AI robot intelligent question and answer method, device and equipment and storage medium

Abstract

The embodiment of the invention provides an AI robot intelligent question-answering method, device and equipment and a storage medium, and is applied to the field of financial science and technology. According to the method, after question data input by a question and answer interaction interface based on a first business line is obtained, card associated information can be extracted from the question data. And downloading the card component in the public component library based on the card association information, and executing dynamic component rendering on the card component to generate answer data. Wherein the card associated information comprises at least one of general information and cross-service information. According to the method, the card component based on the general research and development specification can be downloaded in the public component library by extracting the card association information, so that the general card component and the card components fused with other business lines are quickly loaded, and efficient fusion of the AI robot question and answer processes in multiple business fields is realized; and the cross-business question-answering capability of the financial scene AI robot is improved.

Description

Technical Field

[0001] This application relates to the technical field of AI robots in financial scenarios, and particularly to an AI robot intelligent question-answering method, device, equipment, and storage medium. Background Art

[0002] An artificial intelligence (AI) robot is an intelligent interaction system developed based on artificial intelligence technology. An AI robot can understand various questions raised by users through natural language processing (NLP) technology, and generate accurate and relevant answer information according to a preset knowledge base, data model, or real-time search, etc. An AI robot can simulate the way of human conversation and provide users with convenient and efficient information query and consultation services. For example, in the fintech scenario, an AI robot can serve as a window for fintech enterprises to connect with users and provide services such as financial-related business consultation and business guidance.

[0003] Since the types of questions asked by users based on AI robots are diverse, the answers provided by AI robots also need to be targeted. A simple human-machine interaction interface can hide thousands of reply contents. Therefore, when AI robots are required for external communication with users in multiple business channels or product lines of a large group, according to the product R & D mode, the development workload of AI robots is extremely large. Each product line requires corresponding product design, AI and algorithm R & D, back-end development, front-end development, and so on.

[0004] Multiple business lines or product lines within a large group intersect with each other and are not isolated. For example, in the fintech business, it is possible that the investment consulting window needs to feedback product information of electronic payment. For this reason, the investment consulting project team also needs to develop a set of system feedback contents for electronic payment, and the electronic payment business needs to be developed repeatedly, resulting in a waste of various input costs of the group. For example, two projects are independently developed by two teams respectively. However, when an AI robot wants to access other projects for front-end interface integration later, due to the independent development of the two projects, efficient integration cannot be carried out, and only by rewriting a set of front-end page logics can integration be carried out. Summary of the Invention

[0005] In view of this, the embodiments of this application provide an AI robot intelligent question-answering method, device, equipment, and storage medium to solve the problem that AI robots developed in multiple projects cannot be efficiently integrated.

[0006] According to one aspect of this application, an AI robot intelligent question-answering method is provided. The method includes:

[0007] Obtain question data, where the question data is data input based on the Q&A interaction interface of the first business line;

[0008] Extract card association information from the question data, where the card association information includes at least one of general information and cross-business information;

[0009] Download card components from the public component library based on the card association information, where the card components are public components developed based on general R & D specifications;

[0010] Perform dynamic component rendering on the card components to generate answer data.

[0011] According to another aspect of the present application, there is provided an AI robot intelligent Q&A device, where the device includes:

[0012] A question module for obtaining question data, where the question data is data input based on the Q&A interaction interface of the first business line;

[0013] An information extraction module for extracting card association information from the question data, where the card association information includes at least one of general information and cross-business information;

[0014] A component download module for downloading card components from the public component library based on the card association information, where the card components are public components developed based on general R & D specifications;

[0015] An answer module for performing dynamic component rendering on the card components to generate answer data.

[0016] According to yet another aspect of the present application, there is provided a computer device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, where when the processor executes the program, the above-mentioned AI robot intelligent Q&A method is implemented.

[0017] According to still another aspect of the present application, there is provided a storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned AI robot intelligent Q&A method is implemented.

[0018] With the above technical solution, the embodiments of the present application provide an AI robot intelligent question-answering method, device, equipment and storage medium. After obtaining the question data input through the question-answering interaction interface based on the first business line, the method can extract card association information from the question data. Then, based on the card association information, card components are downloaded from the common component library, and dynamic component rendering is performed on the card components to generate answer data. Among them, the card association information includes at least one of general information and cross-business information. The method can download card components based on general R & D specifications from the common component library by extracting card association information, so as to quickly load general card components and integrate card components of other business lines, realize the efficient integration of the AI robot question-answering process in multiple business fields, and improve the cross-business question-answering ability of the AI robot in the financial scenario.

