A digital virtual human data architecture system and processor suitable for the electric power industry

Abstract

This invention provides a digital virtual human data architecture system and processor suitable for the power industry, belonging to the technical field of the power industry. The digital virtual human data architecture system includes: a business data layer, comprising structured, semi-structured, and unstructured raw data from both inside and outside the enterprise; an industry knowledge base, used to receive the raw data from the business data layer and process the raw data to obtain labeled data, FAQs, rule data, index data, graph data, corpus data, lexicon, and image data; a first database, storing the data processed by the industry knowledge base from the raw data; an interactive data layer, used to acquire data from the first database and search and match questions to be asked based on the data in the first database; and a digital virtual human data layer. This system can generate corresponding digital virtual humans tailored to the architectural characteristics of the power system to address power industry business needs.

Description

Technical Field

[0001] This invention relates to the field of power industry technology, and more specifically to a digital virtual human data architecture system and processor applicable to the power industry. Background Technology

[0002] Currently, digital virtual humans, as a new generation of human-computer interaction platforms, are still in their developmental stage and lack a unified, universal system framework. Based on the production technology of digital virtual humans and the structure of digital human services and products currently available on the market, a digital virtual human system generally consists of five modules: character image, voice generation, animation generation, audio-visual synthesis and display, and interaction. Many existing virtual human models are unsuitable for the complex business needs at the grassroots level of the power industry because they utilize limited data, making it impossible to directly access data from frontline operations, or allowing interaction only through limited knowledge data, resulting in a poor user experience in the power industry. Furthermore, current AI engines providing basic motion capture, driving and rendering, facial modeling, multimodal interaction, and voice customization services for digital humans are costly to customize. Customizing for each specific business operation would be labor-intensive and unable to meet the needs of grassroots users. Therefore, a digital virtual human data architecture system suitable for the power industry is needed. Summary of the Invention

[0003] The purpose of this invention is to provide a digital virtual human data architecture system and processor suitable for the power industry. This system can generate corresponding digital virtual humans based on the architectural characteristics of the power system to cope with the business of the power industry.

[0004] To achieve the above objectives, embodiments of the present invention provide a digital virtual human data architecture system suitable for the power industry, the digital virtual human data architecture system comprising: The business data layer includes structured, semi-structured, and unstructured raw data from both inside and outside the enterprise. An industry knowledge base is used to receive raw data from the business data layer and process the raw data to obtain labeled data, FAQs, rule data, index data, graph data, corpus data, lexicon, and image data. The first database stores the data processed by the industry knowledge base from the original data; The interactive data layer is used to obtain data from the first database and search and match the questions to be asked based on the data in the first database. A digital virtual human data layer, which is used to store digital human facial expressions, voice, movements and 3D asset data; The digital human portrait generation layer is used to generate digital portraits and speech based on the digital human expressions, voices, actions, and 3D asset data included in the digital virtual human data layer. The digital human data interaction layer is used for projects that generate digital human images based on user questions and data within the interaction data layer, including intelligent question answering, task dialogue, intelligent recommendation, and intelligent search.

[0005] Optionally, the raw data includes: feasibility study plan, business requirements plan, software requirements plan, preliminary design plan, security protection plan, Excel data, JSON data, business architecture assets, application architecture assets, data architecture assets, technical architecture assets, and security architecture assets.

[0006] Optionally, the first database includes a relational database and a graph database, which store relational data and graph data respectively.

[0007] Optionally, the interactive data layer acquires data from the first database and searches and matches the questions to be asked based on the data in the first database, including: Obtain data from the first database and extract features from the data in the first database; Obtain the words of the input question and perform feature extraction; The similarity between the words in the input question and the data in the first database is calculated according to formula (1): , formula (1) in, The feature vectors representing the words in the input question. This represents the feature vector of the data in the first database. express and The similarity between them; Sort the similarity calculated by formula (1) and obtain the data in the first database that are ranked in the top three; The filtered data from the first database and the words of the input question are fed into the Attention model to obtain the weights of the filtered data from the first database. Filter data and its context in the first database whose weight is greater than a preset threshold; The data and its context from the first database are fed into the GPT model to generate coherent and fluent statements relevant to the problem.

