融合增强协同信息和知识图谱信息的推荐模型

将知识图谱引入推荐系统,能一定程度解决数据稀疏和冷启动问题,但是往往忽略了高阶协同信息和不同协同信息的重要程度对探索用户潜在偏好的重要性,由此提出了一种融合增强协同信息和知识图谱信息的推荐模型(CIKG)。该模型首先利用用户和项目的历史交互数据,获取一阶协同信息和高阶协同信息,同时使用注意力机制捕获重要信息,得到增强协同信息,用来补充用户和项目的特征表示。其次通过将用户交互的项目与知识图谱中的实体对应,在知识图谱中执行传播操作,得到知识图谱信息,用于挖掘用户的偏好并且增强模型的可解释性。最后通过聚合器将增强协同信息和知识图谱信息结合得到用户和项目的最终表示,从而进行预测。在Last-fm和Book-crossing两个数据集上进行的实验结果表明CIKG相比其他对比的模型推荐效果有较大提升。     

Knowledge types: a practical approach to guide knowledge engineering in domains of building design

Abstract: This paper presents experiences and findings achieved through several expert system projects in domains of technical buiding design. The main focus is on knowledge acquisition and knowledge decomposition. Different approaches are presented and discussed. Four types of knowledge are presented: object knowledge, performance knowledge, event knowledge and metaknowledge, and the applicability of this classification to building design is justified. The paper includes a listing of experiences gained by real world testing of an expert system, and finally we outline guidelines to knowledge engineering in the domain of technical building     

Active learning through collaborative knowledge building using an automatic free-text scoring system in a b-learning environment

According to active learning, students should be responsible for their own learning. Automatic free-text scoring allows teachers to provide open-ended questions with their correct answers to a computer system, so when students answer the questions, they get immediate feedback (a score, a comment, or both). However, teachers are usually overloaded with many tasks, and they may not have time to create the questions with the correct answers. Therefore, in the 2012/2013 academic year, we asked a group of 124 Pre-Primary and Primary Education students to become the creators of the questions and their correct answers in groups in a free-text scoring system, so the questions use learners’ language, not teachers’ language. From them, 41 students (group of involved students, GIS) fulfilled all the requirements during the course. Our hypothesis was that GIS would be able to increase their academic performance and levels of engagement compared to the rest of the students. The results gathered provide statistic evidence to support that hypothesis. This study pretends to help teachers who want to increase the academic performance and levels of engagement of their students in courses that they may find boring and unrelated to the main topic of their degree, or not directly related with their main academic interests.     

工艺设计知识库的建造与维护

文中根据工艺设计知识的特点,构造了层次化的知识表达、组织与知识库模型,分析了工艺知识库不一致的表现形式,并给出了相应的一致性验证算法,提出了基于广义决策表的知识库完备性检查方法。     

Collaborative knowledge building using the Design Principles Database

In this study we describe a mechanism for supporting a community of learning scientists who are exploring educational technologies by helping them to share and collaboratively build design knowledge. The Design Principles Database (DPD) is intended to be built and used by this community to provide an infrastructure for participants to publish, connect, discuss and review design ideas, and to use these ideas to create new designs. The potential of the DPD to serve as a collaborative knowledge-building endeavor is illustrated by analysis of a CSCL study focused on peer-evaluation. The analysis demonstrates how the DPD was used by the researchers of the peer-evaluation study in three phases. In the first phase, design principles were articulated based on a literature review and contributed to the DPD. In the second phase, a peer-evaluation activity was designed based on these principles, and was enacted and revised in a three-iteration study. In the third phase, lessons learned through these iterations were fed back to the DPD. The analysis indicates that such processes can contribute to collaborative development of design knowledge in a community of the learning sciences. Readers of ijCSCL are invited to take part in this endeavor and share their design knowledge with the community.     

基于对抗学习和全局知识信息的关系检测技术研究

针对现有的知识库关系检测任务对于一些不可见关系无法做到准确的向量表示而出现词汇溢出的问题,提出了基于对抗学习和全局知识信息的关系检测模型。该模型使用对抗学习对知识库关系表示模型进行特征强化,使用TransH(translating on hyperplanes)模型提取全局知识信息,同时通过联合训练,将全局知识信息融合进关系表示模型中,进一步提升关系模型的表示能力。实验结果表明,提出的融合模型对于关系检测效果有一定的提升,并且缓解了词汇溢出的问题。     

Coping knowledge boundaries between information system and business disciplines: An intellectual capital perspective

Information system development can be considered a collaboration between users and developers. The inability to leverage the localized knowledge embedded in these two stakeholders hinders software development work to achieve high performance. Exploring the ways to counter this difficulty is then critical. This study applies an intellectual capital perspective to address the issues around spanning the knowledge boundary between developers and users. Our findings highlighted how important effective knowledge boundary spanning is to both product and project quality. Furthermore, three dimensions of intellectual capital increased the degree to which knowledge boundary spanning was effective.     

