1. A smart coaching method, characterized in that the method comprises:

an active coaching step: based on known data input by a user, after entity identification and relation identification are carried out in a knowledge graph, active coach type enlightening is carried out on the user by adopting an enlightening coach technology, a plurality of open type questions are sent, the user answers based on the open type questions, and a final answer is found based on an open type question sorting thought;

a bidirectional training step: after the user inputs known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions, and finds the final answer.

2. The smart coaching method of claim 1, wherein the method further comprises: establishing a knowledge graph: and establishing a knowledge graph based on the historical knowledge base.

3. The smart coaching method of claim 2, wherein the method further comprises:

a passive coaching step: and the coach system carries out entity recognition and relationship recognition based on the knowledge graph, finds a final answer and answers the user.

4. The smart coaching method of claim 2, wherein the knowledge-graph establishing step comprises:

self-learning step: and continuously and autonomously learning from the historical knowledge base to enrich knowledge map information.

5. A smart coaching arrangement, using the smart coaching method of any one of claims 1-4, wherein the arrangement comprises:

an active coaching unit: the knowledge graph module is used for carrying out entity recognition and relationship recognition in the knowledge graph module based on known data input by a user, then adopting a heuristic coach technology to carry out active coach-type heuristic on the user, sending out a plurality of open questions, answering the questions based on the open questions by the user, and finding out a final answer based on an open question sorting thought;

a bidirectional trainer unit: after the user inputs known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions, and finds the final answer.

6. The smart coaching device of claim 5, wherein the device further comprises: a knowledge graph establishing unit: the method is used for establishing the knowledge graph based on the historical knowledge base.

7. The smart coaching device of claim 6, wherein the device further comprises:

a passive coaching unit: and the coach system carries out entity recognition and relationship recognition based on the knowledge graph, finds a final answer and answers the user.

8. The smart trainer device of claim 6, wherein the knowledge graph establishing unit comprises:

a self-learning unit: the knowledge graph learning method is used for continuously and autonomously learning from a historical knowledge base and enriching knowledge graph information.

9. An electronic device, comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus, and wherein the memory is configured to store a computer program; the processor for performing the method steps of any one of claims 1 to 4 by running the computer program stored on the memory.

10. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to carry out the steps of the smart coaching method of any one of claims 1 to 4 when executed.

Background

At present, increasingly, expert systems or question-answer self-service systems are all used for asking questions by users, and background intelligent systems search corresponding answers or organize the answers by themselves according to the questions of the users and then return the answers to the users. Most of these question-answering systems are communicated in text. If voice is available for communication, most of the situations are also entertainment and simple education.

Or when the question-answering service equipment of various query services receives an externally input query statement, the query statement is input into the question-answering system for query, and a reply statement obtained after the query by the question-answering service equipment is obtained, wherein the reply statement contains information to be queried by the query statement. However, in practical applications, if passive query of a query statement is implemented only by relying on a question-answering system, the user can only obtain a reply statement of the query statement, but divergent thinking cannot be performed with the assistance of a coach of the system, and the user cannot comprehensively think and track the final result from the known information.

Therefore, in order to solve the above-mentioned drawbacks of the prior art, the present invention provides an intelligent coaching device, an electronic device and a storage medium, and a universal intelligent coaching method is added to solve the technical problems that the answer efficiency of the question-answering system is low, and only passive answer queries can be performed, which results in the inability of the user to stimulate potential.

Disclosure of Invention

The invention provides an intelligent coaching method and device, electronic equipment and a storage medium, aiming at the technical problems that the answer efficiency of the question-answering system in the prior art is low, and only passive answer query can be carried out, so that a user cannot excite potential.

In order to achieve the purpose, the invention adopts the technical scheme that:

the embodiment of the invention provides an intelligent coaching method, which comprises the following steps:

an active coaching step: based on the known data input by the user, after entity identification and relationship identification are carried out in the knowledge graph, active coach type enlightening is carried out on the user by adopting an enlightening coach technology, a plurality of open type questions are sent out, the user answers based on the open type questions, and a final answer is found based on an open type question sorting thought;

a bidirectional training step: after the user inputs the known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions and finds the final answer.

Preferably, the intelligent coaching method further comprises:

establishing a knowledge graph: and establishing a knowledge graph based on the historical knowledge base.

