预检阶段需检测出较低原始分辨率被篡改为较高分辨率的档案图像。针对图像文件头篡改、图像内容插值等多种分辨率篡改手段,分析它们的篡改机理,设计了由文件头检测、图像频率域分析、图像质量客观评价多方案组成的综合的篡改检测策略。在宁波市档案局的测试数据库上显示,所提出的原始分辨率篡改检测方法取得了很好的效果。     

A neural network-based replacement strategy for high performance computer architectures

We propose a new scheme for the replacement of cache lines in high performance computer systems. Preliminary research, to date, indicates that neural networks (NNs) have great potential in the area of statistical predictions [1]. This attribute of neural networks is used in our work to develop a neural network-based replacement policy which can effectively eliminate dead lines from the cache memory by predicting the sequence of memory addresses referenced by the central processing unit (CPU) of a computer system. The proposed strategy may, therefore, provide better cache performance as compared to the conventional schemes, such as: LRU (Least Recently Used), FIFO (First In First Out), and MRU (Most Recently Used) algorithms. In fact, we observed from the simulation experiments that a carefully designed neural network-based replacement scheme does provide excellent performance as compared to the LRU scheme. The new approach can be applied to the page replacement and prefetching algorithms in virtual memory systems.     

21世纪高校图书馆“形象工程”建设策略研究

高校图书馆的内部形象和外部形象分别影响到高校的教学质量和知名度,文章从图书馆"形象工程"出发,重点探讨了高校图书馆在图书馆"形象工程"建设方面的策略,并对高校图书馆数字化建设方面提出了针对性的建议.     

THE ESPACES STUDY: DESIGNING,DEVELOPING AND MANAGING LEARNING SPACES FOR EFFECTIVE LEARNING

This paper was given as a key note at the North West Academic Libraries conference ‘Designing Spaces for Learning: Flexible learning spaces, libraries and changing roles’ in September 2006 and was adapted from the eSpaces study so that it met the aims and objectives of the North West Academic Libraries conference. The eSpaces study is shorthand for the Joint Information Systems Committee's ‘Study on How Innovative Technologies are Influencing the Design of Physical Learning Spaces in the Post-16 Sector’. The need for this study came from a recognition that most, if not all, education institutions are now integrating learning technologies into the design of new buildings and the refurbishment of existing ones. Managed learning environments, mobile computing, wireless local area networks and broadband are just a few of the technologies that are influencing how we design, use and manage our learning spaces. They are also changing how we design our libraries. The key findings of this study are outlined and the implications for library space and staff roles discussed.     

Neural Network Agents for Learning Semantic Text Classification

The research project AgNeT develops Agents for Neural Text routing in the internet. Unrestricted potentially faulty text messages arrive at a certain delivery point (e.g. email address or world wide web address). These text messages are scanned and then distributed to one of several expert agents according to a certain task criterium. Possible specific scenarios within this framework include the learning of the routing of publication titles or news titles. In this paper we describe extensive experiments for semantic text routing based on classified library titles and newswire titles. This task is challenging since incoming messages may contain constructions which have not been anticipated. Therefore, the contributions of this research are in learning and generalizing neural architectures for the robust interpretation of potentially noisy unrestricted messages. Neural networks were developed and examined for this topic since they support robustness and learning in noisy unrestricted real-world texts. We describe and compare different sets of experiments. The first set of experiments tests a recurrent neural network for the task of library title classification. Then we describe a larger more difficult newswire classification task from information retrieval. The comparison of the examined models demonstrates that techniques from information retrieval integrated into recurrent plausibility networks performed well even under noise and for different corpora.