Knowledge base question and answer method, system and architecture for pottery enterprise system

Abstract

The invention belongs to the technical field of operation and maintenance of a pottery enterprise system, and particularly relates to a knowledge base question answering method, system and architecture for the pottery enterprise system, and the knowledge base question answering method for the pottery enterprise system comprises the steps: responding to a question input by a user, and converting the question into a semantic vector; obtaining a knowledge graph, and obtaining a semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; configuring and / or adjusting a semantic similarity weight based on a preset factor; and screening out answers according to the semantic similarity and the semantic similarity weight. According to the method, the problems can be automatically analyzed, accurate answers or suggestions can be given, compared with a traditional knowledge base, the method has the higher intelligent level and the higher response speed, the enterprise system implementation and operation and maintenance efficiency and quality can be greatly improved, and automatic and intelligent development of enterprise system implementation and operation and maintenance work is promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of ceramic enterprise system operation and maintenance, and in particular, relates to a knowledge base question and answer method, system and architecture for a ceramic enterprise system. Background Art

[0002] In today's rapidly developing information age, the implementation and operation of ceramic enterprise systems have become increasingly complex. With the continuous advancement of artificial intelligence technology, more and more ceramic companies are beginning to explore how to use artificial intelligence technology to improve the implementation and operation efficiency of enterprise systems, thereby reducing costs and improving quality.

[0003] Traditional ceramic enterprise system implementation and operation and maintenance knowledge bases have problems such as delayed information updates and low search efficiency. Although existing technologies provide some solutions, they are often limited by static knowledge base structures and fixed search algorithms, and are unable to update and adapt to the complex and changing system needs of ceramic enterprises in a timely manner.

[0004] In addition, during the implementation and operation of ceramic enterprise systems, manual search and problem solving is inefficient and easily limited by professional knowledge and experience. In the actual implementation process of ceramic enterprises, the knowledge involved in the question-and-answer database is vast and complex, and the traditional knowledge acquisition and sharing methods are inefficient and cannot meet the rapidly changing needs; the traditional problem-solving methods require manual intervention, the response time is long, and the problems in implementation and operation cannot be solved in time; the knowledge of enterprise system implementation and operation is constantly updated, and the traditional knowledge base maintenance method is cumbersome and easy to become outdated, and cannot be updated in time. Summary of the invention

[0005] The present invention proposes a knowledge base question-answering method, system and architecture for a ceramic enterprise system, which can solve at least one technical problem in the above-mentioned background technology.

[0006] A first aspect of the present invention provides a knowledge base question-answering method for a ceramic enterprise system, comprising: Responding to user input questions and converting the questions into semantic vectors; Obtain a knowledge graph, and obtain the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; Configuring and / or adjusting semantic similarity weights based on preset factors; The answer is screened out according to the semantic similarity and the semantic similarity weight.

[0007] As an optional implementation, the obtaining of the knowledge graph includes: Acquire and extract key entity knowledge from entity knowledge; identifying entity relationships within the key entity knowledge; A knowledge graph is constructed based on the key entity knowledge and the entity relationships.

[0008] As an optional implementation manner, the acquiring and extracting key entity knowledge from entity knowledge includes: acquiring a ceramic enterprise data source, and extracting entity knowledge from the ceramic enterprise data source; Natural language processing technology is used to perform entity recognition on the entity knowledge to extract key entity knowledge.

[0009] As an optional implementation manner, the identifying entity relationships within the key entity knowledge includes: Based on the selected dependency syntax parser, each sentence in the key entity knowledge is converted into a dependency tree structure, wherein the dependency tree structure uses entities as nodes and represents the dependency relationship between entities through edges; Get relationship anchor points based on preset scenarios; Extract relational features from dependency trees; Based on the relationship anchor points and the relationship features, entity relationships within the key entity knowledge are identified through a preset training model.

[0010] As an optional implementation, the obtaining the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm includes: The distance between the semantic vector and the knowledge base document in the knowledge graph is calculated based on a preset distance algorithm, and the semantic similarity is evaluated based on the distance.