知识粗糙性和条件信息熵的关系

目前,粗糙集理论存在着两种观点,它们分别是代数观和信息观。在代数观点中,知识粗糙性体现了知识的粒度;而在信息观中,定义了知识的信息熵和条件信息熵。已经有定理证明了信息熵与知识的粗糙性存在对应关系,它建立了代数观和信息观之间的联系,但是这种关系却不是一一对应的。该文通过重新证明知识粗糙性和信息熵的对应关系定理,找到与知识粗糙性存在一一对应关系的是条件信息熵,并给出相关定理及其证明。     

Applying domain knowledge and social information to product analysis and recommendations: an agent-based decision support system

Abstract: The advance of Internet and Web technologies has boosted the development of electronic commerce. More and more people have changed their traditional trading behaviors and started to conduct Internet shopping. However, the exponentially increasing product information provided by Internet enterprises causes the problem of information overload, and this inevitably reduces the customer's satisfaction and loyalty. To overcome this problem, in this paper we propose a multi‐agent system that is capable of eliciting expert knowledge and of recommending optimal products for individual consumers. The recommendations are based on both product knowledge from domain experts and the customer's preferences from system–consumer interactions. In addition, the system also uses behavior patterns collected from previous consumers to predict what the current consumer may expect. Experiments have been conducted and the results show that our system can give sensible recommendations, and it is able to adapt to the most up‐to‐date preferences for the customers.     

A case study analysis of a constructionist knowledge building community with activity theory

This article investigates how activity theory can help research a constructionist community. We present a constructionist activity model called CONstructionism Through ACtivity Theory (CONTACT) model and explain how it can be used to analyse the constructionist activity in knowledge building communities. We then illustrate the model through its application to analysing the Wiki-supported community associated with a computer game. Our analysis focuses mainly on two perspectives: individual and collective actions, as well as individual and collective mediations. Experiences and challenges from the analysis are reported to demonstrate how CONTACT is helpful in analysing such communities.     

Scaffolding wiki‐supported collaborative learning for small‐group projects and whole‐class collaborative knowledge building

While educators value wikis' potential, wikis may fail to support collaborative constructive learning without careful scaffolding. This article proposes literature‐based instructional methods, revised based on two expert instructors' input, presents the collected empirical evidence on the effects of these methods and proposes directions for future refinements. The instructional methods were implemented by an expert instructor teaching a 12‐week 68‐student undergraduate design class in Canada. Data were collected from observations, interviews and content analysis of wikis. The findings revealed that in small‐group project (SGP), the wiki instructional methods enhanced collaborative learning with most instructional methods derived from cooperative learning, but in whole‐class collaborative knowledge building (CKB), the wiki instructional mehtods failed to turn the class into a self‐sustained learning community after the scaffolding faded. We conclude that the genre of wikis should be different for SGP and CKB. While the students easily adopted the ‘reproduced’ genre of wikis for SGP with familiar tasks, they felt overwhelmed or resistant to the unfamiliar ‘emergent’ genre of wikis for CKB in massive collaborative constructive learning. Therefore, we propose that future refinements for wiki‐supported CKB should focus on providing students scaffolding for intersubjectivity (understanding collaborative constructive learning) and transfer of responsibility (developing autonomy).     

航空制造知识图谱构建研究综述

目前航空装备制造企业的设计、制造相关流程中积累了大量数据,基于知识图谱技术可以对这些数据进行有效融合与管理,对不断更新的制造知识进行挖掘,将为航空制造企业智慧化升级提供有力的知识支撑。为探明知识图谱在航空制造领域的理论支撑体系与实际应用情况,通过文献调研分析航空制造知识图谱架构、定义及特点;阐明知识图谱领域构建过程中的核心技术并进行研究综述,对比航空制造知识图谱与通用知识图谱构建技术上的异同,并提出了三个切合实际的航空制造知识图谱应用方向及其解决方案;最后对未来航空制造知识图谱的挑战进行了分析及展望,为后续该领域的研究提供一些思路。     

建设法规政策知识库中后控词表检索设计

编制后控词表是实现智能查询的重要环节.针对建立"智能化建设法规政策知识库",重点研究通过后控词表实现智能化查询的方法.针对自然语言查询涉及政策法规和案例两类查询结果,从而设计了包含有两个主题词表,共6个词表的后控词表模式,分析词表关联关系,说明词表的词汇集合关系,并给出了基于此模式的智能查询算法,对大型文献智能知识库系统开发中后控词表的编制具有一定借鉴意义.     

