Knowledge mapping Overview （ Two ）

The paper 《Knowledge Graphs》 Link to the original text

This part is 《Knowledge Graphs》 The introduction of （ Page number 1） part , It mainly includes six points ： Related work 、KG The concept of 、KG Application 、 The examples used in this article （ Tourism in Chile ）、KG Review literature comparison 、 The structure of this paper .

Every little point below is a paragraph in the paper .

1. Related work

   • 1972 year , Earliest use KG The word" ; but 2012 year Google Put forward modern KG The concept of . Then other big companies （Airbnb、 Amazon 、eBay、 facebook 、IBM、 Led the 、 Microsoft 、 Uber, etc ） And start using KG. Again , Academia has also begun to publish on KG Books 、 The paper 、 summary 、 Definition 、 New technology 、 Overview, etc .
   • With “ Use diagrams to represent knowledge ” The proposal and development of this core concept ,KG be used for Express knowledge clearly , It comes from a large variety of data sources Integrate 、 management and Extract valuable information , To serve various application scenarios . And SQL and NoSQL comparison , Use Graph based knowledge abstraction There's a lot of benefits ：
     • For a / Multiple areas provide concise 、 intuitive The abstraction of ;
     • edge Express （ Social data 、 Biological interaction 、 Bibliographic references 、 co-author 、 Transportation network and so on ） The relationship between entities ;
     • The diagram allows Postpone the definition of a pattern , You can capture incomplete knowledge first , more flexible ;
     • KG Supporting the standard Relational operator （ coupling joins、 relation unions, Projection projections etc. ）, Also supports the use of Navigation operators To recursively find Any length path Connected entities ;
     • Standard knowledge representation —— for example noumenon and The rules , It can be used for Definition and reasoning The semantics of terms in the figure , these The term It's for Mark and describe nodes and edges in the diagram ;
     • An extensible framework for atlas analysis It can be used to calculate Centrality 、 clustering 、 Figure summary etc. , To gain insight into the KG The areas described ;
     • Various Represent Technology The development of , So that it can be applied directly on the graph machine learning technology .
   • Build and use KG The trend of has promoted a series of technology The development of , That is, from various data sources Integration and extraction Valuable information . However , Not yet Universal and unified KG Concept （ That is the problem to be solved in this paper 、 The goal of this paper ） Summarize to describe How to use KG、 Use what technology 、 How to relate to existing data management topics .
   • The target population of this paper is Not familiar with KG Researchers and practitioners （ namely KG The small white ）, The goal of this paper Is a comprehensive introduction to KG, It includes the following aspects ：
     • Describe some KG Of Basic data model 、 How to... Them Inquire about ;
     • KG It means ： Pattern 、 identifier and Context ;
     • Ways to make knowledge clearer ： deductive and inductive ;
     • establish （ structure ） And rich （ Enrich 、 expand ） Various techniques of graph structure data ;
     • understand KG Of Quality assessment , Based on how to KG Conduct improvement ;
     • published KG Of standard and Best practices ;
     • at present existing KG Application Overview .

2. KG The concept of

   • No matter from technical Define or More general The definition of ,KG Is defined as Controversial , Even many definitions are contradictory , This article is in Appendix A These previous definitions are discussed in .
   • This paper adopts an inclusive More general KG Definition , namely KG Aimed at Knowledge that accumulates and conveys real-world knowledge , its node representative Entities of interest , its edge Then represent The relationship between these entities .
   • Data graph It belongs to Graph based data model , It could be a multi graph 、 Attribute map and so on （ The first 2 This chapter will discuss other graphs ）.
   • knowledge yes Known things , It can be from External data sources Accumulated in , Or extracted from KG In itself . Knowledge can be made up of Simple statements （ Such as “ Beijing is the capital of China ”） perhaps Quantitative statement （ Such as “ All capitals are cities ”） form , The former can be directly summarized as... In the data graph edge , The latter needs a more expressive way to express —— Ontology or rule .
   • Deductive method It can be further Derive more knowledge （ Such as “ Beijing is a city ”）.
   • Induction method It can be downloaded from KG in Extract and summarize Other additional knowledge , This extra knowledge is based on Simple declarative and quantitative sentences .
   • KG The data source is diversification Of , So the data itself Structure and granularity Aspects will be more diversified . In order to solve the problem of diversity , Pattern 、 Representation of identifiers and contexts Played a key role ：
     • Pattern Defined KG High level structure of ;
     • identifier Diagram （ Or an external data source ） Medium Which nodes It means the real world The same entity ;
     • Context It may indicate that A background setting stay In some pieces of knowledge Always objective and real .
   • All in all , Want to make a KG Over time, it develops and improves , There must be a reliable way to carry out knowledge extract 、 Enrich 、 Quality assessment and Refine .

