This article is included in the column " Explaining Natural Language Processing in Simple Ways ". This column focuses on the major classic algorithms in the field of natural language processing and will continue to be updated. Welcome to subscribe!
Personal homepage: Program Xingkong with dreams
Personal introduction: The editor is a master in the field of artificial intelligence and a full-stack engineer. He has been deeply involved in Flask back-end development, data mining, NLP, Android development, automation and other fields. He has rich experience in research and development of software systems and artificial intelligence algorithm services.
If the article is helpful to you, welcome 
关注,
点赞,收藏,订阅。
-
Background of Knowledge Graph
On May 17, 2012, Google officially proposed the concept of Knowledge Graph. Its original intention was to optimize the results returned by search engines and improve users' search quality and search experience. In fact, the current artificial intelligence technology can be simply divided into perceptual intelligence (mainly image, video, voice, text, etc. recognition) and cognitive intelligence (involving knowledge reasoning, causal analysis, etc.), knowledge graph technology is the core of the field of cognitive intelligence. The main technology is a component of artificial intelligence technology. Its powerful semantic processing and interconnection organization capabilities provide the basis for intelligent information applications.
-
The definition and three elements of knowledge graph
Knowledge Graph ( KG for short ) is essentially a knowledge base called a semantic network , that is, a knowledge base with a directed graph structure. It consists of nodes (points), edges (edges) and properties (property) . In the knowledge graph, each node represents an 'entity' in the real world, and each edge represents the 'relationship' between entities.
The three elements of the knowledge graph include: entities, relationships, and attributes.
Entity : Also called Ontology, refers to things that exist objectively and can be distinguished from each other. They can be specific people, things, things, or abstract concepts or connections. Entities are the most basic elements in the knowledge map.
Relationship : In the knowledge graph, the edge represents the relationship in the knowledge graph, which is used to represent a certain connection between different entities.
Attributes : Both entities and relations in the knowledge graph can have their own attributes.
Figure 1 Three elements in the knowledge map
-
Classification of Knowledge Graphs
General Knowledge Graph : A "structured encyclopedic knowledge base" for general fields, focusing on building industry commonsense knowledge and using it in search engines and recommendation systems.
Field-specific knowledge graph (industry knowledge graph, vertical knowledge graph): for a specific field, it can be regarded as an "industry knowledge base based on semantic technology", mainly for enterprises, by building knowledge graphs of different industries and enterprises, to Provide knowledge-based services within the enterprise.
-
Technical process of knowledge map
Knowledge sources : Knowledge graph data can be obtained from a variety of sources (text, structured databases, multimedia data, sensor data, etc.).
Knowledge representation : Knowledge representation refers to the method and technology of using computer symbols to describe and represent the knowledge in the human brain to support the reasoning of the machine to simulate the human mind.
Knowledge extraction : Knowledge extraction can be divided into concept extraction, entity recognition, relationship extraction, event extraction and rule extraction according to tasks.
Knowledge fusion : When building a knowledge map, knowledge input can be obtained from third-party knowledge base products or existing structured data.
Knowledge graph completion and reasoning : Commonly used methods include the completion method based on ontology reasoning, link prediction based on representation and knowledge graph embedding, and methods based on graph structure and relationship path features.
Knowledge retrieval and knowledge analysis: The implementation forms of knowledge retrieval based on knowledge graph mainly include semantic retrieval and intelligent question answering. Knowledge graph and semantic technology are also used to assist in data analysis and decision-making.
-
Construction of Knowledge Graph
Figure 2 The construction process of the knowledge map
There are two main ways to build knowledge graphs, top-down and bottom-up.
Top-down : It is necessary to define the ontology (schema) and data schema first, then add the entity to the knowledge base, and complete the process from information extraction to graph construction based on the input data. This construction method needs to utilize some existing structured knowledge base as its basic knowledge base. It is suitable for the construction of professional knowledge graphs, such as enterprise knowledge graphs, which are used by professional users in the field.
Bottom-up : It refers to extracting entities with high confidence from some open-linked data and adding them to the knowledge base, and then constructing the top-level ontology model. It is more suitable for commonsense knowledge, such as the construction of general knowledge graphs such as names of people and organizations. Most knowledge graphs are constructed in a bottom-up manner, the most typical of which are Google's Knowledge Vault and Microsoft's Satori knowledge base. This is also in line with the characteristics of Internet data content knowledge generation.
-
Application of Knowledge Graph
The knowledge map provides a more effective way for expressing, organizing, managing, and utilizing massive, heterogeneous, and dynamic big data on the Internet, making the network more intelligent and closer to human cognitive thinking. The applications of knowledge graphs are mainly concentrated in the fields of search and recommendation . Such as semantic search, intelligent recommendation, knowledge storage, data verification, expert system, customer service robot , etc.
(1) Semantic search
The current keyword-based search technology can be upgraded to entity and relation-based retrieval with the support of KG's knowledge. It can accurately capture the user's search intent, solve the problems of keyword semantic diversity and semantic disambiguation, and directly give answers that meet the user's search intent, rather than links to relevant web pages containing keywords.
Figure 3 Application of knowledge map in search engine
(2) Question answering system
Question answering system is an advanced form of information retrieval system, which can provide users with answers to questions in accurate and concise natural language. Most question answering systems are more inclined to decompose a given question into multiple small questions, and then go to the knowledge base to extract matching answers one by one, and automatically detect their coincidence in time and space, etc., and finally merge the answers. presented to the user in an intuitive manner.
(3) Smart recommendation
In addition to optimizing search results, knowledge graphs can also help e-commerce and social platforms solve some intelligent recommendation problems. For example, the biggest problem for some small and medium-sized platforms in terms of smart recommendation is "what to buy and what to recommend" or "recommended products are not related to customers". The recommended products lack novelty, resulting in mediocre conversion effects.
Knowledge graphs can help e-commerce platforms jump out of this simple recommendation logic, making recommendation results more intelligent and promoting user purchases.
Follow the WeChat public account [Program Starry Sky with Dreams] to learn about cutting-edge knowledge in the field of software systems and artificial intelligence algorithms. Let us learn and make progress together!