[0011] As an optional implementation, the preset factors include at least content factors, context factors and user feedback factors; wherein, Configure semantic similarity weights based on content factors, including: Configure semantic similarity weights based on the importance and usage frequency of knowledge items in key entity content; Adjust semantic similarity weights based on contextual factors, including: Adjusting the semantic similarity weight based on the context information of the user's input question in response; Adjust semantic similarity weights based on user feedback factors, including: The user evaluation and user interaction data are quantified into preset feedback values, and the semantic similarity weights are adjusted based on the preset feedback values.

[0012] As an optional implementation manner, the screening out answers according to the semantic similarity and the semantic similarity weight includes: Retrieving a preset number of candidate answers according to the semantic similarity and the semantic similarity weight; The candidate answers are screened based on preset optimization conditions to obtain a unique answer from the candidate answers.

[0013] A second aspect of the present invention provides a knowledge base question-answering system for a ceramic enterprise system, comprising: A response unit, at least used to respond to a user's input question and convert the question into a semantic vector; An acquisition unit, at least used to acquire a knowledge graph, and acquire the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; A configuration unit, at least for configuring and / or adjusting the semantic similarity weight based on preset factors; The screening unit is at least used to screen out answers according to the semantic similarity and the semantic similarity weight.

[0014] As an optional implementation manner, the acquisition unit includes a first acquisition unit, and the first acquisition unit is at least used for: Acquire and extract key entity knowledge from entity knowledge; identifying entity relationships within the key entity knowledge; A knowledge graph is constructed based on the key entity knowledge and the entity relationships.

[0015] As an optional implementation manner, the acquisition unit includes a second acquisition unit, and the second acquisition unit is at least used for: The distance between the semantic vector and the knowledge base document in the knowledge graph is calculated based on a preset distance algorithm, and the semantic similarity is evaluated based on the distance.

[0016] A third aspect of the present invention provides a knowledge base question-answering architecture for a ceramic system, comprising: The data acquisition module is at least used to obtain the data source of the ceramic enterprise and transmit the data to the data processing module; The data processing module is at least used to extract entity knowledge from the ceramic enterprise data source, perform entity recognition on the entity knowledge using natural language processing technology, and extract key entity knowledge; identify entity relationships within the key entity knowledge; and construct a knowledge graph based on the key entity knowledge and the entity relationships to achieve knowledge generation; The knowledge base management module is used to configure a graph database such as Neo4j to store the knowledge graph and respond to the knowledge retrieval of the question-answering module; The question-and-answer module is at least used to respond to the user's input questions and return answers, and to deliver the responses corresponding to the user's input questions to the question-and-answer module.

[0017] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a computer, the knowledge base question-answering method for a ceramic enterprise system as described in the second aspect of the present invention is executed.

[0018] In the present invention, users can ask questions in natural language, and the present invention will automatically analyze the questions and give accurate answers or suggestions. Compared with traditional knowledge bases, the present invention has a higher level of intelligence and a faster response speed, which can greatly improve the efficiency and quality of enterprise system implementation and operation and maintenance, and promote the automation and intelligent development of enterprise system implementation and operation and maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A flow chart of a knowledge base question-answering method for a ceramic enterprise system in an embodiment of the present invention is shown.

[0020] Figure 2 A flow chart of another knowledge base question-answering method for a ceramic enterprise system in an embodiment of the present invention is shown.

[0021] Figure 3 A flow chart of another knowledge base question-answering method for a ceramic enterprise system in an embodiment of the present invention is shown.

[0022] Figure 4 A flow chart of another knowledge base question-answering method for a ceramic enterprise system in an embodiment of the present invention is shown.

[0023] Figure 5 A block diagram of a knowledge base question-answering system for a ceramic enterprise system in an embodiment of the present invention is shown.

[0024] Figure 6 A block diagram of another knowledge base question and answer system for a ceramic enterprise system in an embodiment of the present invention is shown.

[0025] Figure 7 A block diagram of a knowledge base question-answering architecture for a ceramic enterprise system in an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0027] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and claims, the singular forms of "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations.

