Intelligent voice interaction method suitable for government affair scene based on information management

By introducing intelligent voice interaction methods in government affairs scenarios, using digital people and deep learning algorithms to realize the automation and personalization of government services, the problem of inefficiency of traditional government services is solved, and the processing efficiency and user experience are improved.

CN119065501APending Publication Date: 2024-12-03SHENZHEN RADIO & TELEVISION DIGITAL TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411180043.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Traditional government service methods are inefficient and consume a lot of human resources, making it difficult to meet the growing needs of the people, especially in terms of government interaction with the people and personalized services.

Method used

Adopt intelligent voice interaction method based on information management, guide users to handle business through digital human wake-up and voice, use deep learning algorithms and neural networks to recognize voice, combine head posture interaction, automatically fill in forms and recommend related services, reduce manual intervention, and improve Efficiency and personalized service.

Benefits of technology

It realizes the automated processing of government services, reduces human resource consumption, improves service efficiency and interactive experience, can meet the growing needs of the people, and provides a personalized and convenient government affairs processing experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119065501A_ABST
    Figure CN119065501A_ABST
Patent Text Reader

Abstract

The invention discloses an information management-based intelligent voice interaction method suitable for a government affair scene, and relates to the technical field of AI digital humans, and the interaction steps are as follows: a service name is obtained to enter a service handling interface, a user wakes up a digital person through a client terminal, and the digital person guides the user to autonomously speak out or click to enter the service handling interface through voice; or the user informs the customer service staff of the name of the business to be handled by dialing the manual customer service, and the customer service staff operates to enter the business handling boundary through the transit server; basic information authorization is obtained, and the identity card information of the user is read before the customer service handles the business. According to the method, the form can be filled in based on a voice mode, user operation is facilitated, interaction with the AI digital person is carried out based on the head posture, time and labor are saved in service processing, consumption of a large number of human resources is avoided, efficiency, interaction experience and personalized services are improved, and the ever-increasing requirements of the masses can be met.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of AI digital human technology, and in particular to an intelligent voice interaction method based on information management and applicable to government affairs scenarios. Background Art

[0002] Government affairs scenarios refer to various application scenarios in which the government uses information technology and modern management methods in urban management, public services, etc. to build smart cities, improve government service levels, and meet the needs of the people. These scenarios cover many aspects of government work and aim to improve government service efficiency and enhance interaction and communication between the government and the people through digitalization, networking, and intelligentization. AI digital people refer to virtual people with human characteristics and interactive capabilities developed based on artificial intelligence technology. They can simulate various characteristics such as human language, expression, and voice, and interact with users in a realistic way. With the rapid development of information technology and the deepening of digital transformation, the field of government services is undergoing unprecedented changes. The traditional government service method relies solely on front-end customer service to handle user business, which is time-consuming and labor-intensive, consumes a lot of human resources, is inefficient, has a single interactive experience, and lacks personalized services, making it difficult to meet the growing needs of the people. In this regard, we propose an intelligent voice interaction method based on information management that is suitable for government affairs scenarios. Summary of the invention

[0003] In order to solve the above technical problems, an intelligent voice interaction method based on information management and suitable for government affairs scenarios is provided. This technical solution solves the above problems.

[0004] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0005] An intelligent voice interaction method based on information management and applicable to government affairs scenarios, the interaction steps are:

[0006] Obtain the service name and enter the service handling interface. The user wakes up the digital human through the client terminal. The digital human guides the user to speak or click to enter the service handling interface through voice. Alternatively, the user calls the manual customer service and tells the customer service staff the service name to be handled. The customer service staff then enters the service handling interface through the transfer server.