[0008] Optionally, the data selected from the first database and the words of the input question are fed into the Attention model to obtain the weights of the data selected from the first database, including: Retrieve the filtered data from the first database; Feature extraction is performed on the data in the first database; The weights of the data in the first database selected according to formula (2) are calculated as follows: , formula (2) in, The query vector represents the features of the current word. This is a key vector, representing the weights of other words in the context. It is a value vector, representing the actual information of other words in the context. Indicates weight, The dimension of the key vector. This indicates transpose.

[0009] Optionally, the 3D asset data includes the geometry, size, and positional relationships of the digital virtual human.

[0010] Optionally, the digital virtual human data architecture system includes an application layer, which is used to apply the system to scenarios such as virtual tour guides, digital human customer service, digital human training, digital human human resources, and digital human employees, according to the needs of the company's business scenarios.

[0011] On the other hand, the present invention also provides a processor for running a program, which, when running, performs the steps of a digital virtual human data architecture system applicable to the power industry as described above.

[0012] In one aspect, the present invention also provides a machine-readable storage medium storing instructions for performing the steps of a digital virtual human data architecture system applicable to the power industry as described above.

[0013] Through the above technical solution, the business data layer of the digital virtual human data architecture system and processor for the power industry provided by this invention includes structured, semi-structured, and unstructured raw data from both inside and outside the enterprise. The industry knowledge base can receive the raw data from the business data layer and process it to obtain labeled data, FAQs, rule data, index data, graph data, corpus data, lexicon, and image data. The first database can store the data processed by the industry knowledge base. The interactive data layer can acquire data from the first database and search and match questions based on the data in the first database. The digital virtual human data layer can store digital human facial expressions, voice, actions, and 3D asset data. The digital human avatar generation layer can generate digital avatars and language based on the digital human facial expressions, voice, actions, and 3D asset data included in the digital virtual human data layer. The digital human data interaction layer can generate intelligent question-and-answer, task dialogue, intelligent recommendation, and intelligent search items for digital avatars based on user questions and data within the interaction data layer. This system can generate corresponding digital virtual humans tailored to the architectural characteristics of the power system to address power industry business needs.

[0014] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a system connection diagram of a digital virtual human data architecture system applicable to the power industry according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating data search in the interactive data layer of a digital virtual human data architecture system applicable to the power industry, according to an embodiment of the present invention. Figure 3 This is a flowchart illustrating the acquisition of weights in the interactive data layer of a digital virtual human data architecture system applicable to the power industry, according to an embodiment of the present invention. Detailed Implementation

[0016] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0017] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0018] Figure 1 This is a system connection diagram of a digital virtual human data architecture system applicable to the power industry according to an embodiment of the present invention. In this invention, the digital virtual human data architecture system may include: a business data layer, an industry knowledge base, a first database, an interactive data layer, a digital virtual human data layer, a digital human image generation layer, and a digital human data interaction layer. The business data layer may include structured, semi-structured, and unstructured raw data from within and outside the enterprise. The industry knowledge base can receive the raw data from the business data layer and process it to obtain labeled data, FAQs, rule data, index data, graph data, predictive data, a thesaurus, and image data. The first database can store the data processed by the industry knowledge base from the raw data. The interactive data layer can acquire data from the first database and search and match questions based on the data in the first database. The digital virtual human data layer can store digital human facial expressions, voice, actions, and 3D asset data. The digital human image generation layer can generate digital images and voices based on the digital human facial expressions, voice, actions, and 3D asset data included in the digital virtual human data layer. This digital human data interaction layer can be used to generate digital avatars for intelligent question-and-answer, task dialogue, intelligent recommendation, and intelligent search projects based on user questions and data within the interaction layer. The system can also generate corresponding digital virtual avatars tailored to the architectural characteristics of power systems to address business needs in the power industry.

[0019] In one embodiment of the present invention, the raw data may include: feasibility study plan, business requirement plan, software requirement plan, preliminary design plan, security protection plan, Excel data, JSON data, business architecture assets, application architecture assets, data architecture assets, technical architecture assets, and security architecture assets.