[0019] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the specific embodiments of the present application are hereinafter specifically exemplified. Brief Description of the Drawings

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0021] Figure 1 It is a schematic flowchart of the intelligent question-answering process provided by the embodiments of the present application;

[0022] Figure 2 It is a schematic diagram of the display effect of the answer data provided by the embodiments of the present application;

[0023] Figure 3 It is a schematic flowchart of the AI robot intelligent question-answering method provided by the embodiments of the present application;

[0024] Figure 4 It is a schematic flowchart of the process of extracting card association information provided by the embodiments of the present application;

[0025] Figure 5 It is a schematic flowchart of the process of downloading card components provided by the embodiments of the present application;

[0026] Figure 6 It is a schematic flowchart of the process of developing card components provided by the embodiments of the present application;

[0027] Figure 7 It is a schematic structural diagram of the AI robot intelligent question-answering device provided by the embodiments of the present application;

[0028] Figure 8Schematic diagram of the computer device provided by the embodiment of the present application. Detailed implementation manners

[0029] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other.

[0030] In the embodiment of the present application, an artificial intelligence (AI) robot refers to an AI question-and-answer robot, which is an intelligent interaction system developed based on artificial intelligence technology. The AI robot has generality and can not only refer to a mechanical device with a humanoid appearance structure, but also refer to an application program with a question-and-answer function and a combination of an application program and a hardware device. For example, the AI robot is an intelligent assistant installed on electronic devices such as a computer, a server, a mobile terminal, and a control host.

[0031] The AI robot can understand various questions raised by users through natural language processing (NLP) technology, and generate accurate and relevant answer information according to a preset knowledge base, data model, or real-time search, etc. Therefore, in the embodiment of the present application, the process in which the AI robot generates and displays answer information for the questions raised by users is called intelligent question and answer.

[0032] As Figure 1 shown, the intelligent question-and-answer process may involve text understanding and text generation based on natural language processing. Text understanding is the core of the AI robot, which can parse the natural language text input by the user and understand the intention, keywords, and semantic structure therein. By analyzing and interpreting the semantics, grammar, context, etc. of the text content, the meaning of the text can be understood. Text understanding can use a text understanding model, which is an artificial intelligence model obtained by training with sample data with semantic labels. Such as BERT, OpenNLU, etc.

[0033] In some embodiments, when performing text understanding, tokenization may be performed first to split a continuous text string into meaningful units, such as words, phrases, or symbols, etc. Then, part-of-speech tagging is performed on the tokenization result, that is, to determine the part of speech of each word in the sentence, such as a noun, a verb, an adjective, etc. Then, semantic role labeling is performed to identify the semantic roles of each component in the sentence, such as an agent, a patient, a tool, etc.

[0034] After preprocessing such as word segmentation, stemming, part-of-speech tagging, and named entity recognition, syntax analysis can be performed to analyze the grammatical structure of the sentence and construct a syntax tree to determine the grammatical relationships between words, such as subject, predicate, object, etc. And based on word meaning, syntactic structure, and semantic roles, the overall meaning of the text is inferred. And based on the pragmatic analysis algorithm, the actual intention and effect of the text in a specific context are determined.

[0035] After parsing natural language text and understanding the user's question, the AI robot can generate natural and fluent answers. The text generation process requires the use of language generation models, such as language generation models based on deep learning architectures like Transformer. The language generation model can learn a large amount of language data to generate answers that conform to grammatical rules and semantic logic.

[0036] In some embodiments, when performing text generation, the input data can be prepared according to the task requirements first, and a suitable generation strategy can be selected, such as greedy decoding, beam search, random sampling, etc. Then, based on the input data and the generation strategy, the model generates the target text. Through text understanding and text generation, the AI robot can simulate the human conversation mode to provide users with convenient and efficient information query and consultation services.

[0037] The AI robot can serve as a window for the enterprise to connect with users and provide a more convenient communication method. In some embodiments, in order to use the AI robot as a window for the enterprise to connect with users, after clarifying the main functions of the AI robot and analyzing the user portraits of the target user group, a multi-layer architecture including a user interface layer, a dialogue management layer, and a background processing layer can be adopted to construct a user connection window application. Among them, the user interface layer is responsible for interacting with users, the dialogue management layer is used to process dialogue logic, and the background processing layer integrates the enterprise internal system.

[0038] Based on the constructed connection window application, technology selection and model training are also required. That is, by selecting suitable NLP technologies, including word segmentation, semantic understanding, etc., to accurately parse user input. Then, use the large language model for training and fine-tuning to improve language understanding and generation capabilities. And construct a domain knowledge graph to store business knowledge and common question answers to quickly locate user questions. Then, by integrating materials such as historical conversation records and product manuals, a structured knowledge base is established, and the knowledge base is automatically updated through AI technology to ensure the timeliness of information. Among them, the knowledge base can be a structured database or a large amount of text data.