Preferably, the intelligent coaching method further comprises:

a passive coaching step: the user directly puts forward a question to the system, and the coach system carries out entity recognition and relation recognition based on the knowledge graph, finds a final answer and answers the user.

Preferably, the knowledge-graph establishing step includes:

self-learning step: and the user can continuously and autonomously learn from a historical knowledge base, so that knowledge map information is enriched.

The embodiment of the invention provides an intelligent coaching device, which adopts the intelligent coaching method, and the intelligent coaching device comprises:

an active coaching unit: the system is used for carrying out active coach type enlightening on a user by adopting enlightening coach technology after entity recognition and relation recognition are carried out in a knowledge graph module based on user input known data, and sending out a plurality of open type questions, wherein the user answers the questions based on the open type questions and finds a final answer based on an open type question sorting thought;

a bidirectional trainer unit: after the user inputs the known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions and finds the final answer.

Preferably, the intelligent coaching device further comprises:

a knowledge graph establishing unit: the method is used for establishing the knowledge graph based on the historical knowledge base.

Preferably, the intelligent coaching device further comprises:

preferably, the intelligent coaching device further comprises:

a passive coaching unit: the user directly puts forward a question to the system, and the coach system carries out entity recognition and relation recognition based on the knowledge graph, finds a final answer and answers the user.

Preferably, the knowledge-graph creating unit includes:

a self-learning unit: the method is used for continuous autonomous learning from the historical knowledge base and enriching knowledge map information.

The embodiment of the invention provides electronic equipment, which comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are used for completing mutual communication through the communication bus, and the memory is used for storing a computer program; a processor for performing the method steps in any of the above embodiments by executing the computer program stored on the memory.

An embodiment of the present invention provides a computer-readable storage medium, in which a computer program is stored, where the computer program is configured to execute the method steps in any of the above embodiments when the computer program is executed.

Compared with the prior art, the invention has the advantages and positive effects that:

(1) the intelligent coach method supports the user to achieve achievements in a question asking mode, and is willing to act to complete the excitation of potential energy, so that the user can finally clear the thought of the user and finally find answers;

(2) the intelligent training method mainly solves the problems of the open type training method and combines machine thinking with user intelligence;

(3) the trainer of the intelligent training method can actively ask a problem and guide the trainee to think;

(4) the intelligent coaching method disclosed by the invention combines three working modes, so that the expandability of a use scene is greatly improved;

(5) the intelligent coaching method is based on knowledge graph technology, and through autonomous learning of a knowledge graph library, the answer hit rate and the heuristic question number of the coaching method are greatly improved;

(6) the content of the knowledge graph is continuously and iteratively updated through continuous autonomous learning, the existing knowledge is expanded, new knowledge is added, and knowledge fusion is realized.

Drawings

FIG. 1 is a flowchart illustrating a method for intelligent coaching according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a scenario in accordance with an embodiment of the present invention;

FIG. 3 is a schematic flow chart of an active mode according to an embodiment of the present invention;

FIG. 4 is a diagram illustrating scenario two according to an embodiment of the present invention;

FIG. 5 is a schematic view of a passive mode process according to an embodiment of the present invention;

FIG. 6 is a diagram illustrating a scenario three according to an embodiment of the present invention;

FIG. 7 is a schematic flow chart of a bi-directional mode of an embodiment of the present invention;

FIG. 8 is a diagram illustrating an autonomous learning scenario in accordance with an embodiment of the present invention;

FIG. 9 is a schematic view of a knowledge graph;

FIG. 10 is a diagram of a knowledge-graph technology architecture;

FIG. 11 is a schematic diagram of an ontology construction;

FIG. 12 is a schematic diagram of logic-based reasoning;

FIG. 13 is a diagram based reasoning diagram;

FIG. 14 is a schematic diagram of deep learning based reasoning;

FIG. 15 is a schematic diagram of an intelligent trainer according to the invention;

fig. 16 is a block diagram of an alternative electronic device according to an embodiment of the present application.

In the above figures:

100. intelligent training device

10. Knowledge graph establishing unit 20 and active coaching unit

30. Bidirectional training unit 40 and passive training step unit

401. Processor 402, communication interface

403. Memory 404, communication bus

Detailed Description

The technical solutions in the embodiments of the present invention will be fully described in detail below with reference to the accompanying drawings. It is obvious that the described embodiments are only some specific embodiments, not all embodiments, of the general technical solution of the present invention. All other embodiments, which can be derived by a person skilled in the art from the general idea of the invention, fall within the scope of protection of the invention.