[0020] In one embodiment of the present invention, the first database may include a relational database and a graph database. The relational database and the graph database may respectively store corresponding relational data and graph data. After processing the original data, it can be divided into relational data and graph data according to data type and then stored in the first database.

[0021] In one embodiment of the present invention, such as Figure 2 As shown, the data search process in the interactive data layer may include: In step S1, data from the first database is acquired and features are extracted from the data in the first database.

[0022] In step S2, the words of the input question are obtained and their features are extracted.

[0023] In step S3, the similarity between the words of the input question and the data in the first database is calculated according to formula (1): , formula (1) in, The feature vectors representing the words in the input question. This represents the feature vector of the data in the first database. express and The similarity between them.

[0024] In step S4, the similarity calculated by formula (1) is sorted and the data in the first database that ranks in the top three is obtained.

[0025] In step S5, the data from the first database and the words of the input question are fed into the Attention model to obtain the weights of the data from the first database.

[0026] In step S6, data and its context within the first database with a weight greater than a preset threshold are filtered.

[0027] In step S7, the data from the first database and its context are fed into the GPT model to generate coherent and fluent statements related to the problem.

[0028] In this invention, when performing a relevant search for a question, data from the first database can be obtained first, and features can be extracted from the data in the first database. Simultaneously, the words of the input question can also be obtained and their features extracted. After feature extraction, the data from the first database and the words of the input question can be converted into vectors. The similarity between the words of the input question and the data in the first database can be calculated using formula (1). Based on this similarity, the data in the first database most similar to the question can be determined. Therefore, the similarity calculated by formula (1) can be sorted, and the top three data from the first database can be obtained. The selected data from the first database and the words of the input question can be fed into the Attention model to obtain the weights of the selected data from the first database. Based on these weights, the relevance of the words and context related to the question can be determined. Therefore, data and context with weights greater than a preset threshold can be selected. These data and context are more relevant to the question. Therefore, the data from the first database and its context can be fed into the GPT model to generate a coherent and fluent sentence related to the question. This sentence is most relevant to the question.

[0029] In one embodiment of the present invention, such as Figure 3 As shown, the process of obtaining weights in this interactive data layer may include: In step S8, the data from the first filtered database is obtained.

[0030] In step S9, feature extraction is performed on the data in the first database.

[0031] In step S10, the weights of the data in the first database selected are calculated according to formula (2): , formula (2) in, The query vector represents the features of the current word. This is a key vector, representing the weights of other words in the context. It is a value vector, representing the actual information of other words in the context. Indicates weight, The dimension of the key vector. This indicates transpose.

[0032] In this invention, when obtaining the weights of the data in the first selected database, the data in the first selected database can be obtained first, and then the data in the first database can be feature extracted to convert it into a vector. After converting the data in the first database into a vector, the weights of the data in the first selected database can be calculated according to the formula (2).

[0033] In one embodiment of the present invention, the 3D asset data may include the geometry, size and positional relationships of the digital virtual human. Based on the geometry, size and positional relationships of the digital virtual human, the approximate shape of the digital virtual human can be constructed, and it can be further refined based on more detailed data.

[0034] In one embodiment of the present invention, the digital virtual human data architecture system may include an application layer. This application layer can be applied to scenarios such as virtual tour guides, digital human customer service representatives, digital human HR personnel, and digital human employees, according to the needs of the company's business scenarios. Different parameters can be adjusted for different scenarios to better adapt to the corresponding situations.

[0035] On the other hand, the present invention also provides a processor that can be used to run a program, which, when running, performs the steps of a digital virtual human data architecture system applicable to the power industry as described above.

[0036] In one aspect, the present invention also provides a machine-readable storage medium on which instructions can be stored. These instructions can be used to perform the steps of a digital virtual human data architecture system applicable to the power industry as described above.

