基于本体的BML命令语义标注方法

作战管理语言（Battle Management Language,BML）是一种无歧义的语言规范,用于解决指控和仿真系统间的互操作问题。针对BML命令缺乏语义信息,难以被计算机理解和自动处理的问题,分析了BML的五视图体系结构,引入了BNF形式化语法,并在此基础上提出了一种语义标注方法,对命令中的对象、时间和地理位置等要素进行标注,然后整合各要素来表达整条命令的语义,系统通过解析标注文档可以实现对命令的无歧义理解,最后给出实例予以说明。论文为表达BML命令的语义提出了一种借鉴思路。     

Automatic annotation of tennis games: An integration of audio,vision, and learning

Fully automatic annotation of tennis game using broadcast video is a task with a great potential but with enormous challenges. In this paper we describe our approach to this task, which integrates computer vision, machine listening, and machine learning. At the low level processing, we improve upon our previously proposed state-of-the-art tennis ball tracking algorithm and employ audio signal processing techniques to detect key events and construct features for classifying the events. At high level analysis, we model event classification as a sequence labelling problem, and investigate four machine learning techniques using simulated event sequences. Finally, we evaluate our proposed approach on three real world tennis games, and discuss the interplay between audio, vision and learning. To the best of our knowledge, our system is the only one that can annotate tennis game at such a detailed level.     

基于条件随机场的中医临床病历命名实体抽取

中医临床病历是中医重要的科研数据资源,但目前临床病历仍以文本为主要表达形式,对病历数据深入分析的前提是进行结构化信息抽取,而命名实体抽取是其基础性步骤。针对中医临床病历的命名实体,如症状、疾病和诱因等的抽取问题,通过手工标注的413份病历数据(以中文字为特征)与4类特征模版,将条件随机场(CRF)、隐马尔科夫模型(HMM)和最大熵马尔科夫模型(MEMM)用于中医病历命名实体抽取的实验,并进行比较分析。结果表明,结合合适的特征模版,CRF命名实体抽取方法取得了较好的性能,F1值的症状达到0.80,疾病名称达到0.74,诱因0.74。与HMM和MEMM相比,CRF有最高的准确率和召回率,是一种较为适用的中医临床病历命名实体抽取方法。     

基于话题链的汉语语篇连贯性描述体系

汉语简洁灵活的意合型篇章组合结构,对传统的基于关联词的篇章连贯性描述体系提出了新的挑战。该文引入话题链描述形式,设计不同类型的话题评述关系集,构建了以话题链为主,融合关联词语和其他连贯形式描述机制,覆盖话题评述、并列、因果、转折四大类关系的汉语语篇连贯性描述体系。在清华句法树库TCT上进行的验证实验,发现话题链和关联词语分别覆盖了约76%和50%的汉语复句,并且两者经常同时使用,初步证明了这个体系在句子连贯性描述方面的可行性和有效性。     

汉语中介语的依存句法标注规范及标注实践

汉语中介语是伴随着汉语国际教育产生的,随着汉语学习在全球的不断开展,汉语中介语的规模不断增长,由于这些语料在语言使用上有其独特性,使得中介语成为语言信息处理和智能语言辅助学习的独特资源。依存语法分析是语言信息处理的重要步骤,英语中介语的依存语法标注语料已经有很好的应用,目前汉语中介语语料库对句法的关注度较低,缺乏一个充分考虑汉语中介语特点的依存句法标注规范。该文着眼于汉语中介语的依存句法标注语料库的建构,探讨依存标注规范,在充分借鉴国际通用依存标注体系(Universal Dependencies)的基础上,制定了汉语中介语的依存标注规范,并进行了标注实践,形成了一个包括汉语教学语法点的中介语依存语料库。     