[0028] FastGPT is a natural language processing technology based on a large-scale pre-trained model. It is evolved from the Transformer model, an advanced technology in the field of deep learning. FastGPT performs well in processing natural language tasks and has strong language understanding and generation capabilities. It can automatically learn and understand large-scale text data and generate natural language text with logic and coherence.

[0029] The enterprise system implementation and operation and maintenance knowledge base refers to a database or knowledge base that contains a large amount of enterprise system implementation and operation and maintenance related knowledge. This knowledge includes experience and skills in system configuration, troubleshooting, performance optimization, etc. Traditional knowledge bases are usually static and require manual continuous updating and maintenance. They are inefficient and easily outdated. The solution provided by the present invention can be combined with the enterprise system implementation and operation and maintenance knowledge base for implementation.

[0030] For ceramic companies, since the ceramic industry involves multiple fields such as material science, design, and manufacturing, the demand for professional knowledge is relatively high. The ceramic production process includes multiple links such as raw material preparation, molding, drying, and sintering. Each link has its specific technical requirements and quality control standards.

[0031] Due to the continuous emergence of new technologies, the existing knowledge base may not be updated in a timely manner, resulting in inaccurate or outdated information provided; ceramic knowledge is scattered in different documents, patents, standards and expert experience, lacking systematic organization and summary; the knowledge of the ceramic industry is relatively professional, and non-professionals may find it difficult to understand and apply it, which increases the difficulty for ordinary users to obtain effective information; ceramic companies mainly focus on intelligent manufacturing, the level of collaboration between companies is low, and the knowledge sharing mechanism is not sound, making the knowledge dissemination efficiency within the industry inefficient; for some knowledge points involving patents or trade secrets, companies may be reserved when sharing publicly, which limits the richness of the content of the knowledge base; different regions and countries may have different standards for ceramic products, which also brings challenges to the establishment of the knowledge base.

[0032] like Figure 1 As shown, the first aspect of the present invention provides a knowledge base question-answering method for a ceramic enterprise system, comprising: Step S100: respond to the user's input question and convert the question into a semantic vector. Specifically, the question is input as text, and a pre-trained language model (such as BERT, FastGPT) is used to convert the text into a high-dimensional vector representation. The user's input question and the document in the knowledge base are converted into a semantic vector through an encoder; Step S200: obtaining a knowledge graph, and obtaining the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; Step S300: configuring and / or adjusting semantic similarity weights based on preset factors; Step S400: Filter out answers according to the semantic similarity and the semantic similarity weight.

[0033] Specifically, documents in the knowledge base are converted into semantic vectors to enable large models to retrieve and match more efficiently. Semantic vectors can capture the context and meaning in documents, allowing the model to understand and compare the similarities between texts. This conversion helps the model find relevant documents more accurately when facing queries, rather than relying solely on traditional keyword matching, thereby providing more accurate and contextual answers.

[0034] In the present invention, users can ask questions in natural language, and the present invention will automatically analyze the questions and give accurate answers or suggestions. Compared with traditional knowledge bases, the present invention has a higher level of intelligence and a faster response speed, which can greatly improve the efficiency and quality of enterprise system implementation and operation and maintenance, and promote the automation and intelligent development of enterprise system implementation and operation and maintenance work.

[0035] like Figure 2 As shown, in one embodiment of the present invention, the step S200 of acquiring the knowledge graph includes: Step S210: Acquire and extract key entity knowledge from entity knowledge, such as device name, fault type, solution, etc.; Step S220: identifying entity relationships within the key entity knowledge; Step S230: Construct a knowledge graph based on the key entity knowledge and the entity relationship. In the construction of the knowledge graph, a graph database such as Neo4j can be configured to store the knowledge graph to achieve efficient graph data management and query. The knowledge graph can be automatically updated by regularly acquiring the latest knowledge from the data source, and the accuracy and completeness of the knowledge graph can be ensured through manual review or other feedback mechanisms.

[0036] The present invention constructs a knowledge graph for enterprise system implementation and operation and maintenance to structure key entity knowledge so that relevant knowledge can be quickly located based on user questions. It also monitors database changes and updates them in a timely manner to ensure the timeliness and accuracy of knowledge by implementing an automated knowledge update and maintenance mechanism.