[0007] Obtain basic information authorization. Before the customer service handles the business, read the user's ID card information and inform the user of the list of materials and business handling instructions required for the current business handling. After reading and understanding, the user confirms the authorization on the terminal touch screen to handle the business;

[0008] Assist customers to fill in form information. Backstage customer service staff assist customers to fill in business form information through voice communication in the customer service system. Customers fill in the form by speaking. Based on deep learning algorithm technology, the customer's voice input is recognized and converted into text and transcribed in the corresponding position on the form. When the voice is not standard, it is matched with the vocabulary in the vocabulary library for similarity, and the vocabulary corresponding to the current voice is inferred and generated. The user judges whether it is correct and interactively judges right or wrong based on the action;

[0009] Submit materials. After filling in the information, upload the materials.

[0010] Sign and confirm the form information. After collecting the materials, the user previews the form information, confirms that it is correct, signs and confirms, and then handles the business;

[0011] Business recommendation, based on the current business type, screens out similar businesses from the business database and automatically recommends relevant business types to customers for them to check whether they need to handle them.

[0012] Preferably, the client terminal is software installed on the client terminal device for users to interact with the customer service system. The client application includes interface design, user authentication and information interaction functions to provide customers with a friendly user interface for sending consultations, asking questions and obtaining services.

[0013] Preferably, the transit server is used for intermediate forwarding in network communications, receiving requests from clients or source servers, forwarding the requests to the target server according to set rules or logic, and then returning the response of the target server to the original requester; the back-end server is the support for the client application, and is used to receive, process and respond to client requests. The back-end server includes database management, business logic processing, user authentication and authorization, and multi-channel message receiving and sending functions. The back-end server is integrated with databases, knowledge bases, third-party services, etc. to provide comprehensive and accurate services.

[0014] Preferably, the digital human wake-up method is based on user voice wake-up. The user wakes up and activates the digital human by saying a preset wake-up word. After the digital human is woken up, it interacts with the customer through voice and guides the user to handle the business. When the user cannot obtain relevant business information from the digital human, the digital human calls the backend customer service staff to operate.

[0015] Preferably, before handling the business, the user places the ID card information at the corresponding position on the customer terminal to read the user's ID card information; the customer service provides a form for the corresponding business to fill in, and when filling in the form, the user generates and fills in the statement based on voice interaction.

[0016] Preferably, the client terminal collects the user's voice signal in advance, converts the voice signal into a digital signal, preprocesses the converted digital signal, extracts feature information therein, selects a neural network as a model, prepares a data set, trains the model to optimize and improve the model, inputs the preprocessed voice signal into the model, performs conversion processing, and converts the voice signal into text information corresponding to the output on the form to be filled out.

[0017] Preferably, when faced with a non-standard voice signal, the current voice signal is compared with the vocabulary signal in the vocabulary library to determine the consistency, and the vocabulary signal that best solves the currently collected voice signal is selected as output and output on the form to be filled in; the consistency determination method directly calculates the correlation coefficient of the two groups of signals, and the calculation formula is:

[0018]

[0019] Where n is the number of sample points collected in the signal, X i With Y i are the observed values ​​of the two variables at the i-th sample point on the spectrum, and are the sample signal means of X and Y, r is the similarity of the two sets of signal data for comparison, where r's value of 1 indicates complete correlation, that is, the two sets of signals are consistent, and all other values ​​indicate inconsistency. The closer the r value is to 1, the more consistent the two sets of signals are, and vice versa. The similarity is determined by calculation, and the most similar vocabulary signals are screened out. The currently screened vocabulary signals are output and written on the form to be filled out.

[0020] Preferably, the action interaction behavior is a head movement. The client terminal obtains the head movement of the current user based on the camera, and determines whether the current form is filled out correctly based on the head posture. Shaking the head means an error, and nodding means correct.

[0021] Preferably, different types of services in the current scenario are classified based on a decision tree, and information gain is used as a standard for classifying service features to classify services. The information gain formula is:

[0022] Ga(A, D)=H(D)-H(DA)

[0023] Where Ga(A, D) is the information gain value, H(D) is the information entropy of data set D, H(DA) is the conditional entropy of data set D under the condition of given feature A, and the uncertainty of data set D after the value of feature A is known. By calculating the information gain of the feature, the splitting characteristics of the business are evaluated, and the splitting characteristics are used as the basis for business classification. Based on this basis, the business in this scenario is classified and processed.