[0039] After obtaining the language processing model of the AI robot through model training, an intelligent question-answering system integrating the docking window application and the language processing model can be obtained. Then, the intelligent question-answering system is integrated and deployed according to the specific docking window requirements. For example, the AI robot can be integrated with backend systems such as the enterprise's Customer Relationship Management (CRM) and order management system through an Application Programming Interface (API). It can also be deployed through containerization, that is, using technologies such as Docker and Kubernetes to deploy the service to a cloud server to ensure the high availability and scalability of the system.

[0040] After the intelligent question-answering system is integrated and deployed, users can input question data through the terminal device after docking the application to the terminal device through multiple channels such as web pages, applications, and social media. The intelligent question-answering system can then receive the question data input by the user in real time. After receiving the question data input by the user, the question data is input into the language understanding model to understand the overall meaning of the question data. Then, relevant information is retrieved from the knowledge base based on the language understanding result. For example, a retrieval-based question-answering system will match the user's question with the questions in the knowledge base through semantic search technology to find the most relevant answer. After retrieving the relevant information, the intelligent question-answering system needs to generate an answer text. The intelligent question-answering system can generate an answer text based on the retrieved relevant information, that is, it can directly select the most matching answer data from the knowledge base. The intelligent question-answering system can also obtain the answer text based on the generated text data, that is, use a deep learning model to generate the answer data.

[0041] After obtaining the answer data, the intelligent question-answering system can display the answer data through the question-and-answer interaction interface. Among them, the question-and-answer interaction interface is the front-end human-computer interaction interface of the AI robot intelligent question-answering system. As the layer closest to the user in the entire intelligent question-answering system, it can interact through the dialogue communication of intelligent assistants and question-and-answer web pages. The display method of the answer data can include various forms. For example, as Figure 2 shown, the answer data can be presented in the form of ordinary text, pictures, and can also be presented in various ways such as data tables, data charts (such as pie charts, bar charts, line charts, etc.), various links and buttons.

[0042] In some embodiments, the question-and-answer interaction interface can present the answer data by combining card components. Among them, the card component is a functional component generated by unifying and modularizing display carriers such as ordinary text, pictures, tables, charts, links, buttons, etc. Therefore, what the intelligent question-and-answer system communicates with the user on the front-end question-and-answer interaction interface are multiple card components in different forms, that is, various answers are formed by combining multiple card components.

[0043] Since the types of questions asked by users based on AI robots are diverse, the answers given by AI robots also need to be targeted. A simple human-computer interaction interface can hide thousands of reply contents. Therefore, when AI robots are required for external windows for user communication in multiple business channels or product lines of a large group, according to the product R & D model, the development workload of AI robots is extremely large. Each product line requires corresponding product design, AI and algorithm R & D, back-end development, front-end development, etc.

[0044] Multiple business lines or product lines of a large group intersect with each other and are not isolated. For example, in the case of fintech business, it is possible that the investment consulting window needs to feedback product information of electronic payment. For this reason, the investment consulting project team also needs to develop a set of system feedback content for electronic payment, and the electronic payment needs to be developed repeatedly, resulting in a waste of various input costs of the group. For example, two projects are developed independently by two teams respectively. However, when the subsequent AI robot wants to access other projects for front-end interface integration, due to the independent development of the two projects, efficient integration cannot be carried out, and only by rewriting a set of front-end page logic can integration be carried out.

[0045] To solve the problem that AI robots developed in multiple projects cannot be efficiently integrated, some embodiments of the present application provide an AI robot intelligent question-and-answer method. The method can be applied to an intelligent question-and-answer system and is specifically applied to an electronic device deployed with the intelligent question-and-answer system. For example, in a financial scenario, the electronic device can be a question-and-answer terminal set in the service hall. For the sake of convenience of description, in the embodiments of the present application, the intelligent question-and-answer system is used as the execution subject of the AI robot intelligent question-and-answer method. It should be understood that the method can also be applied to other types of execution subjects, such as computers, servers, cloud computing network applications, mobile terminals, etc., which are not shown one by one in the embodiments of the present application. As Figure 3 shown, the method includes:

[0046] S101. Obtain question data.

[0047] To implement the intelligent question-and-answer function, the intelligent question-and-answer system can obtain question data, which is the data input based on the question-and-answer interaction interface of the first business line. Herein, the first business line is a business line in the group business architecture that can operate and maintain the intelligent question-and-answer system, and can be specifically divided according to the group's business projects. For example, for the property insurance business of PA Group, it may include the AI operation assistant (branch housekeeper) project and the small and micro assistant project. The intelligent question-and-answer systems of these two projects can be independently developed by two teams respectively. Then the first business line can be the business line of the AI operation assistant project, and the business line of the small and micro assistant project is the second business line.