[0037] like Figure 3 As shown, in one embodiment of the present invention, the step S210 of acquiring and extracting key entity knowledge from entity knowledge includes: Step S211: Acquire the ceramic enterprise data source, and extract entity knowledge from the ceramic enterprise data source; specifically, the ceramic enterprise data source may be a variety of data sources such as ceramic enterprise implementation and operation and maintenance documents, fault reports and solutions, and entity knowledge may be collected from the above-mentioned data sources. The entity knowledge may be presented in the form of text, and further, natural language processing technology may be used to perform data cleaning and data preprocessing to remove irrelevant information and obtain the final entity knowledge. Step S212: Use natural language processing technology to perform entity recognition on the entity knowledge and extract key entity knowledge. Specifically, natural language processing technology can use technologies such as CRF (conditional random field), BiLSTM-CRF (bidirectional long short-term memory network + conditional random field) to extract key entity knowledge in the entity knowledge, such as equipment name, fault type, solution, etc.

[0038] Specifically, entity knowledge refers to detailed information about specific entities (such as names of people, places, organizations, etc.). Key entities are important objects or concepts in the text, and these entities play a key role in understanding and answering questions. They are usually presented in a structured form, for example: Entity knowledge: usually exists in the form of database entries, knowledge graph nodes, or dictionary entries, containing detailed information on attributes and relationships. Key entities: are usually nouns or phrases in the text, such as "Einstein", "Paris", "United Nations". These entities can be identified and extracted in sentences to enhance semantic understanding and information retrieval.

[0039] Specifically, entity recognition is a task in natural language processing that aims to identify and classify specific types of entities from text, such as names of people, places, dates, organizations, etc. Its core is to extract these entities from the text and label their categories. For example, in the sentence "Einstein published the theory of relativity in 1905", the goal of entity recognition is to identify "Einstein" as a name and "1905" as a date. Doing so helps enhance the ability of text understanding and information extraction.

[0040] As an alternative implementation method, natural language processing technology can identify and classify key entity knowledge from entity knowledge, analyze the dependency relationships in sentences based on dependency syntax, extract semantic relationships between entities, understand the semantic information in the text, and assist in relationship extraction.

[0041] Preferably, the present invention can use supervised learning models (such as CRF, LSTM) to extract entities and relations, and use pre-trained language models (such as BERT, GPT) to improve semantic understanding capabilities.

[0042] Specifically, Figure 4 As shown, in one embodiment of the present invention, the identifying entity relationships within the key entity knowledge in step S220 includes: Step S221: Based on the selected dependency syntax parser, each sentence in the key entity knowledge is converted into a dependency tree structure. The dependency tree structure uses entities as nodes and represents the dependency relationship between entities through edges. It has strong scalability. Step S222: obtaining a relationship anchor point based on a preset scenario; Step S223: extracting relationship features from the dependency tree; Step S224: Based on the relationship anchor points and the relationship features, entity relationships within the key entity knowledge are identified through a preset training model.

[0043] As an alternative, in an embodiment of the present invention, entity relationships within key entity knowledge, such as the association relationship between faulty equipment and solutions, may also be identified based on coreference resolution technology.

[0044] In one embodiment of the present invention, the obtaining of the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm includes: The distance between the semantic vector and the knowledge base document in the knowledge graph is calculated based on a preset distance algorithm, and the semantic similarity is evaluated based on the distance.

[0045] The preset distance algorithm may be cosine similarity, Euclidean distance, or Mahalanobis distance.

[0046] Among them, knowledge base documents are texts or records stored in the knowledge base, which may contain detailed entity knowledge and background information. Entity knowledge and key entities are components of knowledge base documents. Entity knowledge is detailed information about a specific entity, while key entities are entities that are important in a specific context. Using semantic vectors to calculate the distance between documents can help evaluate and compare their semantic similarity, thereby achieving more accurate information retrieval and answer generation.