In the description of the present invention, it is to be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience of description and simplicity of description, and do not indicate or imply that the referenced devices or elements must have a particular orientation, be constructed and operated in a particular orientation, and thus, are not to be construed as limiting the present invention.

The terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.

The invention discloses an intelligent coaching method, which aims to realize an intelligent coaching method, adopts an enlightening coaching technology, carries out a brain storm type man-machine intelligent communication mode based on a knowledge graph, combines user thinking and machine intelligence to find rules and relations among things by diverging thinking, and assists a user to finally clear thinking by questioning in a coaching mode to obtain a final answer.

The inspired training technique provided by the invention is a way of asking questions, supports the user to achieve the result, and is willing to act to complete the excitation of potential energy. The user finally clears the thought of the user by himself and finally finds the answer. Furthermore, the problem of coach technique is mainly open-ended, such as how do you see? What is wanted? If understood? Without asking a closure question, e.g.,? Is there a pair of pairs? In summary, the intelligent coaching technology or coaching-type communication of the present invention is a way to help users to clear their thoughts in an intelligent analysis decision system.

The following detailed description of specific embodiments of the invention refers to the accompanying drawings in which:

fig. 1 is a flowchart illustrating a smart coaching method according to an embodiment of the present invention, and as shown in fig. 1, the smart coaching method according to the embodiment of the present invention includes:

active coaching step S20: based on the known data input by the user, after entity identification and relation identification are carried out in the knowledge graph, active coach type enlightening is carried out on the user, a plurality of open type questions are sent out, the user answers based on the open type questions, and a final answer is found based on an open type question sorting thought;

bidirectional coaching step S30: after the user inputs the known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions and finds the final answer.

Further, the intelligent coaching method of the present invention further comprises:

a knowledge map creation step S10: establishing a knowledge graph based on a historical knowledge base;

further, the intelligent coaching method of the present invention further comprises:

passive coaching step S40: the user directly puts forward a question to the system, and the coach system carries out entity recognition and relation recognition based on the knowledge graph, finds a final answer and answers the user.

Wherein the knowledge-graph establishing step S10 includes:

self-learning step: and the user can continuously and autonomously learn from a historical knowledge base, so that knowledge map information is enriched.

The specific embodiment of the invention is as follows:

assuming that a case has occurred in the public security system, the clerk has some information: such as suspect, time, location, vehicle, lodging conditions, etc. Deep analysis is needed to find the connection and lock the suspect.

FIG. 2 is a schematic diagram of a scenario of an embodiment of the present invention, as shown in FIG. 2, illustrating an intelligent robot, in which the intelligent coaching method of the present invention is implemented. The small and clear role is a communication robot which has the function of acting as a case analysis intelligence library and is highly independent. Meanwhile, the Xiaoming supports voice input and output, and the background is provided with a knowledge graph technology for correlation analysis. The invention intelligent coaching system: actively ask questions to guide the party being coached to think. Trainee (i.e., user): in the process of answering the coach question, the user can keep on the way of thinking and thinking.

There are three modes that can be selected: active mode (coaching), passive mode (coached), bidirectional mode (both coaching and coached);

the first mode of operation of the Xiaoming: active mode, as shown in fig. 2:

questions to the user are asked at little initiative:

what is a suspect? Motivation is?

What is the relationship between suspects a and B? The role in this case is?

C suspects on the timeline, the role in this case is? Why? What is the potential risk?

After the user answers the questions, the Xiaoming can continue to raise the deeply dug questions according to the answers of the user, and the questions can be also raised.

Fig. 3 is a schematic flow chart of an active mode according to an embodiment of the present invention, and as shown in fig. 3, the detailed flow chart includes:

(1) inputting known data into the knowledge graph by a user;

(2) the Xiaoming carries out entity recognition and relation recognition based on the knowledge graph;

(3) the Xiaoming asks the user in a coaching mode based on historical questions searched in the knowledge graph;

(4) the user carries out brainstorming and deep thinking based on coaching questions;

(5) the user answers the question after thinking and finds the final answer.

And (2) a small and clear working mode II: fig. 4 is a schematic diagram of a scenario two according to a specific embodiment of the present invention, as shown in fig. 4:

xiaoming passively answers the user's question:

simple answer:

the relationship between the A suspect and the B suspect is friend and relative.