[0048] After the user runs the application of the intelligent question-and-answer system or accesses the web page of the intelligent question-and-answer system through the terminal device, the user can control the terminal device to display the question-and-answer interaction interface of the first business line. The question-and-answer interaction interface may include at least one text input control. The user can input question data based on the text input control. For example, after the user clicks the AI operation assistant application icon, the user can control the terminal device to run the AI operation assistant application and display the corresponding question-and-answer interaction interface. The corresponding question-and-answer interaction interface of the AI operation assistant may include a text input box, and the user can click the text input box and input question data.

[0049] It should be noted that the question data can be obtained from the text directly input by the user, or can be obtained by converting other forms of signals input by the user into text. That is, in some embodiments, the intelligent question-and-answer system can obtain the question signal input by the user based on the question-and-answer interaction interface of the first business line, and perform signal preprocessing such as amplification, noise reduction, filtering, and effective signal extraction on the question signal to obtain an effective input signal.

[0050] Then perform feature extraction on the effective input signal to convert the effective input signal into a feature representation that can be processed by the machine. After obtaining the feature representation of the signal, the intelligent question-and-answer system can, based on the feature recognition model, convert the features into the probability distribution of data units, and thus determine the data unit whose probability distribution meets the preset distribution threshold requirement as the transcription result. Then, according to the statistical rules of the language, the language model provides a prior probability score for the candidate transcription results to obtain a word sequence that conforms to the language habit. Finally, by synthesizing the probabilities of the feature recognition model and the language model, find the most likely recognition result among all possible text sequences, that is, obtain the question data in text format.

[0051] For example, the question-and-answer interaction interface of the first business line may include a voice input control. After the user clicks the voice input control and inputs a voice signal, the intelligent question-and-answer system can perform noise reduction and silence detection on the voice signal, remove environmental noise and silent segments, improve the quality of the voice signal, and can also perform pre-emphasis filtering to enhance high-frequency signals and reduce the loss of high-frequency signals during transmission.

[0052] Then, through feature extraction methods such as Mel Frequency Cepstral Coefficients (MFCC) and spectrogram, the speech signal is converted into a feature representation that can be processed by a machine. And an acoustic model is used to convert the features into a probability distribution of speech units, where the acoustic model is one or a combination of models such as Hidden Markov Model (HMM), Deep Neural Network (DNN), Recurrent Neural Network (RNN), and Transformer model.

[0053] By calling language models such as n-gram model and neural network language model, and using the language model to provide a prior probability score for the candidate transcription results according to the statistical laws of the language, to prefer word sequences that are more in line with language habits. Then, based on decoding algorithms such as the Viterbi algorithm and beam search algorithm built into the decoder, the probabilities of the acoustic model and the language model are combined to find the most likely recognition result among all possible text sequences, so as to convert the speech signal into question data in text format.

[0054] Since the question content input by the user is data in natural language form, and data in natural language form is difficult to be directly recognized and processed by a machine. Therefore, in some embodiments, in order to facilitate subsequent processing, data processing can also be performed on the data input by the user to convert the data in natural language form into structured data. That is, the intelligent question answering system can, after receiving the text data input by the user, first perform text cleaning on the text data to remove the noise in the text, such as programming language tags, special symbols, extra spaces, etc. Then, through word segmentation processing, the input text is segmented into words and phrases, and the word segmentation results are converted into a unified form.

[0055] For the word segmentation results, the intelligent question answering system can also perform information extraction, that is, the intelligent question answering system can extract keywords from the word segmentation results based on information extraction methods such as rules, statistics, and deep learning to obtain a keyword set. For example, when the intelligent question answering system extracts keywords based on the rule-based information extraction method, it can use regular expressions to match specific patterns and extract keywords that meet the regular expressions, such as phone numbers, dates, email addresses, etc.

[0056] After information extraction, the intelligent question answering system also constructs structured data according to the keyword set obtained from information extraction. For this purpose, the intelligent question answering system can perform entity recognition and linking, identify the entities in the text, and link them to the corresponding entities in the knowledge base. Then perform relation extraction and event extraction, where relation extraction aims to identify the relationships between entities. Event extraction is used to identify the events in the text and their participants, time, location and other elements. Then organize the extracted information into a structured format, such as JSON, XML or database table, so as to convert the text data in natural language form into structured text for machine recognition.

[0057] S102. Extract card association information from the said question data.