[0047] In one embodiment of the present invention, the preset factors include at least content factors, context factors and user feedback factors; wherein configuring the semantic similarity weight based on the content factors includes: The semantic similarity weights are configured based on the importance and usage frequency of knowledge items in the key entity knowledge content. For example, frequently appearing knowledge items or answers that have been verified correct multiple times have higher weights. Among them, the semantic similarity weight is adjusted based on contextual factors, including: Adjust the semantic similarity weight based on the context information of the user's input question, for example, increase the weight of the answer with high context relevance to ensure the accuracy and coherence of the answer; Among them, the semantic similarity weight is adjusted based on user feedback factors, including: User evaluation and user interaction data are quantified into preset feedback values, and the semantic similarity weights are adjusted based on the preset feedback values. The weights of each knowledge item can be continuously adjusted according to user feedback and system learning, and the sorting and recommendation of knowledge items can be dynamically optimized through user evaluation and interaction data.

[0048] In one embodiment of the present invention, the step of selecting an answer according to the semantic similarity and the semantic similarity weight includes: Retrieving a preset number of candidate answers according to the semantic similarity and the semantic similarity weight. Specifically, the preset optimization condition can be performed by setting a threshold value, and only the answers with similarity higher than a certain standard are selected to enter the candidate set, become candidate answers, and selectively returned to the user; The candidate answers are screened based on preset optimization conditions, and a unique answer is obtained from the candidate answers and returned to the user. If achievable, the candidate answers can be further processed and optimized using FastGPT technology to generate a final answer that meets user needs, and a personalized and accurate answer can be generated based on contextual information and the user's historical question records, and returned to the user.

[0049] Based on the user's input questions and knowledge graphs, the present invention utilizes intelligent reasoning and recommendation technology to quickly generate accurate answers or suggestions and improve question response efficiency.

[0050] like Figure 5 As shown, the second aspect of the present invention provides a knowledge base question-answering system for a ceramic enterprise system, comprising: A response unit 51, at least used to respond to a question input by a user and convert the question into a semantic vector; An acquisition unit 52 is at least used to acquire a knowledge graph, and acquire a semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; A configuration unit 53, at least for configuring and / or adjusting the semantic similarity weight based on preset factors; The screening unit 54 is at least used to screen out answers according to the semantic similarity and the semantic similarity weight.

[0051] like Figure 6 As shown, in one embodiment of the present invention, the acquisition unit 52 includes a first acquisition unit 521, and the first acquisition unit 521 is at least used for: Acquire and extract key entity knowledge from entity knowledge; identifying entity relationships within the key entity knowledge; A knowledge graph is constructed based on the key entity knowledge and the entity relationships.

[0052] like Figure 6 As shown, in one embodiment of the present invention, the acquisition unit 52 includes a second acquisition unit 522, and the second acquisition unit 522 is at least used to: calculate the distance between the semantic vector and the knowledge base document in the knowledge graph based on a preset distance algorithm, and evaluate the semantic similarity based on the distance.

[0053] like Figure 6 As shown, in one embodiment of the present invention, the allocation unit 53 includes a first allocation unit 531 , a second allocation unit 532 and a third allocation unit 533 .

[0054] Specifically, the preset factors at least include content factors, context factors and user feedback factors, and each factor acts on the first allocation unit 531, the second allocation unit 532 and the third allocation unit 533 respectively.

[0055] The first allocation unit 531 is at least used to configure the semantic similarity weight based on content factors, including: Semantic similarity weights are configured based on the importance and usage frequency of knowledge items in the key entity knowledge content.

[0056] The second adjustment unit 532 is at least used to adjust the semantic similarity weight based on context factors, including: The semantic similarity weight is adjusted based on the context information of the user's input information received in response.

[0057] The third adjustment unit 533 is at least used to adjust the semantic similarity weight based on the user feedback factor, including: The user evaluation and user interaction data are quantified into preset feedback values, and the semantic similarity weights are adjusted based on the preset feedback values.

[0058] In one embodiment of the present invention, the screening unit includes retrieving a preset number of candidate answers according to the semantic similarity and the semantic similarity weight; The candidate answers are screened based on preset optimization conditions to obtain a unique answer from the candidate answers.

