基于动态分类的隐喻识别方法

隐喻计算是自然语言处理领域中的重要问题.尝试以差异性计算为基础,结合语言、心理和认知的角度对英语隐喻识别进行深入分析和探索.对人类而言,隐喻识别是一个动态分类的过程,动态分类是从多个角度来度量事物之间的差异性.研究了如何模仿人类来获取概念的特征、选择分类角度、在特定分类角度下计算差异性,并进行了英语名词性隐喻识别的实验.该方法对隐喻/常规表达识别的准确率达到85.4%,实验结果表明,该方法是有效的.     

Neural networks for learning in the real world: representation, reinforcement and dynamics

It is argued that the backpropagation learning algorithm is unsuited to tackling real world problems such as sensory-motor coordination learning or the encoding of large amounts of background knowledge in neural networks. One difficulty in the real world - the unavailability of ‘teachers’ who already know the solution to problems, may be overcome by the use of reinforcement learning algorithms in place of backpropagation. It is suggested that the complexity of search space in real world neural network learning problems may be reduced if learning is divided into two components. One component is concerned with abstracting structure from the environment and hence with developing representations of stimuli. The other component involves associating and refining these representations on the basis of feedback from the environment. Time-dependent learning problems are also considered in this hybrid framework. Finally, an ‘open systems’ approach in which subsets of a network may adapt independently on the basis of spatio-temporal patterns is briefly discussed.     

面向分层结构的网页分类与抓取

传统网络爬虫为基于关键字检索的通用搜索引擎服务,无法抓取网页类别信息,给文本聚类和话题检测带来计算效率和准确度问题。本文提出基于站点分层结构的网页分类与抽取,通过构建虚拟站点层次分类树并抽取真实站点分层结构,设计并实现了面向分层结构的网页抓取;对于无分类信息的站点,给出了基于标题的网页分类技术,包括领域知识库构建和基于《知网》的词语语义相似度计算。实验结果表明,该方法具有良好的分类效果。     

Sandy ideas and coloured days: Some computational implications of embodiment

This paper is an exploration of the relationship between language and vision from the perspective of language evolution on the one hand, and metaphor on the other. Recent research has suggested that the origins of human language capacity can be traced to the evolution of a region in the brain that permits the interaction of information from sensory and motor cortices. In light of this, it is hypothesised that the computational mechanisms of language are derived from those of the sensory-motor domain, and that the pervasiveness of metaphor is one manifestation of language's computational antecedants. A variety of cognitive and computational implications are drawn from these hypotheses.     

Distributed Constraint Optimization Problems: Review and perspectives

Intelligent agents is a research area of the Artificial Intelligence intensely studied since the 1980s. Multi-agent systems represent a powerful paradigm of analyzing, projecting, and developing complex systems. One of the main difficulties in modeling a multi-agent system is defining the coordination model, due to the autonomous behavior of the agents. Distributed Constraint Optimization Problems (DCOP) have emerged as one of most important formalisms for coordination and distributed problem solving in multi-agent systems and are capable of modeling a large class of real world problems naturally. This work aims to provide an overview and critical review of DCOP, addressing the most popular methods and techniques, the evolution and comparison of algorithms, and future perspectives on this promising research area.     

基于Stacking组合分类方法的中文情感分类研究

情感文本分类(简称情感分类)是一种面向主观信息分类的文本分类任务。目前,由于其广泛的应用前景,该任务在自然语言处理研究领域中得到了普遍关注,相继出现多种用于情感文本分类的有监督的分类方法。该文具体研究四种不同的分类方法在中文情感分类上的应用,并且采用一种基于Stacking的组合分类方法,用以组合不同的分类方法。实验结果表明,该组合方法在所有领域都能够获得比最好基分类方法更好的分类效果。从而克服了分类方法领域依赖的困境(不同领域需要选择不同基分类方法才能获得更好的分类结果)。     