Complex answers:

as a role in this case? Why? What is the potential risk?

This problem places high demands on the machine intelligence. It is necessary to extract learned knowledge from historical knowledge.

Fig. 5 is a schematic diagram of a passive mode process according to an embodiment of the present invention, and as shown in fig. 5, the detailed process includes:

(1) the user directly asks questions to the Xiaoming with voice;

(2) performing voice recognition and semantic analysis on the Xiaoming;

(3) the Xiaoming carries out entity recognition and relation recognition based on the knowledge graph;

(4) searching answers in the knowledge graph by Xiaoming;

(5) the Xiaoming converts the answer into voice, and the voice is output to tell the user the answer.

And (3) a small and clear working mode III: fig. 6 is a schematic view of a scenario three according to a specific embodiment of the present invention, as shown in fig. 6:

the Xiaoming can not only raise the question, but also answer the question.

Fig. 7 is a schematic diagram of a bidirectional mode process according to an embodiment of the present invention, and as shown in fig. 7, the detailed process is:

(1) inputting known data into the knowledge graph by a user;

(2) if the xiaoming is based on the knowledge graph, entity recognition and relationship recognition are carried out;

(3) then the xiaoming asks the user in a coaching mode based on the historical questions searched in the knowledge graph;

(4) the user carries out brainstorming and deep thinking based on coaching questions;

(5) the user answers the question after thinking and finds the final answer.

(6) Or, if the user is performing a coach-type question;

(7) the minuscule answers the user's question based on the knowledge graph and finds the final answer.

For small and clear autonomous learning, fig. 8 is a schematic view of an autonomous learning scenario according to an embodiment of the present invention, as shown in fig. 8:

xiaoming learns from the historical knowledge base independently and enriches knowledge map information of the Xiaoming learning knowledge base. The service is output to the outside with stronger analysis capability.

The historical knowledge base of the knowledge graph comprises: historical knowledge, historical cases, historical conclusions and historical relationships.

The small and clear knowledge graph capability comprises event association, event inference, character association, character recommendation, character analysis, potential problems, possible directions, habit inference, probability analysis and the like.

First, knowledge graph definition, essentially, is a semantic network that reveals relationships between entities. For quickly describing concepts and their interrelationships in the physical world.

The knowledge graph is converted into a simple and clear triple of entities, relations and entities by effectively processing, processing and integrating the data of the complicated document, and finally a great deal of knowledge is aggregated, so that the quick response and reasoning of the knowledge are realized.

The knowledge graph has two construction modes of top-down and bottom-up. The top-down construction is to extract ontology and mode information from high-quality data by means of structured data sources such as encyclopedic websites and the like, and add the ontology and mode information into a knowledge base; the bottom-up construction is that a resource mode is extracted from publicly acquired data by a certain technical means, a new mode with higher confidence coefficient is selected, and the new mode is added into a knowledge base after manual review.

FIG. 9 is a diagram of a knowledge graph, as shown in FIG. 9, if there is a Relationship between two nodes, they are connected by an undirected edge, and then this node is called an Entity (Entity), and this edge between them is called a Relationship (Relationship).

The basic unit of the knowledge graph is a triple formed by an Entity (Entity) -Relationship (Relationship) -Entity (Entity), and the triple is also the core of the knowledge graph.

Entity refers to something that is distinguishable and exists independently. The entity is the most basic element in the knowledge graph, and different relationships exist among different entities. Such as "china", "beijing", "16410 square kilometers", etc. in the figure.

Relationship is a connection between different entities, referring to a connection between entities. And connecting the nodes in the knowledge graph through the relation nodes to form a large graph. Such as "population", "capital", "area", etc. in the figure

Data type and storage mode

The raw data types of a knowledge graph generally have three classes (also three classes of raw data on the internet):

structured Data (structured Data): such as a relational database

Semi-structured Data (Semi-structured Data): such as XML, JSON, encyclopedia

Unstructured Data (unstructured Data): such as pictures, audio, video, text

Logic and technology architecture

1. Logic architecture

The knowledge graph can be logically divided into two levels, namely a mode layer and a data layer.

The mode layer is built on the data layer and is the core of the knowledge graph, and an ontology base is generally adopted to manage the mode layer of the knowledge graph. The ontology is a concept template of the structured knowledge base, and the knowledge base formed by the ontology base has a strong hierarchical structure and a small redundancy degree.