[0058] After obtaining the question data, the intelligent Q&A system can extract card association information from the question data, and the card association information includes at least one of general information and cross-business information. Among them, the general information is data having an association relationship with the general card component. For example, when the first business line is the branch steward project and the group business is the insurance business, when keywords such as "insurance" and "branch" are included in the question data input by the user, the keyword "insurance" related to the group's general business can be determined as the general information, and the keyword "branch" related to the branch steward project can be determined as the current business information.

[0059] The cross-business information is data having an association relationship with the card components of other business lines other than the first business line. In the embodiments of the present application, the second business line is used as an example to refer to other business lines other than the first business line. It should be understood that the second business line generally refers to a business line different from the first business line, and does not specifically refer to a certain specific business line. For example, when the first business line is the branch steward project and the second business line is the small and micro assistant project, when keywords such as "loan" are included in the question data input by the user based on the intelligent Q&A system of the branch steward project, the keyword "loan" related to the small and micro assistant project can be determined as the cross-business information.

[0060] Obviously, the said cross-business information is relative, that is, when judging whether the question data contains cross-business information, it is necessary to determine the input channel of the question data. When the user inputs question data related to the second business line based on the Q&A interaction interface of the first business line, it can be determined that the question data includes cross-business information. Similarly, when the user inputs question data related to the first business line based on the Q&A interaction interface of the second business line, it can also be determined that the question data includes cross-business information. When the user inputs question data related to the first business line or question data related to the group's general business based on the Q&A interaction interface of the first business line, it can be determined that the question data does not contain cross-business information.

[0061] In some embodiments, in order to extract cross-business information, when the intelligent question-answering system extracts card association information from the question data, it may first extract at least one keyword from the question data, then extract alternative keywords from at least one of the keywords according to the part of speech, and respectively obtain the associated word libraries of the first business line and the second business line. Then, based on the associated word libraries, mark the first keyword and the second keyword from the alternative keywords, and generate the cross-business information according to the second keyword. Wherein, the first keyword is a keyword included in the associated word library corresponding to the first business line; the second keyword is a keyword included in the associated word library corresponding to the second business line.

[0062] For example, as Figure 4 shown, after the user inputs the question data with the content "What is the current loan interest rate" in the intelligent assistant built in the Network Point Butler (the first business line) app, the intelligent question-answering system corresponding to the intelligent assistant can perform word segmentation on the question data to obtain the keyword set "current / loan / interest rate / is / what". And according to the part of speech of the keywords, extract the noun keywords from the keyword set and form an alternative keyword set, that is, "current / loan / interest rate". According to the pre-set associated word library WW1 corresponding to the Network Point Butler (the first business line) project and the keyword library WW2 corresponding to the Small Micro Assistant (the second business line) project, and based on the associated word libraries WW1 and WW2 for keyword matching, determine the associated word libraries to which the keywords in the alternative keyword set belong respectively. That is, the alternative keywords "loan" and "interest rate" belong to the associated word library WW2, then determine "loan" and "interest rate" as the second keywords, and generate cross-business information according to the second keywords.

[0063] In some implementations, in order to extract general information, when the intelligent question-answering system extracts card association information from the question data, it may obtain the general word library corresponding to the group business architecture, and then, based on the general word library, mark the third keyword in the alternative keywords, and generate the general information according to the third keyword. Wherein, the third keyword is a keyword included in the general word library.

[0064] Since the general word library can be shared by multiple business lines, the keywords included in the general word library are also included in the associated word library of the first business line and the associated word library of the second business line at the same time. Then, when marking the third keyword in the alternative keywords, it may first determine the keywords included in the associated word library of the first business line, and then determine whether the keywords included in the associated word library of the first business line are also included in the associated word library of the second business line at the same time. If a keyword is included in the associated word libraries corresponding to two or more business lines, then the keyword can be marked as the third keyword. That is, the general word library is the intersection of the associated word libraries corresponding to multiple business lines.

[0065] Still taking the question data where the user inputs the content "What is the current loan interest rate" through the corresponding Q&A interaction interface of the branch network steward as an example. Among the alternative keyword set "current / loan / interest rate", the keyword "current" is included in both the associated word library of the branch network steward and the associated word library corresponding to the small and micro assistant. Therefore, the keyword "current" can be marked as the third keyword, and general information can be generated based on the third keyword.

[0066] S103. Download card components from the public component library based on the card association information.

[0067] After extracting the card association information, the intelligent Q&A system can download card components from the public component library based on the card association information. Among them, the card components are public components developed based on the general R & D specification. The general R & D specification is jointly customized by the first business line and the second business line according to the group business architecture.