[0024] Preferably, after the user has completed the service, the client terminal makes intelligent recommendations on the same type of services from the database based on the current service type, and generates a personalized service recommendation list. The recommendations include directly related service types, potential value-added services, and potential needs predicted by historical transaction behaviors. The presentation of the recommendation list is intuitive and easy to understand, making it convenient for customers to quickly browse and select.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] Based on the practical experience of government affairs, the present invention records customer information, communication content and processing results for unified management, and uses AI digital humans to automatically process part of the customer consultation process, which is completed through template replies and automatic triggering of processes. Through the intervention of manual customer service, it helps customers handle various government affairs, and can fill in forms based on voice, which is convenient for user operation. It interacts with AI digital humans based on head posture, saves time and effort in handling business, avoids the consumption of a large amount of human resources, improves efficiency, interactive experience and personalized services, and can meet the growing needs of the people. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flowchart of the interactive steps of the present invention;

[0028] Figure 2 This is a flowchart of the steps of voice filling in a form according to the present invention. DETAILED DESCRIPTION

[0029] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0030] Reference Figure 1 As shown, an intelligent voice interaction method based on information management and applicable to government affairs scenarios, the interaction steps are:

[0031] Obtain the service name and enter the service handling interface. The user wakes up the digital human through the client terminal. The digital human guides the user to speak or click to enter the service handling interface through voice. Alternatively, the user calls the manual customer service and tells the customer service staff the service name to be handled. The customer service staff then enters the service handling interface through the transfer server.

[0032] Obtain basic information authorization. Before the customer service handles the business, read the user's ID card information and inform the user of the list of materials and business handling instructions required for the current business handling. After reading and understanding, the user confirms the authorization on the terminal touch screen to handle the business;

[0033] Assist customers to fill in form information. Backstage customer service staff assist customers to fill in business form information through voice communication in the customer service system. Customers fill in the form by speaking. Based on deep learning algorithm technology, the customer's voice input is recognized and converted into text and transcribed in the corresponding position on the form. When the voice is not standard, it is matched with the vocabulary in the vocabulary library for similarity, and the vocabulary corresponding to the current voice is inferred and generated. The user judges whether it is correct and interactively judges right or wrong based on the action;

[0034] Submit materials. After filling in the information, upload the materials.

[0035] Sign and confirm the form information. After collecting the materials, the user previews the form information, confirms that it is correct, signs and confirms, and then handles the business;

[0036] Business recommendation, based on the current business type, screens out similar businesses from the business database and automatically recommends relevant business types to customers for them to check whether they need to handle them.

[0037] After the user sends out the required service through voice, the device automatically identifies the service type based on the voice and displays the corresponding service interface on the device screen. First, the user can choose the login method according to different methods, such as ID card sensing or mobile phone code scanning. After successful login, the user can fill in the corresponding information according to the business steps on the device screen. Finally, the user submits all the filled information. After successful submission, the device will voice prompt the processing result, and the terminal screen will display the successful processing interface. If the submission fails, the user will be informed by the digital human voice broadcast according to the specific error message;

[0038] In special business scenarios, users can request online support from manual customer service through voice or touch, and communicate with customer service through real-time voice. Manual customer service appears on the terminal device as a unified digital human image and voice, and asks the user about business needs. After clarifying the user's business needs, the customer service guides the user to identify the identity and determines whether the user meets the processing requirements based on the identity information. If the requirements are met, the customer service will continue to assist in the processing and require the user to take a photo of the corresponding materials and upload them through the terminal peripherals. After the information and materials are verified, the collected user information and materials will be submitted, and the returned results will be returned to the user in the form of printed receipts or text messages.