[0059] Based on the same idea as the method in the above-mentioned embodiment, the knowledge base question and answer system for the ceramic enterprise system provided by the present application can implement the method of the above-mentioned embodiment. For the convenience of explanation, the structural diagram of the system embodiment only shows the part related to the embodiment of the present application. Those skilled in the art can understand that the illustrated structure does not constitute a limitation on the system, and may include more or fewer modules than shown in the diagram, or a combination of certain modules, or a different module arrangement.

[0060] like Figure 7 As shown, the third aspect of the present invention provides a knowledge base question-answering architecture for a ceramic system, comprising: The data acquisition module is at least used to obtain the data source of the ceramic enterprise and transmit the data to the data processing module; The data processing module is at least used to extract entity knowledge from the ceramic enterprise data source, perform entity recognition on the entity knowledge using natural language processing technology, and extract key entity knowledge; identify entity relationships within the key entity knowledge; and construct a knowledge graph based on the key entity knowledge and the entity relationships to achieve knowledge generation; The knowledge base management module is used to configure a graph database such as Neo4j to store the knowledge graph and respond to the knowledge retrieval of the question-answering module; The question-and-answer module is at least used to respond to the user's input questions and return answers, and to deliver the responses corresponding to the user's input questions to the question-and-answer module.

[0061] Specifically, Figure 7 As shown, it shows the process of data from collection to processing, storage and utilization, and it shows in detail the working logic and data flow direction between modules: (Taoqi) data source → data collection → data preprocessing → information extraction → knowledge generation → knowledge base storage → knowledge retrieval → user query.

[0062] Specifically, in an implementable manner, the architecture can also be expressed as: 1. User interface module: provides an interface for users to interact with the system, including a text input box and a question submission button. 2. Natural language processing module: Use natural language processing technologies such as FastGPT to perform semantic analysis and understanding of questions input by users. 3. Knowledge graph module: Build and maintain the knowledge graph for enterprise system implementation and operation, and represent knowledge in the form of a graph, including nodes and relationships. 4. Intelligent reasoning and recommendation module: Based on user questions and knowledge graphs, it uses intelligent reasoning and recommendation technology to generate accurate answers or suggestions. 5. Knowledge update and maintenance module: monitors changes in the knowledge base and automatically updates it to ensure the timeliness and accuracy of knowledge. 6. Answer display module: displays the answers or suggestions generated by the system to the user, including text and visual display.

[0063] Specifically, the corresponding process is as follows: Step 1. User inputs question; Step 2. The natural language processing module performs semantic analysis and understanding of the question; Step 3. Retrieve relevant knowledge from the knowledge graph module based on the understood problem; Step 4. Generate accurate answers or suggestions using intelligent reasoning and recommendation technology; Step 5. Display the generated answers or suggestions to the user; Step 6. If the user is not satisfied or needs further information, return to step 1 to continue.

[0064] This architecture uses a variety of technical means such as natural language processing technology, knowledge graph construction, intelligent reasoning and recommendation technology, and knowledge update and maintenance mechanism to achieve the goal of quickly and accurately responding to user questions.

[0065] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a computer, the method described in any of the above embodiments is executed.

[0066] It can be understood that computer-readable storage media may include: any entity or device capable of carrying a computer program, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc. A computer program includes computer program code. The computer program code may be in source code form, object code form, executable file, or some intermediate form, etc. A computer-readable storage medium may include: any entity or device capable of carrying a computer program code, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc.

[0067] The processor may refer to a processor included in the controller. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0068] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0069] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A knowledge base question-answering method for a ceramic enterprise system, characterized in that: include: Responding to user input questions and converting the questions into semantic vectors; Obtain a knowledge graph, and obtain the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; Configuring and / or adjusting semantic similarity weights based on preset factors; The answer is screened out according to the semantic similarity and the semantic similarity weight.

2. The knowledge base question-answering method for a ceramic enterprise system according to claim 1, characterized in that: The obtaining of the knowledge graph includes: Acquire and extract key entity knowledge from entity knowledge; identifying entity relationships within the key entity knowledge; A knowledge graph is constructed based on the key entity knowledge and the entity relationships.