Wonderful events automatic detection and location of the HCRF and webcast text

Based on the powerful function of the hidden conditional random fields (HCRF) model in the expression and identification of semantic events and combining the webcast text information, a new framework for wonderful events automatic detection is put forward. Firstly, by analyzing and processing the webcast text, keyword combinations corresponding to each exciting event are obtained. Secondly, by classifying the webcast text to be detected, the happening time labels of each wonderful event are obtained. Thirdly, an HCRF model for semantic shot annotation is built to realize the semantic annotation of multiple types of semantic shots simultaneously, and the semantic shot sequence of the video clip is obtained. Finally, combining the multi-modal semantic clues, an HCRF model for the wonderful events detection and localization is effectively built in the case of small-scale training samples. Based on the mapping relationship among video low-level features, the multi-modal semantic clues and the wonderful semantic events, the inherent patterns of the wonderful events are excavated deeply in the multiple dimensions of the semantic structure, and then the wonderful events automatic detection, localization and segmentation are precisely achieved. Experiments show the effectiveness of this model.     

三维标注的产品数字化定义标准及其管理系统开发

三维标注新技术引起了数字化定义方法的变革。简要介绍三维标注技术的内涵,详细说明了三维标注信息的分类和表达方法,构建了产品数字化定义的标准体系。为减轻产品数字化定义的工作量,开发了基于CATIA V5的产品定义标准信息管理系统,实现了产品数字化定义标准信息的统一管理,并详细阐述了基于CATIA的产品数字化定义应用流程。     

Impact analysis and change propagation in service-oriented enterprises: A systematic review

ContextThe adoption of Service-oriented Architecture (SOA) and Business Process Management (BPM) is fairly recent. The major concern is now shifting towards the maintenance and evolution of service-based business information systems. Moreover, these systems are highly dynamic and frequent changes are anticipated across multiple levels of abstraction. Impact analysis and change propagation are identified as potential research areas in this regard.ObjectiveThe aim of this study is to systematically review extant research on impact analysis and propagation in the BPM and SOA domains. Identifying, categorizing and synthesizing relevant solutions are the main study objectives.MethodThrough careful review and screening, we identified 60 studies relevant to 4 research questions. Two classification schemes served to comprehend and analyze the anatomy of existing solutions. BPM is considered at the business level for business operations and processes, while SOA is considered at the service level as deployment architecture. We focused on both horizontal and vertical impacts of changes across multiple abstraction layers.ResultsImpact analysis solutions were mainly divided into dependency analysis, traceability analysis and history mining. Dependency analysis is the most frequently adopted technique followed by traceability analysis. Further categorization of dependency analysis indicates that graph-based techniques are extensively used, followed by formal dependency modeling. While considering hierarchical coverage, inter-process and inter-service change analyses have received considerable attention from the research community, whereas bottom-up analysis has been the most neglected research area. The majority of change propagation solutions are top-down and semi-automated.ConclusionsThis study concludes with new insight suggestions for future research. Although, the evolution of service-based systems is becoming of grave concern, existing solutions in this field are less mature. Studies on hierarchical change impact are scarce. Complex relationships of services with business processes and semantic dependencies are poorly understood and require more attention from the research community.     

基于Hessian半监督特征选择的网络图像标注

针对半监督特征选择算法进行了研究,采用有标签图像和无标签图像的半监督特征选择方法来提升网络图像标注的性能。基于二阶Hessian能提出一个新的半监督特征选择方法,该方法具有更好的局部拓扑结构保持特性和推断能力,从而能够克服基于图拉普拉斯半监督学习方法的缺点。将所提出的半监督特征选择算法应用到网络图像标注任务中,在两个大规模网络图像数据库上进行了实验,结果表明Hessian半监督特征选择方法优于拉普拉斯半监督特征选择方法,适合大规模网络图像标注。     

汉语复合名词语义信息标注词库: 基于生成词库理论

复合名词分析一直是语言研究和自然语言信息处理中的一个重要问题,涉及未登录词的识别、自动释义以及词典编纂等。生成词库理论是目前较新的一种语义学理论,我们拟利用该理论对汉语复合名词进行语义信息标注,建立一个词库。该文首先介绍了这一词库的标注框架,然后通过对“纸”、“石”构成的复合词的对比分析,展示了这一词库在复合名词构词、语义研究中的应用。研究结果显示,物性角色、自然类和人造类是非常重要的语义信息,能揭示复合名词构词和语义的一些模式和规律。