[0037] Through the above technical solution, the business data layer of the digital virtual human data architecture system and processor for the power industry provided by this invention includes structured, semi-structured, and unstructured raw data from both inside and outside the enterprise. The industry knowledge base can receive the raw data from the business data layer and process it to obtain labeled data, FAQs, rule data, index data, graph data, corpus data, lexicon, and image data. The first database can store the processed data from the industry knowledge base. The interactive data layer can acquire data from the first database and search and match questions based on the data in the first database. The digital virtual human data layer can store digital human facial expressions, voice, actions, and 3D asset data. The digital human avatar generation layer can generate digital avatars and language based on the digital human facial expressions, voice, actions, and 3D asset data included in the digital virtual human data layer. The digital human data interaction layer can generate intelligent question-and-answer, task dialogue, intelligent recommendation, and intelligent search items for the digital avatar based on user questions and data within the interaction data layer. This system can generate corresponding digital virtual humans tailored to the architectural characteristics of the power system to address power industry business needs.

[0038] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0039] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0040] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0041] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0042] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0043] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0044] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0045] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0046] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A digital virtual human data architecture system suitable for the electric power industry, characterized in that, The digital virtual human data architecture system includes: The business data layer includes structured, semi-structured, and unstructured raw data from both inside and outside the enterprise. An industry knowledge base is used to receive raw data from the business data layer and process the raw data to obtain labeled data, FAQs, rule data, index data, graph data, corpus data, lexicon, and image data. The first database stores the data processed by the industry knowledge base from the original data; The interactive data layer is used to obtain data from the first database and search and match the questions to be asked based on the data in the first database. A digital virtual human data layer, which is used to store digital human facial expressions, voice, movements and 3D asset data; The digital human portrait generation layer is used to generate digital portraits and speech based on the digital human expressions, voices, actions, and 3D asset data included in the digital virtual human data layer. The digital human data interaction layer is used for projects that generate digital human images based on user questions and data within the interaction data layer, including intelligent question answering, task dialogue, intelligent recommendation, and intelligent search.

2. The digital virtual human data architecture system of claim 1, wherein, The raw data includes: feasibility study plan, business requirements plan, software requirements plan, preliminary design plan, security protection plan, Excel data, JSON data, business architecture assets, application architecture assets, data architecture assets, technical architecture assets, and security architecture assets.

3. The digital virtual human data architecture system of claim 1, wherein, The first database includes a relational database and a graph database, which store relational data and graph data respectively.

4. The digital virtual human data architecture system of claim 1, wherein, The interactive data layer acquires data from the first database and searches and matches the questions to be asked based on the data in the first database, including: Obtain data from the first database and extract features from the data in the first database; Obtain the words of the input question and perform feature extraction; The similarity between the words in the input question and the data in the first database is calculated according to formula (1): Equation (1) wherein, represents a feature vector of a word inputted as a question, represents a feature vector of data within the first database, represents and a similarity between them; Sort the similarity calculated by formula (1) and obtain the data in the first database that are ranked in the top three; The filtered data from the first database and the words of the input question are fed into the Attention model to obtain the weights of the filtered data from the first database. Filter data and its context in the first database whose weight is greater than a preset threshold; The data and its context from the first database are fed into the GPT model to generate coherent and fluent statements relevant to the problem.

5. The digital virtual human data architecture system of claim 4, wherein, The data selected from the first database and the words in the input question are fed into the Attention model to obtain the weights of the data selected from the first database, including: Retrieve the filtered data from the first database; Feature extraction is performed on the data in the first database; The weights of the data in the first database selected according to formula (2) are calculated as follows: Equation (2) wherein, is a query vector, representing the features of the current word, is a key vector, representing the weights of other words in the context, is a value vector, representing the actual information of other words in the context, denotes a weight, denotes the dimension of the key vector, denotes the transpose.

6. The digital virtual human data architecture system of claim 1, wherein, The 3D asset data includes the geometry, size, and positional relationships of the digital virtual human.

7. The digital virtual human data architecture system of claim 1, wherein, The digital virtual human data architecture system comprises an application layer, which is used to apply the system in virtual tour guide, digital human customer service, digital human training, digital human personnel and digital human worker scenes according to the needs of company business scenes.

8. A processor, comprising: A program for running, which is used to execute the steps of the digital virtual human data architecture system for the power industry according to any one of claims 1-7 when the program is running.

9. A machine-readable storage medium, characterized in that, The machine readable storage medium has instructions stored thereon, which are used to execute the steps of the digital virtual human data architecture system for the power industry according to any one of claims 1-7.

Images