3. The knowledge base question-answering method for a ceramic enterprise system according to claim 2 is characterized in that: The step of acquiring and extracting key entity knowledge from entity knowledge includes: Obtaining a ceramic enterprise data source, and extracting entity knowledge from the ceramic enterprise data source; Natural language processing technology is used to perform entity recognition on the entity knowledge to extract key entity knowledge.

4. The knowledge base question-answering method for a ceramic enterprise system according to claim 2, characterized in that: The identifying entity relationships within the key entity knowledge includes: Based on the selected dependency syntax parser, each sentence in the key entity knowledge is converted into a dependency tree structure, wherein the dependency tree structure uses entities as nodes and represents the dependency relationship between entities through edges; Get relationship anchor points based on preset scenarios; Extract relational features from dependency trees; Based on the relationship anchor points and the relationship features, entity relationships within the key entity knowledge are identified through a preset training model.

5. The knowledge base question-answering method for a ceramic enterprise system according to claim 4 is characterized in that: The obtaining the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm includes: The distance between the semantic vector and the knowledge base document in the knowledge graph is calculated based on a preset distance algorithm, and the semantic similarity is evaluated based on the distance.

6. The knowledge base question-answering method for a ceramic enterprise system according to claim 1, characterized in that: The preset factors include at least content factors, context factors and user feedback factors; wherein, Configure semantic similarity weights based on content factors, including: Configure semantic similarity weights based on the importance and usage frequency of knowledge items in key entity content; Adjust semantic similarity weights based on contextual factors, including: Adjusting the semantic similarity weight based on the context information of the user's input question in response; Adjust semantic similarity weights based on user feedback factors, including: The user evaluation and user interaction data are quantified into preset feedback values, and the semantic similarity weights are adjusted based on the preset feedback values.

7. The knowledge base question-answering method for a ceramic enterprise system according to claim 1, characterized in that: The step of selecting the answer according to the semantic similarity and the semantic similarity weight includes: Retrieving a preset number of candidate answers according to the semantic similarity and the semantic similarity weight; The candidate answers are screened based on preset optimization conditions to obtain a unique answer from the candidate answers.

8. A knowledge base question-answering system for a ceramic enterprise system, characterized in that: include: A response unit, at least used to respond to a user's input question and convert the question into a semantic vector; An acquisition unit, at least used to acquire a knowledge graph, and acquire the semantic similarity between the semantic vector and the knowledge graph based on a preset distance algorithm; A configuration unit, at least for configuring and / or adjusting the semantic similarity weight based on preset factors; The screening unit is at least used to screen out answers according to the semantic similarity and the semantic similarity weight.

9. The knowledge base question-answering system for a ceramic enterprise system according to claim 8, characterized in that: The acquisition unit includes a first acquisition unit, and the first acquisition unit is at least used for: Acquire and extract key entity knowledge from entity knowledge; identifying entity relationships within the key entity knowledge; A knowledge graph is constructed based on the key entity knowledge and the entity relationships.

10. The knowledge base question-answering system for a ceramic enterprise system according to claim 8, characterized in that: The acquisition unit includes a second acquisition unit, and the second acquisition unit is at least used for: The distance between the semantic vector and the knowledge base document in the knowledge graph is calculated based on a preset distance algorithm, and the semantic similarity is evaluated based on the distance.

11. A knowledge base question-answering architecture for ceramic systems, characterized in that: include: The data acquisition module is at least used to obtain the data source of the ceramic enterprise and transmit the data to the data processing module; The data processing module is at least used to extract entity knowledge from the ceramic enterprise data source, perform entity recognition on the entity knowledge using natural language processing technology, extract key entity knowledge, and identify entity relationships within the key entity knowledge; Constructing a knowledge graph based on the key entity knowledge and the entity relationship to achieve knowledge generation; The knowledge base management module is used to configure a graph database such as Neo4j to store the knowledge graph and respond to the knowledge retrieval of the question-answering module; The question-and-answer module is at least used to respond to the user's input questions and return answers, and to deliver the responses corresponding to the user's input questions to the question-and-answer module.

12. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is run by a computer, the knowledge base question-answering method for a ceramic enterprise system as described in any one of claims 1 to 7 is executed.

Images