[0068] As Figure 5 shown, in some embodiments, in order to download card components, when the intelligent Q&A system downloads card components from the public component library based on the card association information, it can generate a component download request according to the card association information. Among them, the component download request is used to control the intelligent Q&A system of the current business line to download card components from the public component library. The component download request may include card identification information, and the card identification information may include at least one of card name, card ID, card component package download path, and card component input parameters.

[0069] Therefore, when the intelligent Q&A system generates a component download request according to the card association information, it can first obtain the specification file corresponding to the general R & D specification, and read the template protocol for defining the card identification information from the specification file. Then, based on the template protocol, query the card identification information corresponding to the card association information, and encapsulate the card identification information into the request instruction to generate a component download request.

[0070] For example, after extracting the card association information "loan / interest rate" from the question data, the specification document corresponding to the general R & D specification can be obtained first, and the template agreement can be read from the specification document. Since the template agreement contains content such as the card name, card ID (uniqueness), card component package download path, input parameters required by the card components, etc. to uniquely identify the front-end card components. Therefore, the content in the template agreement that identifies the uniqueness of the card components for the card association information "loan / interest rate" can be compared respectively. That is, according to "loan / interest rate", the card components with the card name "loan" and the card name "interest rate" can be queried, and at least one of the card identification information of the corresponding card components, such as the card name, card ID (uniqueness), card component package download path, input parameters required by the card components, etc., can be obtained and encapsulated into the request instruction to generate a request for downloading the card components with the card names "loan" and "example".

[0071] After generating the component download request, the intelligent Q&A system can send the component download request to the public component library, so that the public component library queries the component download link of the card component in response to the component download request and sends the component download link of the card component to the intelligent Q&A system. The intelligent Q&A system then downloads the card component by accessing the component download link after receiving the component download link sent by the public component library.

[0072] S104. Perform dynamic component rendering on the card component to generate answer data.

[0073] After downloading the card component, the intelligent Q&A system can perform dynamic component rendering to generate answer data and present it in the Q&A interaction interface corresponding to the first business line. In some embodiments, in order to generate answer data, when the intelligent Q&A system performs dynamic component rendering on the card component, it can first call the parsing algorithm process based on the template agreement in the general R & D specification and use the parsing algorithm process to parse the card component to obtain the data to be rendered. Then, obtain the script library to which the card component belongs, and the dynamic component rendering method information corresponding to the script library. Thus, according to the dynamic component rendering method information, the data to be rendered is rendered into an answer card component to generate the answer data.

[0074] For example, when the AI robot intelligent Q&A system of the network point steward business line needs the UI card components common to the group, or when the AI robot intelligent system of the network point steward business line needs to reuse the business logic of the small and micro assistant business line to answer users, the common UI card components or business components can be directly downloaded and installed into the front-end project according to the required card ID, name, and installation path. Then, the front-end project of the network point steward business line uses the corresponding parsing algorithm to parse the installed components according to the fixed rules and template protocols agreed upon in the common R & D rules, and uses rendering methods such as dynamic component rendering with the vue or react framework to render the reply card for the robot to reply. Moreover, when the user clicks on the corresponding card information, they can interact and jump to the small and micro assistant business system in the network point steward business system.

[0075] By applying the technical solutions of the above embodiments, after obtaining the question data input from the Q&A interaction interface based on the first business line, the AI robot intelligent Q&A method provided by the above embodiments can extract card association information from the question data. Then, based on the card association information, download card components from the common component library and perform dynamic component rendering on the card components to generate answer data. Among them, the card association information includes at least one of general information and cross-business information. The method can download card components based on the common R & D specifications from the common component library by extracting card association information, so as to quickly load common card components and integrate card components of other business lines, realizing the efficient integration of the AI robot Q&A process in multiple business fields and improving the cross-business Q&A ability of the AI robot in the financial scenario.

[0076] In some embodiments, as a refinement and extension of the specific implementation manner of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, some embodiments of the present application also provide an AI robot intelligent Q&A method, as Figure 6 shown, this method includes:

[0077] S201. Obtain the component requirement information uploaded by the first business line and the second business line;

[0078] S202. Formulate the common R & D specifications based on the component requirement information and the business architecture;

[0079] S203. Obtain the to-be-tested components constructed by different business lines according to the common R & D specifications;

[0080] S204. Perform development tests on the to-be-tested components, and after the to-be-tested components pass the development tests, package the common components and business components into public components;

[0081] S205. Store the public components in the public component library.