Retrieving information on the World Wide Web: effects of domain specific knowledge

In this study, we intend to examine information retrieval behaviors from a psychological point of view using a search engine on the World Wide Web (WWW). We investigated information retrieving behaviors in detail based on both the recorded data of retrievers’ web browsing actions and their thinking processes by the “think aloud” method. We focused on selected keywords for retrieving and compared them between retrievers who had enough knowledge about their task and those who did not. Our goal was to learn about the literacy needed for finding required information efficiently on the WWW.

Cooperation between expert knowledge and data mining discovered knowledge: Lessons learned

Expert systems are built from knowledge traditionally elicited from the human expert. It is precisely knowledge elicitation from the expert that is the bottleneck in expert system construction. On the other hand, a data mining system, which automatically extracts knowledge, needs expert guidance on the successive decisions to be made in each of the system phases. In this context, expert knowledge and data mining discovered knowledge can cooperate, maximizing their individual capabilities: data mining discovered knowledge can be used as a complementary source of knowledge for the expert system, whereas expert knowledge can be used to guide the data mining process. This article summarizes different examples of systems where there is cooperation between expert knowledge and data mining discovered knowledge and reports our experience of such cooperation gathered from a medical diagnosis project called Intelligent Interpretation of Isokinetics Data, which we developed. From that experience, a series of lessons were learned throughout project development. Some of these lessons are generally applicable and others pertain exclusively to certain project types.     

知识表示学习方法研究综述

近年来,知识表示学习已经成为知识图谱领域研究的热点。为了及时掌握当前知识表示学习方法的研究现状,通过归纳与整理,将具有代表性的知识表示方法进行了介绍和归类,主要分为传统的知识表示模型、改进的知识表示模型、其他的知识表示模型。对每一种方法解决的问题、算法思想、应用场景、评价指标、优缺点进行了详细归纳与分析。通过研究发现,当前知识表示学习主要面临关系路径建模、准确率、复杂关系处理的挑战。针对这些挑战,展望了采用关系的语义组成来表示路径、采用实体对齐评测指标、在实体空间和关系空间建模,以及利用文本上下文信息以扩展KG的语义结构的解决方案。     

Community knowledge assessment in a knowledge building environment

This study investigates ways of using key terms to represent and assess community knowledge in an online knowledge building environment. Knowledge Forum – an environment specially designed to support advances in community knowledge – incorporates key-term analytic tools. In the current study these tools were used to determine if key-term measures complement conventional online behavioral measures in assessing community knowledge advances. Discourse rated as more reflective and depth-oriented showed higher percentages of shared key terms and higher frequency use of shared key terms than less reflective, shallower discourse. Limitations and possibilities for using key terms for automated assessment and visual representation of community knowledge are discussed.     

Fostering collective and individual learning through knowledge building

The purpose of this study was to design and examine a computer-supported knowledge-building environment and to investigate both collective knowledge-building dynamics and individual learning in the context of a tertiary education course in mainland China. The participants were 102 students in four intact Year-one tertiary business classes. Two classes experienced a knowledge-building environment (CKB) and the other two were taught using a regular project-based approach (RPBL). Data were obtained from interactions in the forum, writing quality, group-learning portfolios, and surveys. Quantitative analyses indicated that the knowledge-building groups outperformed the comparison groups on academic literacy assessed in terms of conceptual understanding and explanation, and obtained higher scores on beliefs about collaboration. Within-group analyses indicated that the students’ engagement in Knowledge Forum was a significant predictor of their academic literacy. Qualitative contrastive analyses of high- and low-performance groups identified different patterns of conceptual, metacognitive and social processes, and showed that student groups engaging in more collective and meta-discourse discourse moves performed better on individual scores in academic literacy. The implications of examining both collaborative dynamics and individual learning and designing computer-supported knowledge building for tertiary students are discussed.     

Fast searching for information on the Internet to use in a learning context: The impact of domain knowledge

The purpose of the study was to examine the role of domain knowledge when retrieving and using information from the Internet as a resource for essay tasks, as well as to investigate the quality of Internet searches and its relation to essay performance. In two experiments, 100 undergraduates searched the Internet for 30 min and completed two essays; one in which they had high domain knowledge and one in which domain knowledge was low. Two control groups of 70 undergraduates just wrote the essays. Searching the Internet for information enhanced essay performance relative to the control groups only for the topic for which participants had high domain knowledge. In the second experiment, analyses of Internet searches revealed large individual differences in search behaviors and these behaviors did not relate to essay performance, although individuals highlighted the importance of domain knowledge in making their searches easier. Domain knowledge is one factor that educators should pay attention to when using the Internet for learning tasks, particularly when study time is limited, in order to maximize the ability of students to successfully retrieve and use information from the Internet.