Construction and application of industrial knowledge atlas (I): overview of industrial knowledge atlas

This chapter first introduces the overview and classification of knowledge maps in industrial scenarios , Then it briefly introduces some basic principles and methods in the practice of knowledge map , Finally, the construction is introduced 、 Systems needed to maintain large-scale knowledge maps , And the corresponding architecture design of the system .

1.1 An overview of Knowledge Mapping

In recent years , Thanks to the development of big data and the significant improvement of cloud service computing power , Based on the information of the whole network 、 Massive knowledge of structured objective world becomes possible ; Besides , Large scale structured knowledge construction , It can greatly improve the efficiency of Internet access to information . Specific to specific industries , Digital industry knowledge helps to automate business decisions , At the same time, it can make business decisions globally optimized . It is under the double wheel drive of technological progress and industrial application , Knowledge map is in full swing in industry . The present , From the perspective of the whole industry , Knowledge maps can be roughly divided into general knowledge maps and industry knowledge maps . Building a knowledge map 、 Using knowledge map is a systematic project , It's about natural language processing 、 The semantic web 、 Logical reasoning and other academic frontier scientific research topics . what's more , Whether it is general encyclopedic knowledge , Or in-depth knowledge of the industry , Are massive knowledge systems . The huge scale of knowledge makes the construction of knowledge atlas involve graph database 、 Massive knowledge management tools 、 Construction of large-scale engineering systems such as big data production system . Last , Knowledge often comes from big data mining 、 Integration with existing knowledge base and input of multi role experts , therefore , The construction of knowledge map requires a lot of manual participation , It's very complicated .

1.1.1 General knowledge map

The earliest industrial application of knowledge map is search engine . Mainstream search engines have built a large-scale general knowledge map , To support interactive search . in addition , Smart speakers such as tmall Genie provide interactive Q & A based on voice and natural language , Rely more on the precise answers provided by the knowledge map . These knowledge maps usually include common sense of the real world , Users can use this to query people 、 place 、 Description and connection of things and organizations .

Entities in the general knowledge map 、 The types of relationships and attributes are defined in the ontology . Contains about billions of major entities , Tens of billions of triples know