Mode layer: entity-relationship-entity, entity-attribute-property values

The data layer is mainly composed of a series of facts, and knowledge is stored in the unit of facts. If facts are expressed in triplets of (entity 1, relationship, entity 2), (entity, attribute value), graph databases may be selected as storage media, such as open-source Neo4j, Twitter's FlockDB, sones' GraphDB, and so on.

And (3) a data layer: Belgium-wife-Meilin-Gaiz, Belgium-President-Microsoft

2. Technical architecture

Fig. 10 is a schematic diagram of a technical architecture of a knowledge graph, and the overall architecture of the knowledge graph is as shown in fig. 10, wherein a part inside a dashed box is a construction process of the knowledge graph and is also a process of updating the knowledge graph.

The construction and application of the large-scale knowledge base need the support of various intelligent information processing technologies. Knowledge elements such as entities, relationships, attributes and the like can be extracted from some published semi-structured and unstructured data through knowledge extraction technology. Through knowledge fusion, ambiguity between the referent items such as entities, relations and attributes and the fact objects can be eliminated, and a high-quality knowledge base is formed. Knowledge reasoning is to further mine implicit knowledge on the basis of the existing knowledge base, so that the knowledge base is enriched and expanded. The comprehensive vector formed by the distributed knowledge representation has important significance for the construction, reasoning, fusion and application of the knowledge base.

And (3) knowledge extraction: the method is mainly oriented to open link data, and can extract available knowledge units through an automatic technology, wherein the knowledge units mainly comprise 3 knowledge elements of entities (concept extension), relations and attributes, and a series of high-quality fact expressions are formed on the basis of the knowledge units, so that a foundation is laid for constructing an upper mode layer. Knowledge extraction has three main tasks:

and (3) entity extraction: in the art we refer more to the NER (named entity recognition), which refers to the automatic recognition of named entities from the original corpus. Since the entity is the most basic element in the knowledge-graph, the completeness, accuracy, recall rate and the like of the extraction directly influence the quality of the knowledge base. Therefore, entity extraction is the most basic and critical step in knowledge extraction;

and (3) extracting the relation: the aim is to solve the problem of semantic linkage between entities, and early relation extraction mainly identifies entity relations by a method of manually constructing semantic rules and templates. Subsequently, the relationship model between the entities gradually replaces the manually predefined grammar and rules.

And (3) extracting attributes: the attribute extraction is mainly for the entity, and a complete sketch of the entity can be formed through the attribute. Since the attribute of the entity can be regarded as a name relationship between the entity and the attribute value, the extraction problem of the entity attribute can be converted into a relationship extraction problem.

Knowledge representation

In recent years, an expression learning technology represented by deep learning is advanced significantly, semantic information of an entity can be expressed as a dense low-dimensional real-valued vector, and further, the entity, a relationship and complex semantic association between the entities and the relationship are calculated efficiently in a low-dimensional space, so that the method has important significance for construction, reasoning, fusion and application of a knowledge base. The graph embedding is a representation of learning, as friends who pay attention to our public numbers must read the last blog.

Knowledge fusion

Knowledge in the knowledge map has a wide range of knowledge sources, and has problems of good knowledge quality, overlapping knowledge from different data sources, and unclear relation between knowledge, so that it is necessary to perform knowledge fusion. Knowledge fusion is a high-level knowledge organization, so that knowledge from different knowledge sources is subjected to steps of heterogeneous data integration, disambiguation, processing, reasoning verification, updating and the like under the same frame specification, fusion of data, information, methods, experiences and human ideas is achieved, and a high-quality knowledge base is formed.

Among them, knowledge updating is an important part. The cognitive abilities, knowledge reserves and business needs of humans are increasing over time. Therefore, the content of the knowledge graph needs to be advanced, and the knowledge graph, whether a general knowledge graph or an industry knowledge graph, needs to be continuously updated in an iterative mode, existing knowledge is expanded, and new knowledge is added.

Application of knowledge graph

The knowledge graph provides a more effective mode for the expression, organization, management and utilization of massive, heterogeneous and dynamic big data on the Internet, so that the intelligent level of the network is higher and is closer to the cognitive thinking of human beings.