[0082] Due to the R & D architecture according to business projects, the front - end interfaces of each project line and project entity are developed independently. There is no problem in the scenario where there is no commonality between the two project interfaces and no cross - cooperation in business. However, within a group, often among multiple business lines or multiple projects, there is commonality in the front - end interface and the business often involves cross - cooperation, and they are not completely isolated. In this scenario, a general R & D specification can be formulated according to the group's business architecture. After each business line develops card components based on the general R & D specification, they are uniformly stored in a common component library so that the card components developed by multiple business lines can be mutually used. The general UI components and business components can avoid duplicate development across projects and improve efficiency.

[0083] Therefore, the intelligent Q&A system can obtain the component requirement information uploaded by the first business line and the second business line, and formulate a general R & D specification based on the component requirement information and the business architecture. Then, the formulated general R & D specification is sent to the terminal devices corresponding to the first business line and the second business line so that the first business line and the second business line can build card components based on the general specification.

[0084] After sending the formulated general R & D specification to the first business line and the second business line, the intelligent Q&A system can also obtain the to - be - tested components constructed by different business lines according to the general R & D specification. Among them, the to - be - tested components include general components and business components. General components are components that have nothing to do with a specific business line; business components are components related to a specific business line. Then, development tests are performed on the to - be - tested components, and after the to - be - tested components pass the development tests, the general components and business components are packaged into public components and stored in the public component library.

[0085] In order to achieve efficient co - construction of multiple projects at the front - end level of the intelligent Q&A system, the front - end R & D entities of multiple business channels or product lines belonging to the same group can jointly formulate a set of general front - end component R & D specifications, that is, formulate some fixed rules and template protocols. Among them, the template protocol can include card name, card ID (uniqueness), card component package download path, input parameters required by the card component, etc., which are used to uniquely identify the front - end card component.

[0086] After formulating the general R & D specification, each R & D entity can develop card components according to the formulated R & D specification. Among them, the card components can include general components for forming the front - end UI of each business line and business components for representing business content. General components are group - general components, so general components can have nothing to do with a specific business line and can be jointly used by each business line. Business components are related to specific businesses and each business line needs to use them interspersed.

[0087] After each R & D entity develops a card component according to the established R & D specifications, it can either perform component testing on its own or upload the developed card component to the intelligent Q&A system for component testing. Since the developed card component has not been tested yet, and component testing is required before the card component is applied, the developed card component is also called a component to be tested.

[0088] When performing development testing on the component to be tested, the intelligent Q&A system can first design test cases. For example, by drawing a business process flowchart, it can clarify the input, output, and processing logic of each business component. And by using methods such as scenario method and path method, according to the importance, usage frequency, and functional requirements of the business process, test cases are designed. It is also possible to set priorities for the test cases and give priority to testing key business processes. For the designed test cases, component testing is performed on the developed component to be tested. Component testing can include functional testing, integration testing, performance testing, security testing, etc. Functional testing is used to verify whether the functions of each business component meet the requirements, including positive testing and negative testing. Integration testing is used to test the interfaces and interactions between business components to ensure that the components can work together correctly. Performance testing is used to evaluate the response time and processing capacity of the system to ensure that the system can still run stably under high concurrency. Security testing is used to check the security of the system, including data encryption, user authentication, etc.

[0089] After the card component passes the development testing, it can be packaged into a common component and uploaded to the common repository deployed by the group. For example, the card component that has passed the development testing can be packaged into a common component and uploaded to the group's common Node Package Manager (NPM) repository. Since NPM is the default package manager for JavaScript and also the standard package management tool for Node.js. Therefore, the NPM repository is an online repository used to host open-source Node.js packages. Through the NPM repository, powerful package management and dependency management functions can be provided for each business line, simplifying the development process of JavaScript and Node.js projects.

[0090] By applying the technical solutions of the above embodiments, based on the AI robot intelligent Q&A method provided in the above embodiments, it is possible to achieve development collaboration for the front-end solutions of multiple business lines. For example, in the case of integrating the small and micro assistant project into the property insurance AI operation assistant project, there is no need for repeated development, which can save various R & D investments of the group, and can also better strengthen the win-win cooperation between each business line, enable customer drainage between multiple projects, and increase the exposure of product information.

[0091] In some embodiments, as a specific implementation of the AI robot intelligent question-and-answer method described in the above embodiments, some embodiments of the present application further provide an AI robot intelligent question-and-answer device, as follows Figure 7 As shown, the device includes:

[0092] A question module, configured to obtain question data, where the question data is data input based on a question-and-answer interaction interface of a first business line;

[0093] An information extraction module, configured to extract card association information from the question data, where the card association information includes at least one of general information and cross-business information;

[0094] A component download module, configured to download card components in a public component library based on the card association information, where the card components are public components developed based on general R & D specifications;

[0095] An answer module, configured to perform dynamic component rendering on the card components to generate answer data.