Sensing through body dynamics

It has been shown that sensory morphology and sensory–motor coordination enhance the capabilities of sensing in robotic systems. The tasks of categorization and category learning, for example, can be significantly simplified by exploiting the morphological constraints, sensory–motor couplings and the interaction with the environment. This paper argues that, in the context of sensory–motor control, it is essential to consider body dynamics derived from morphological properties and the interaction with the environment in order to gain additional insight into the underlying mechanisms of sensory–motor coordination, and more generally the nature of perception. A locomotion model of a four-legged robot is used for the case studies in both simulation and real world. The locomotion model demonstrates how attractor states derived from body dynamics influence the sensory information, which can then be used for the recognition of stable behavioral patterns and of physical properties in the environment. A comprehensive analysis of behavior and sensory information leads to a deeper understanding of the underlying mechanisms by which body dynamics can be exploited for category learning of autonomous robotic systems.     

Situated Computing: A Paradigm for the Mobile User-Interaction with Multimedia Sources

Today, users interact with computers in an explicit manner and the system's response is independent from their situations. Hence, it is difficult to integrate computers with working life as embedded tools, which can facilitate users to accomplish real world objectives easily. Situated computing is a new paradigm for mobile computer users based on their physical context and activities carried out as a part of their working business. It provides the mechanism to have a mobile computer as a utility to satisfy the user's real world requirements as well as an infrastructure for the situated interaction using applications. In this paper, we are presenting a metaphor called situation metaphor to model interaction between the user and mobile computers in order to achieve expectations of situated computing. A three-layered schema is followed in developing situation metaphor. We discuss extensively the theoretical foundation and framework for the situation metaphor, and followed by applications developed based on the framework.     

一种基于特征重要度的文本分类特征加权方法

针对文本分类中的特征加权问题,提出了一种基于特征重要度的特征加权方法.该方法基于实数粗糙集理论,通过定义特征重要度,将特征对分类的决策信息引入到特征权重中.然后,在标准文本数据集Reuters-21578 Top10和WebKB上进行了实验.结果表明,该方法能改善样本空间的分布状态,使同类样本更加紧凑,异类样本更加松散,从而简化从样本到类别的映射关系.最后,使用Nave Bayes,kNN和SVM分类器在上述数据集上对该方法进行了实验.结果表明,该方法能提高分类的准确率、召回率和F1值.     

Stakeholders, soft systems and technology: separation and mediation in the analysis of information system requirements

This paper describes the application of stakeholder analysis and soft systems thinking for an investigation of information system requirements. It is argued that it is appropriate to approach IS development as an exercise in managing complexity (soft systems) and pluralism (stakeholder analysis). A framework for investigating IS requirements is proposed that contrasts the current situation with the future situation and the real world with conceptual thinking about the real world. These aspects are viewed as outcomes of the process of IS requirements analysis, for which the metaphor of mediation is adopted in preference to presenting requirements analysis as a binary distinction between social construction and objectivity. The IS requirements analysis framework is applied in action research and the findings and learning that arose are presented.     

WEAKEST LINK,OR…

Metaphorizing is the principal means with which to conduct our thinking. We use metaphors to make meaning and sense about the world around us. We use them, furthermore, to form a view on the nature of the subject of the study. Thus, they play an important role in understanding a subject. Our point of view will determine what insights we will make and what inferences we draw. It is essential to understand the meaning of the metaphor and what mental pictures—including unintended ones—it conveys. Using an inappropriate mental picture can cause serious misconceptions. The focal point of this article is metaphors used to describe information security. The first one is a commonly used metaphor in the information security industry—the link metaphor. The meaning and the problems evident with using this metaphor are discussed first. An alternative is then provided—the Formula 1 (F1) car metaphor. A detailed rationale for using this metaphor instead of the first one completes the article.     

A note on web intelligence, world knowledge and fuzzy logic

Existing search engines––with Google at the top––have many remarkable capabilities; but what is not among them is deduction capability––the capability to synthesize an answer to a query from bodies of information which reside in various parts of the knowledge base.