1. Intelligent search

After a user's query is entered, the search engine does not merely look for keywords, but rather first performs semantic understanding. For example, after the query is participled, the description of the query is normalized so that it can be matched with the knowledge base. The returned result of the query is a complete knowledge system given by the search engine after searching the corresponding entity in the knowledge base.

2. Deep question and answer

The question-answering system is a high-level form of an information retrieval system and can provide users with answers to questions in accurate and concise natural language. Most of the question-answering systems tend to decompose a given question into a plurality of small questions, then extract matched answers one by one from a knowledge base, automatically detect the matching degree of the answers in time and space and the like, and finally combine the answers and display the answers to users in a visual mode.

The intelligent voice assistant Siri of the apple can provide services such as answers, introduction and the like for the user, and is a result of introducing the knowledge graph. Knowledge maps allow machine-to-human interaction to appear more intelligent.

Fourthly, construction of knowledge graph

The first step is as follows: information extraction

Information extraction (information extraction) is the 1 st step of knowledge graph construction, and the key problems are as follows: how to automatically extract information from heterogeneous data sources to obtain candidate indication units.

Information extraction is a technique for automatically extracting structured information such as entities, relationships, and entity attributes from semi-structured and unstructured data.

The related key technologies comprise: entity extraction, relationship extraction and attribute extraction.

The second step is that: knowledge fusion

Knowledge fusion includes 2 parts of content: entity linking, knowledge consolidation

1. Entity linking (entity linking): the method refers to an operation of linking an entity object extracted from a text to a corresponding correct entity object in a knowledge base.

The basic idea is to first select a set of candidate entity objects from the knowledge base according to a given entity designation, and then link the designation to the correct entity object through similarity calculation.

Study history:

only how to link the entities extracted from the text into the knowledge base is concerned, ignoring the semantic connections that exist between entities located in the same document.

Attention is being turned to exploiting co-occurrence relationships of entities while linking multiple entities into a knowledge base. Namely, integration entity linking (collecting entity linking)

Entity linking process:

and extracting entity name items from the text through the entities.

And carrying out entity disambiguation and coreference resolution, and judging whether the same-name entities in the knowledge base represent different meanings and whether other named entities exist in the knowledge base and represent the same meanings.

After confirming the corresponding correct entity object in the knowledge base, linking the entity designation chain to the corresponding entity in the knowledge base.

Entity disambiguation: the technology specially used for solving the problem of ambiguity generated by the same-name entity can accurately establish entity link according to the current context through entity disambiguation, and the entity disambiguation mainly adopts a clustering method. It can also be thought of as a context-based classification problem, similar to part-of-speech disambiguation and sense disambiguation.

Performing coreference resolution: the method is mainly used for solving the problem that a plurality of designations correspond to the same entity object. In a session, multiple references may point to the same entity object. These terms can be associated (merged) to the correct entity object using coreference resolution techniques, which attracts a lot of research efforts due to the special importance of this problem in the fields of information retrieval and natural language processing. Coreference resolution also has some other names such as object alignment, entity matching, and entity synonymity.

2. And (3) knowledge merging: in constructing the knowledge graph, knowledge input may be obtained from third party knowledge base products or existing structured data.

Common knowledge consolidation requirements are two, one is to consolidate external knowledge bases and the other is to consolidate relational databases.

Fusing the external knowledge base to the local knowledge base requires handling two levels of problems:

the fusion of data layers, including the designation, attribute, relationship and belonging category of the entity, has the main problem of how to avoid the conflict between instances and relationships, resulting in unnecessary redundancy

Fusing newly obtained ontology into the existing ontology library through the fusion of the mode layer

Then, the relational database is merged, and an important high-quality knowledge source is the own relational database of an enterprise or an organization in the process of constructing the knowledge graph. In order to incorporate these structured historical data into the knowledge graph, a Resource Description Framework (RDF) may be employed as a data model. This data conversion process is referred to figuratively by the industry and academia as RDB2RDF, which is essentially the conversion of data from a relational database into triple data of RDF.

The third step: knowledge processing

In the foregoing, we have extracted knowledge elements such as entities, relationships, attributes, and the like from the original corpus through information extraction, and have removed ambiguity between entity names and entity objects through knowledge fusion, to obtain a series of basic fact expressions.

However, the fact itself is not equal to knowledge. To finally obtain a structured and networked knowledge system, a knowledge processing process is required.