[0096] It should be noted that for other corresponding descriptions of the various functional units involved in the AI robot intelligent question-and-answer device provided in the embodiments of the present application, reference may be made to the corresponding descriptions in the AI robot intelligent question-and-answer method provided in the above embodiments, and details are not described herein again.

[0097] As Figure 8 shown, embodiments of the present application further provide a computer device, specifically a personal computer, a server, a network device, etc. The computer device includes a bus, a processor, a memory, and a communication interface, and may further include an input / output interface and a display device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in the method embodiments are implemented.

[0098] Those skilled in the art can understand that the structure of the above computer device is only a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components, or combine some components, or have different component arrangements.

[0099] In one embodiment, a computer-readable storage medium is further provided. The computer-readable storage medium may be non-volatile or volatile, and a computer program is stored thereon. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0100] In one embodiment, a computer program product is further provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0101] Those of ordinary skill in the art can understand that all or part of the processes in the above method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it may include the processes of the above method embodiments.

[0102] Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application may include at least one of non-volatile and volatile memories. Non-volatile memories may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc.

[0103] Volatile memories may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0104] The databases involved in the embodiments provided in the present application may include at least one of relational databases and non-relational databases. Non-relational databases may include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application may be general-purpose processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0105] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0106] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. An AI robot intelligent question answering method, characterized in that: The method comprises: Acquire question data, where the question data is data input based on a question-and-answer interaction interface of the first business line; Extracting card-related information from the question data, wherein the card-related information includes at least one of general information and cross-business information; Downloading a card component from a public component library based on the card association information, wherein the card component is a public component developed based on a general research and development specification; Dynamic component rendering is performed on the card component to generate answer data.

2. The method according to claim 1, characterized in that Extracting card-related information from the question data includes: Extracting at least one keyword from the question data; Extracting candidate keywords from at least one of the keywords according to part of speech; Acquire a vocabulary associated with the first business line and the second business line; Based on the associated word library, marking a first keyword and a second keyword from the candidate keywords, wherein the first keyword is a keyword included in the associated word library corresponding to the first business line; and the second keyword is a keyword included in the associated word library corresponding to the second business line; The cross-business information is generated according to the second keyword.

3. The method according to claim 2, characterized in that Extracting card-related information from the question data includes: Get the common vocabulary corresponding to the business architecture; Based on the general vocabulary, marking a third keyword in the candidate keywords, wherein the third keyword is a keyword included in the general vocabulary; The general information is generated according to the third keyword.

4. The method according to claim 1, characterized in that Downloading a card component from a public component library based on the card association information includes: generating a component download request according to the card association information, wherein the component download request includes card identification information; Sending the component download request to the public component library, so that the public component library queries the component download link of the card component in response to the component download request; Receive the component download link issued by the public component library, and download the card component by accessing the component download link.

5. The method according to claim 4, characterized in that Generating a component download request according to the card association information includes: Obtain the specification file corresponding to the general R&D specification; Reading a template protocol from the specification file, the template protocol is used to define the card identification information, the card identification information including at least one of a card name, a card ID, a card component package download path, and a card component input parameter; Based on the template protocol, query the card identification information corresponding to the card association information; The card identification information is encapsulated into a request instruction to generate the component download request.

6. The method according to claim 1, characterized in that Performing dynamic component rendering on the card component to generate answer data includes: Calling a parsing algorithm process based on a template protocol in the general R&D specification; Parse the card component using the parsing algorithm process to obtain data to be rendered; Obtain the script library to which the card component belongs, and the dynamic component rendering mode information corresponding to the script library; According to the dynamic component rendering mode information, the data to be rendered is rendered into an answer card component to generate the answer data.

7. The method according to claim 1, characterized in that The method further comprises: Obtaining component requirement information uploaded by the first business line and the second business line; Formulate the general R&D specification based on the component requirement information and business architecture; Obtaining components to be tested constructed by different business lines according to the general R&D specification, wherein the components to be tested include general components and business components, wherein the general components are components unrelated to specific business lines; and the business components are components related to specific business lines; Performing development testing on the component to be tested, and after the component to be tested passes the development testing, packaging the general component and the business component into a common component; The common components are stored in the common component library.

8. An AI robot intelligent question-answering device, characterized in that: The device comprises: A questioning module, used to obtain questioning data, where the questioning data is data input based on the question-answering interactive interface of the first business line; An information extraction module, used to extract card-related information from the question data, wherein the card-related information includes at least one of general information and cross-business information; A component download module, used to download a card component from a public component library based on the card association information, wherein the card component is a public component developed based on a general R&D specification; The answer module is used to perform dynamic component rendering on the card component to generate answer data.

9. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images