[0039] The client terminal is software installed on the client terminal device for users to interact with the customer service system. The client application includes interface design, user authentication and information interaction functions to provide customers with a friendly user interface for sending consultations, asking questions and obtaining services.

[0040] The transit server is used for intermediate forwarding in network communication, receiving requests from clients or source servers, forwarding requests to the target server according to the set rules or logic, and then returning the response of the target server to the original requester; the back-end server is the support of the client application, which is used to receive, process and respond to the client's request. The back-end server includes database management, business logic processing, user authentication and authorization, and multi-channel message receiving and sending functions. The back-end server is integrated with databases, knowledge bases, third-party services, etc. to provide comprehensive and accurate services;

[0041] Common application scenarios of transit servers include:

[0042] Load balancing: The transit server can distribute requests to multiple target servers according to the load balancing algorithm to achieve load balancing and improve system performance and availability;

[0043] Cache server: The transit server can cache some frequently accessed data or pages. When the same request arrives, the response can be returned directly from the cache to reduce the pressure on the target server;

[0044] Reverse proxy: The transit server can act as a reverse proxy server, receiving client requests and forwarding the requests to the backend target server, and then returning the target server's response to the client. This can hide the real server information, improve network security, and also perform request filtering, load balancing, caching and other functions;

[0045] Data processing: The transit server can be used for data encryption, decryption, compression, decompression, etc. It can convert or process data during data transmission to meet specific needs.

[0046] The digital human wake-up method is based on user voice wake-up. The user wakes up and activates the digital human by saying the preset wake-up word. After the digital human is awakened, it interacts with the customer through voice and guides the user to handle business. When the user cannot obtain relevant business information from the digital human, the digital human calls the backend customer service staff to operate.

[0047] Through voice wake-up and interaction, users can easily communicate with digital humans without using complex keyboard input or touch operations. This natural interaction method not only improves the user experience, but also enables users to quickly obtain the information or services they need anytime, anywhere; digital humans can respond to user requests instantly, quickly answer user questions or guide users to complete business processing through preset processes and knowledge bases. This immediacy greatly improves service efficiency, reduces user waiting time, and improves overall service satisfaction.

[0048] Before handling the business, the user places the ID card information at the corresponding position on the customer terminal to read the user's ID card information; the customer service provides a form for the corresponding business to fill in. When filling in the form, the user generates and fills in the form based on voice interaction.

[0049] Automatic reading of ID card information reduces the error rate of manual input and avoids security risks caused by inaccurate handwriting recognition. Accurate reading of ID card information ensures the legality and compliance of subsequent business handling. Voice interaction eliminates the need for users to manually enter large amounts of information, especially for users with poor eyesight or inconvenient operation. This method is more friendly and convenient. Users only need to use voice commands to complete the form filling, which greatly simplifies the operation process.

[0050] Reference Figure 2 As shown, the client terminal collects the user's voice signal in advance, converts the voice signal into a digital signal, preprocesses the converted digital signal, extracts feature information therein, selects a neural network as a model, prepares a data set, trains the model to optimize and improve the model, inputs the preprocessed voice signal into the model, performs conversion processing, and converts the voice signal into text information corresponding to the output on the form to be filled in.

[0051] When faced with non-standard speech signals, the current speech signal is compared with the vocabulary signals in the vocabulary library to determine the consistency, and the vocabulary signal that best solves the currently collected speech signal is selected as output and output on the form to be filled in; the consistency judgment method directly calculates the correlation coefficient of the two groups of signals, and the calculation formula is:

[0052]

[0053] Where n is the number of sample points collected in the signal, X i With Y i are the observed values ​​of the two variables at the i-th sample point on the spectrum, and are the sample signal means of X and Y, r is the similarity of the two sets of signal data for comparison, where r's value of 1 indicates complete correlation, that is, the two sets of signals are consistent, and all other values ​​indicate inconsistency. The closer the r value is to 1, the more consistent the two sets of signals are, and vice versa. The similarity is determined by calculation, and the most similar vocabulary signals are screened out. The currently screened vocabulary signals are output and written on the form to be filled out.