The knowledge processing mainly comprises 3 aspects: ontology construction, knowledge reasoning and quality assessment.

1. Ontology construction

Ontology refers to a concept set, concept framework, such as "people", "things", etc., of a worker.

The ontology can be manually constructed in a manual editing mode (by means of ontology editing software) or can be constructed in a data-driven automatic mode. Because the workload of the manual mode is huge and experts meeting the requirements are difficult to find, the current mainstream global ontology library products are obtained by gradually expanding the existing ontology libraries oriented to specific fields by adopting an automatic construction technology.

The automated ontology building process comprises three phases:

similarity calculation of entity parallel relation

Entity context extraction

Generation of ontologies

Fig. 11 is a schematic diagram of the construction of the body, as shown in fig. 11: when the knowledge graph obtains three entities, namely the entities of the Alibab, the Tencent and the mobile phone, the knowledge graph may think that the three entities have no difference, but when the knowledge graph calculates the similarity between the three entities, the knowledge graph finds that the Alibab and the Tencent are likely to be more similar and have larger difference with the mobile phone.

This is the first step, but in this way, the knowledge map does not actually have the concept of upper and lower layers, and it is unknown that the agilawood and the mobile phone are not in the same type at all and cannot be compared. Therefore, we need to complete the above work in the step of extracting the context in the entity, so as to generate the ontology of the third step.

When the three steps are completed, the knowledge graph may understand that "Alibara and Tencent, which are actually the subdivisions of the company as one entity. They are not generic to handsets. "

Knowledge reasoning

After we have completed this step of ontology construction, a prototype of the knowledge-graph has been built. However, most of the relations between knowledge maps are incomplete at this time, and the missing values are very serious, so that at this time, a knowledge reasoning technology can be used to complete further knowledge discovery.

We can find that: if A is the spouse of B, B is the chairman of C, and C is located at D, then we can consider that A lives in the city of D.

According to the rule, we can dig out whether there are other paths in the graph to satisfy the condition, and then we can associate two AD. In addition, we can think that a ring in the series is the chairman of B being C, then CEO of B being C, COO of B being C, is not a ring that can also be used as the inference strategy?

Of course, the object of knowledge inference is not limited to the relationship between entities, and may be the attribute values of the entities, the conceptual hierarchical relationship of the ontology, and the like.

Reasoning attribute values: knowing the birthday attribute of an entity, the age attribute of the entity can be obtained through reasoning;

reasoning concept: it is known that (Tiger, family, Felidae) and (Felidae, order, Carnivora) can be deduced (Tiger, order, Carnivora)

Fig. 12 is a schematic diagram of logic-based reasoning, fig. 13 is a schematic diagram of graph-based reasoning, and fig. 14 is a schematic diagram of deep learning-based reasoning, and as shown in fig. 12-14, the algorithm of the block can be mainly classified into 3 categories, logic-based reasoning, graph-based reasoning and deep learning-based reasoning.

Fifth, summarize

Through the knowledge map, the information of the internet can be expressed into a form closer to the human cognitive world, and a mode for better organizing, managing and utilizing mass information is provided. The current knowledge-graph technology is mainly used for intelligent semantic search, mobile personal assistant (Siri) and deep question and answer system (Watson), and the core technology supporting the applications is the knowledge-graph technology.

In the intelligent semantic search, when a user initiates a query, a search engine can analyze and reason the keywords queried by the user by the aid of a knowledge graph, then map the keywords to one or a group of concepts in the knowledge graph, and then return a graphical knowledge structure to the user according to the concept hierarchy of the knowledge graph, namely a knowledge card seen in Google and hundred-degree search results.

In deep question-and-answer applications, the system also performs semantic analysis and syntax analysis on questions posed by a user using natural language with the help of a knowledge graph, converts the questions into query sentences in a structured form, and queries answers in the knowledge graph. For example, if the user asks: "how to determine if an ebola virus infection is present? If so, then the query may be equivalently transformed into? And then performing reasoning transformation to finally form equivalent triple query sentences, such as (ebola, symptom. If the user's question cannot be answered by reasoning due to the imperfect knowledge base, the deep question-answering system can also feed back the search result to the user by using the search engine, and update the knowledge base according to the search result, thereby preparing in advance for answering subsequent questions.

The key point of the present invention is to use open questions to guide users or robots (Xiaoming) to use knowledge graph technology to perform deep summarization to obtain answers.