In recent years, impressive progress has been made in enhancing performance of search engines through the use of methods based on bivalent logic and bivalent-logic-based probability theory. But can such methods be used to add nontrivial deduction capability to search engines, that is, to upgrade search engines to question-answering systems? A view which is articulated in this note is that the answer is “No.” The problem is rooted in the nature of world knowledge, the kind of knowledge that humans acquire through experience and education.

It is widely recognized that world knowledge plays an essential role in assessment of relevance, summarization, search and deduction. But a basic issue which is not addressed is that much of world knowledge is perception-based, e.g., “it is hard to find parking in Paris,” “most professors are not rich,” and “it is unlikely to rain in midsummer in San Francisco.” The problem is that (a) perception-based information is intrinsically fuzzy; and (b) bivalent logic is intrinsically unsuited to deal with fuzziness and partial truth.

To come to grips with fuzziness of world knowledge, new tools are needed. The principal new tool––a tool which is briefly described in this note––is Precisiated Natural Language (PNL). PNL is based on fuzzy logic and has the capability to deal with partiality of certainty, partiality of possibility and partiality of truth. These are the capabilities that are needed to be able to draw on world knowledge for assessment of relevance, and for summarization, search and deduction.     

Individual learning of coordination knowledge

Abstract. Social agents, both human and computational, inhabiting a world containing multiple active agents, need to coordinate their activities. This is because agents share resources, and without proper coordination or ‘rules of the road’, everybody will be interfering with the plans of others. As such, we need coordination schemes that allow agents to effectively achieve local goals without adversely affecting the problem-solving capabilities of other agents. Researchers in the field of Distributed Artificial Intelligence (DAI) have developed a variety of coordination schemes under different assumptions about agent capabilities and relationships. Whereas some of these researchers have been motivated by human cognitive biases, others have approached it as an engineering problem of designing the most effective coordination architecture or protocol. We evaluate individual and concurrent learning by multiple, autonomous agents as a means for acquiring coordination knowledge. We show that a uniform reinforcement learning algorithm suffices as a coordination mechanism in both cooperative and adversarial situations. Using a number of multi-agent learning scenarios with both tight and loose coupling between agents and with immediate as well as delayed feedback, we demonstrate that agents can consistently develop effective policies to coordinate their actions without explicit information sharing. We demonstrate the viabilityof using both the Q-learning algorithm and genetic algorithm based classifier systems with different pay-off schemes, namely the bucket brigade algorithm (BBA) and the profit sharing plan (PSP), for developing agent coordination on two different multi-agent domains. In addition, we show that a semi-random scheme for action selection is preferable to the more traditional fitness proportionate selection scheme used in classifier systems.     

基于机器学习的文本分类技术研究进展

文本自动分类是信息检索与数据挖掘领域的研究热点与核心技术,近年来得到了广泛的关注和快速的发展.提出了基于机器学习的文本分类技术所面临的互联网内容信息处理等复杂应用的挑战,从模型、算法和评测等方面对其研究进展进行综述评论.认为非线性、数据集偏斜、标注瓶颈、多层分类、算法的扩展性及Web页分类等问题是目前文本分类研究的关键问题,并讨论了这些问题可能采取的方法.最后对研究的方向进行了展望.     