3. KG Application

   • In an organization or community ,KG It's a Evolving Share the basic knowledge base .
   • At the application level , There are mainly Open source KG and Enterprises KG Two kinds of . Open source KG Publish... On the Internet , Everyone can use , Like the most famous DBpedia、Freebase、Wikidata、YAGO etc. , They cover a variety of areas , They are all from Wikipedia Extracted from , perhaps Constructed by a volunteer community . Besides , Open source KG It will be Publish in a specific field , For example, media 、 The government 、 Geography 、 tourism 、 Life science and so on . Enterprises KG stay Internal use And for Business use cases . Use enterprise KG Of Major industries Include Web Search for （ For example, Bing 、 Google ）、 business （ for example Airbnb、 Amazon 、eBay 、 Uber ）、 Social networks （ for example Facebook、LinkedIn）、 Finance （ For example, Accenture 、 Bank of Italy 、 Bloomberg 、 First capital 、 Wells Fargo Bank ） wait . Specific application software Including search engines 、 Recommendation system 、 Personal agent 、 advertisement 、 Business analysis 、 risk assessment 、 Automation and so on . This paper will discuss the specific application details in Chapter 10 .

4. The examples used in this article （ Tourism in Chile ）

   • For discussion , This paper is based on Chilean Tourism dependent KG Provide specific examples in a hypothetical context . The KG Managed by a Tourism Bureau , The aim is Promote the development of tourism in the country , as well as Explore new attractions in the planning area . The KG Itself will describe Tourist attractions 、 Cultural Activity 、 service and business , as well as City and Intercity travel routes . The internal plan of the organization applies it to ：
     • In many languages establish One Portal site , So that tourists can Search for scenic spots 、 lately Activities to be held And other related services ;
     • In season 、 In depth analysis of nationality and other dimensions Tourist demographics ;
     • analysis Tourists' interest in attractions and activities Love （ Sentiment analysis ）, Including positive comments 、 Summary of complaints about activities and services 、 Crime reports, etc ;
     • analyse Travel trajectory ： Tourists visit scenic spots 、 Sequence of activities ;
     • Cross reference the flight and bus tracks to Design a new journey planning route ;
     • Personalized attractions recommend .

5. KG Review literature comparison

   • In Table 1 present About KG Three types of relevant literature —— review 、 Books and handouts , The topics they cover are compared with this paper Compare . It can be seen from the figure that some literatures tend to pay attention to KG A particular aspect of , Different from these documents , This paper is to provide a reference to KG Broad and easy to understand Introduce the tutorial . The research results of some of these papers provide more in-depth technical details than this paper ; In the whole process of this paper , We will refer to these surveys for further reading as appropriate .

• Symbol description ：
  • Circle and √： In depth discussion ;
  • circle ： Express a brief discussion ;
  • *： It means that arXiv Informal release on .

6. The structure of this paper

   • The first 2 This chapter outlines Figure data model And can be used for Inquire about Their Language .
   • The first 3 Chapter describes KG in Pattern 、 Identifier and context It means .
   • The first 4 This chapter introduces the knowledge Express and contains Deductive formal expression of .
   • The first 5 Chapter describes Extract additional knowledge Of Inductive technology .
   • The first 6 Chapter discusses the use of External data sources structure and Enrich Knowledge map .
   • The first 7 This chapter lists the possible assessment KG Of quality dimension .
   • The first 8 Chapter discusses the use of KG Refine All kinds of technology .
   • The first 9 Chapter discussion Release KG Principles and agreements .
   • The first 10 This chapter investigates Some famous KG And Its Applications .
   • The first 11 Chapter is Yes KG Summary of And its Future research direction .
   • appendix A Provides KG Of Historical background and In the past Yes KG Definition .
   • appendix B Lists the... Cited in the main body of the paper Formal definition .