Fig. 15 is a schematic structural diagram of an intelligent coaching device according to the present invention, and as shown in fig. 15, an embodiment of the present invention further provides an intelligent coaching device 100, which uses the above-mentioned intelligent coaching method, and the device uses an heuristic coaching technique, and performs a brainstorming type man-machine intelligent communication mode based on a knowledge graph, and combines user thinking with machine intelligence to find rules and connections between things by divergent thinking, and helps a user to finally clear thinking by a coaching type question, and obtains a final answer.

Wherein, intelligent trainer device 100 includes: active coaching unit 20 and bi-directional coaching unit 30

Active coaching unit 20: the system is used for carrying out active coach type enlightenment on a user after entity identification and relation identification are carried out in a knowledge graph module based on user input known data, sending out a plurality of open type questions, answering the questions based on the open type questions by the user, sorting thought based on the open type questions and finding a final answer;

the bidirectional trainer unit 30: after the user inputs the known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions and finds the final answer.

Further, the smart trainer device 100 further includes:

the knowledge-graph establishing unit 10: the system is used for establishing a knowledge graph based on a historical knowledge base;

the knowledge graph establishing unit 10 includes: the self-learning unit is used for continuously and autonomously learning from a historical knowledge base and enriching knowledge map information;

further, the smart trainer device 100 further includes:

passive coaching step unit 40: the user directly puts forward a question to the system, and the coach system carries out entity recognition and relation recognition based on the knowledge graph, finds a final answer and answers the user.

Fig. 16 is a block diagram of an alternative electronic device according to an embodiment of the present application, and as shown in fig. 16, an embodiment of the present invention provides an electronic device including a processor 401, a communication interface 402, a memory 403, and a communication bus 404, where the processor 401, the communication interface 402, and the memory 403 complete communication with each other through the communication bus 404, and the memory 403 is used for storing a computer program; a processor 401 for performing the method steps of any of the above embodiments by running said computer program stored on the memory.

Active coaching step S20: based on the known data input by the user, after entity identification and relation identification are carried out in the knowledge graph, active coach type enlightening is carried out on the user, a plurality of open type questions are sent out, the user answers based on the open type questions, and a final answer is found based on an open type question sorting thought;

bidirectional coaching step S30: after the user inputs the known data, the user carries out coach type questioning, and a coach system finds a final answer and answers based on the knowledge graph; or the coaching system carries out coaching type questioning based on the knowledge graph, inspires the user to think and answer the questions and finds the final answer.

In a specific embodiment of the present invention, the device implementing the intelligent coaching method may be an intelligent robot or other terminal Devices, and the terminal Devices may be terminal Devices such as a smart phone (e.g., an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, a Mobile Internet Device (MID), a PAD, and the like. Fig. 16 does not limit the structure of the electronic device. For example, the terminal device may also include more or fewer components (e.g., network interfaces, display devices, etc.) than shown in fig. 16, or have a different configuration than shown in fig. 16.

Embodiments of the present invention provide a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of the intelligent coaching method in any of the above embodiments when the computer program is executed. In this embodiment, the storage medium may include, but is not limited to: various media capable of storing program codes, such as a U disk, a ROM, a RAM, a removable hard disk, a magnetic disk, or an optical disk.

The intelligent coaching method supports the user to achieve the result in a questioning mode, is willing to act and completes the excitation of potential. The user finally clears the thought of the user and finally finds an answer; the intelligent training method mainly solves the problems of the open type training method and combines machine thinking with user intelligence; the trainer of the intelligent training method can actively ask a problem and guide the trainee to think; the intelligent coaching method disclosed by the invention combines three working modes, so that the expandability of a use scene is greatly improved; the intelligent coaching method is based on knowledge graph technology, and through autonomous learning of the knowledge graph library, the answer hit rate and the heuristic question number of the coaching method are greatly improved.

The above description is for the purpose of describing particular embodiments of the present invention, and is not intended to limit the scope of the present invention, which is defined by the claims and their equivalents, and all changes and modifications that can be made therein without departing from the spirit and scope of the present invention. It is noted that in the drawings and in the description, implementations not shown or described are all in a form known to those of ordinary skill in the art and are not described in detail. Furthermore, the above definitions of the various components and processes are not intended to be limited to the specific structures, shapes, or configurations shown in the examples.