基于Agent的文本分类系统

信息检索的一个核心问题是自动文本分类。基于分类体系的文本分类需要全文抽取主题词、计算权重,再根据分类体系对文献进行分类。文中构建一种基于Agent技术的文本自动分类系统,仅需要对文档头进行信息处理就可以进行快速文本分类,有效地减少了文本分类过程中的时间和空间的消耗。     

Coordination in information exchange between organizationaldecision units

The information technology (IT) infrastructure of today's organization is often one of distributed networked computing, with a flexible architecture to incorporate diverse software and hardware components to meet the needs of different user groups. One problem that may be encountered in this environment is the lack of information exchange across user groups or teams because of incompatible data definitions, diversity of formats, and differences in information processing capabilities and IT resource allocation across the organization. Successful information flows involve a user group providing information attribute levels (e.g., levels of precision or detail, timeliness, etc.), data definitions, formats and presentation styles required by the recipient group(s). When information exchange occurs without a fit between the attribute required and that provided, we have the problem of “information coordination.” It is thus critical to design and implement a set of organizational mechanisms which will encourage units within an organization to share relevant information in a coordinated manner. We develop a stylized model of coordination based information exchange between two decision units, and show that structured communication regarding the units' intended information sharing choices can partly mitigate the problem of coordination. When there is uniformity in information management capability and IT resources across the organization, the positive impact of such communication is reinforced, These results imply an emerging role of MIS, where the latter functions as a coordinator of communication and understanding between interdependent derision units, and a facilitator of equity in information processing capability across the units     

基于词典的名词性隐喻识别

隐喻是用一个事物来类比另外一个事物的语言表达,在自然语言中非常普遍,要实现自然语言理解隐喻处理不可避免。该文针对最基本的隐喻类型——名词性隐喻,提出基于词典的识别方法。结合同义词词林的语义距离与HowNet的语义关系来识别隐喻,考察隐喻与语义距离及语义关系之间的关联。     

Coordination in hospitals: organized or emergent process?

Hospitals have become more and more complex organizations that require increasing degrees of vertical, lateral and longitudinal coordination among their staff. As organizations, hospitals positively organize the coordination between the activities of its different agents, as well as the associated communication between them. The overall goal is to construct a “common ground” between the agents about the work process. In order to achieve this, a series of management tools are called upon. However these coordination mechanisms may fail. In the case analyzed, the agents seem to organize their behavior through direct and local individual interactions with their work environment, rather than through a global representation of the work. The case study demonstrates that intentionally organized coordination mechanisms interact with, and may be superseded by, the “emergence-through use” of coordination mechanisms in real time. These two mechanisms are clearly embedded at work, and can both be beneficial in promoting coordination in large scale systems.     

隐喻理解的计算模型综述

国际上,隐喻在思维及语言中所处的中心地位正逐渐引起人工智能研究者的重视。但在国内学术界,还鲜有开展隐喻计算化这方面研究的;实际上,作为异常用法的隐喻现象是自然语言中的普遍情况,因此隐喻问题若得不到很好的解决,将成为制约自然语言理解和机器翻译的瓶颈问题。本文结合相关的隐喻理论基础,根据不同的计算路线对已有隐喻理解计算模型进行分类,包括基于语义优先方法、基于知识表示的方法、基于逻辑的方法和基于统计语料库的方法,并在分析这些方法的适用范围和优缺点的基础上,对隐喻的计算理解方法以及面向汉语的隐喻理解计算模型研究提出了展望和建议。     

分布式协作模型及应用研究

With the development of the Web technology,the application environment has acquired many new characters such as dynamic, openness,distribution and information uncertainty.The processing mode of application systems is more complicated than ever.For example,it requires application systems to have more community processing ability,interactive ability,distributed processing ability and collaborative ability.Accordingly the research and development of the computer application system transited from client/server information processing system into distributed collaborative processing system based on Web.Especially in the environment where the information and resources are highly distributed,the accomplishment of complicated tasks is dependent more on the resources coordination,information sharing and coordinator collaboration.The collaboration is one aspect of the group behavior and its goal is to provide a optimal method to utilize the resource through the information interaction and to solve the task which couldn't be accomplished by each coordinator alone and get the more total benefits than the sum of each benefit.The collaboration problem is the important one for distributed tasks processing.This paper surveys,the research and application status of distributed collaborative models and several representative architectures of distributed collaborative processing are proposed.However,the existing problems and the future researching direction are presented.