[0054] The digital signal preprocessing, feature extraction and neural network model in this application are all existing technologies and will not be elaborated here. The extracted speech feature vector is converted into text information based on the neural network model. In the output layer of RNN or DNN, the softmax function is used to predict the next word or character, and the generated text information is post-processed, such as part-of-speech tagging, named entity recognition, and dependency parsing, to improve recognition accuracy and comprehension ability.

[0055] The action interaction behavior is the movement of the head. The client terminal obtains the current user's head movement based on the camera, and determines whether the current form is filled out correctly based on the head posture. Shaking the head means wrong, and nodding means correct.

[0056] In this application, key points of the face, such as eyes, nose, and mouth, are detected in the captured image. These points will be used for subsequent three-dimensional posture estimation. The detected 2D face key points are matched with the predefined 3D face model to establish a corresponding relationship in the three-dimensional space. According to the conversion relationship between the 3D points and the corresponding 2D points, the posture angles of the head are calculated, including the pitch angle, yaw angle, and roll angle. These angles describe the rotation state of the head in the three-dimensional space. According to the calculated head posture angles, combined with the preset interaction rules or machine learning model, the user's intention is identified. The user confirms by nodding and denies by shaking his head.

[0057] Based on the decision tree, different types of services in the current scenario are classified and processed. Information gain is used as the standard for service feature classification to classify services. The information gain formula is:

[0058] Ga(A, D)=H(D)-H(DA)

[0059] Where Ga(A, D) is the information gain value, H(D) is the information entropy of data set D, H(DA) is the conditional entropy of data set D under the condition of given feature A, and the uncertainty of data set D after the value of feature A is known. By calculating the information gain of the feature, the splitting characteristics of the business are evaluated, and the splitting characteristics are used as the basis for business classification. Based on this basis, the business in this scenario is classified and processed.

[0060] After the user completes the service, the client terminal makes intelligent recommendations on the same type of services from the database based on the current service type, and generates a personalized service recommendation list. The recommendations include directly related service types, potential value-added services, and potential needs predicted by historical transaction behaviors. The recommendation list is presented in an intuitive and easy-to-understand manner, allowing customers to quickly browse and select.

[0061] This application adopts the presentation method of recommendation list, which is intuitive and easy to understand, convenient for customers to quickly browse and select, reducing the time cost of customers to find the required services in massive information and improving service efficiency. At the same time, personalized recommendations also reduce the possibility of customers being disturbed by irrelevant information, further optimizing the customer experience. The intelligent recommendation system can predict the potential needs of customers based on historical transaction behavior, so as to plan recommendation strategies in advance.

[0062] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An intelligent voice interaction method based on information management and applicable to government affairs scenarios, characterized in that: The interaction steps are: Obtain the service name and enter the service handling interface. The user wakes up the digital human through the client terminal. The digital human guides the user to speak or click to enter the service handling interface through voice. Alternatively, the user calls the manual customer service and tells the customer service staff the service name to be handled. The customer service staff then enters the service handling interface through the transfer server. Obtain basic information authorization. Before the customer service handles the business, read the user's ID card information and inform the user of the list of materials and business handling instructions required for the current business handling. After reading and understanding, the user confirms the authorization on the terminal touch screen to handle the business; Assist customers to fill in form information. Backstage customer service staff assist customers to fill in business form information through voice communication in the customer service system. Customers fill in the form by speaking. Based on deep learning algorithm technology, the customer's voice input is recognized and converted into text and transcribed in the corresponding position on the form. When the voice is not standard, it is matched with the vocabulary in the vocabulary library for similarity, and the vocabulary corresponding to the current voice is inferred and generated. The user judges whether it is correct and interactively judges right or wrong based on the action; Submit materials. After filling in the information, upload the materials. Sign and confirm the form information. After collecting the materials, the user previews the form information, confirms that it is correct, signs and confirms, and then handles the business; Business recommendation, based on the current business type, screens out similar businesses from the business database and automatically recommends relevant business types to customers for them to check whether they need to handle them.

2. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: The client terminal is software installed on the client terminal device for users to interact with the customer service system. The client application includes interface design, user authentication and information interaction functions to provide customers with a friendly user interface for sending consultations, asking questions and obtaining services.

3. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: The relay server is used to perform intermediate forwarding in network communications, receiving requests from clients or source servers, forwarding the requests to the target server according to the set rules or logic, and then returning the response of the target server to the original requester; The back-end server is the support for the client application, which is used to receive, process and respond to client requests. The back-end server includes database management, business logic processing, user authentication and authorization, and multi-channel message receiving and sending functions. The back-end server is integrated with databases, knowledge bases, third-party services, etc. to provide comprehensive and accurate services.

4. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: The digital human wake-up method is based on user voice wake-up. The user wakes up and activates the digital human by saying the preset wake-up word. After the digital human is awakened, it interacts with the customer through voice and guides the user to handle business. When the user cannot obtain relevant business information from the digital human, the digital human calls the backend customer service staff to operate.

5. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: Before handling the business, the user places the ID card information at the corresponding position on the customer terminal to read the user's ID card information; the customer service provides a form for the corresponding business to fill in. When filling in the form, the user generates and fills in the form based on voice interaction.

6. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: The client terminal collects the user's voice signal in advance, converts the voice signal into a digital signal, preprocesses the converted digital signal, extracts feature information therein, selects a neural network as a model, prepares a data set, trains the model to optimize and improve the model, inputs the preprocessed voice signal into the model, converts it, and outputs the voice signal into text information on the form to be filled out.

7. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: When faced with non-standard speech signals, the current speech signal is compared with the vocabulary signals in the vocabulary library to determine the consistency, and the vocabulary signal that best solves the currently collected speech signal is selected as output and output on the form to be filled in; the consistency judgment method directly calculates the correlation coefficient of the two groups of signals, and the calculation formula is: Where n is the number of sample points collected in the signal, X i With Y i are the observed values ​​of the two variables at the i-th sample point on the spectrum, and are the sample signal means of X and Y, r is the similarity of the two sets of signal data for comparison, where r's value of 1 indicates complete correlation, that is, the two sets of signals are consistent, and all other values ​​indicate inconsistency. The closer the r value is to 1, the more consistent the two sets of signals are, and vice versa. The similarity is determined by calculation, and the most similar vocabulary signals are screened out. The currently screened vocabulary signals are output and written on the form to be filled out.

8. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: The action interaction behavior is the movement of the head. The client terminal obtains the current user's head movement based on the camera, and determines whether the current form is filled out correctly based on the head posture. Shaking the head means wrong, and nodding means correct.

9. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 1 is characterized in that: Based on the decision tree, different types of services in the current scenario are classified and processed. Information gain is used as the standard for service feature classification to classify services. The information gain formula is: Ga(A, D)=H(D)-H(D|A) Where Ga(A, D) is the information gain value, H(D) is the information entropy of data set D, H(DA) is the conditional entropy of data set D under the condition of given feature A, and the uncertainty of data set D after the value of feature A is known. By calculating the information gain of the feature, the splitting characteristics of the business are evaluated, and the splitting characteristics are used as the basis for business classification. Based on this basis, the business in this scenario is classified and processed.

10. The intelligent voice interaction method based on information management and applicable to government affairs scenarios according to claim 9 is characterized in that: After the user completes the service, the client terminal makes intelligent recommendations on the same type of services from the database based on the current service type, and generates a personalized service recommendation list. The recommendations include directly related service types, potential value-added services, and potential needs predicted by historical transaction behaviors. The recommendation list is presented in an intuitive and easy-to-understand manner, allowing customers to quickly browse and select.