Method and system for carrying out government affair service interaction by using large model

Abstract

The invention discloses a method and a system for carrying out government affair service interaction by utilizing a large model, and relates to a government affair service optimization technology in the field of government affair. Comprising the steps of 1, directly obtaining OCR data and ASR records from existing communication records of a cloud hall, 2, carrying out RAG updating and timed fine tuning on a large model according to the OCR data and the ASR records, and 3, deploying a client based on a pc end of the cloud hall, obtaining communication information in real time through the client, and sending the communication information to the cloud hall. Wherein screen information in the government affair service handling process is recorded through screen recording of the client side, the screen information comprises key steps and an operation interface, and voice content of a client is converted into text content in real time through a voice recognition method; and 5, automatically arranging and executing a government affair handling process based on a matching result, and carrying out authenticity judgment when the certificate is involved.

Description

Technical Field

[0001] The present invention discloses a method and system for government service interaction using a large model, and relates to a government service optimization technology in the field of government affairs. Background Art

[0002] In the current government service process, service personnel must manually answer questions through face-to-face communication or video conferencing in a cloud lobby. This method requires a large workload, making it difficult for service personnel to handle multiple matters simultaneously. Due to the heavy workload, citizens are forced to wait in line at the lobby or on the service website, resulting in long wait times. Summary of the Invention

[0003] In response to the problems of the prior art, the present invention provides a method and system for government service interaction using a large model, which uses the large model to realize the intelligence and automation of government services, thereby improving processing efficiency and service quality.

[0004] The specific scheme proposed by the present invention is:

[0005] The present invention provides a method for government service interaction using a large model, comprising:

[0006] Step 1: Directly obtain OCR data and ASR records from the existing communication records in the cloud lobby.

[0007] Step 2: Update the RAG and adjust the timing of the large model based on the OCR data and ASR records.

[0008] Step 3: Deploy the client on the PC side of the cloud lobby and obtain communication information in real time through the client. The screen information of the government service process is recorded through the client, including key steps and operation interface. The voice content of the customer is converted into text content in real time using the voice recognition method.

[0009] Step 4: Use the fine-tuned large model to analyze communication information in real time, extract key information from screen information and text content, and match it with the knowledge base.

[0010] Step 5: Automatically organize and execute government affairs processing procedures based on the matching results, and make authenticity judgments when certificates and licenses are involved.

[0011] Furthermore, in step 1 of the method for interacting with government services using a large model, the acquired OCR data and ASR records are desensitized and preprocessed to ensure the quality and consistency of the data.

[0012] Furthermore, in step 2 of the method for using a large model for government service interaction, regular screenshots and voice transcription are performed based on the cloud lobby, and the acquired data is used as the RAG industry knowledge expansion of the large model intelligent body, and new communication records are acquired and added to the knowledge base in real time, and the RAG content is updated in real time.

[0013] Regularly use RAG content as fine-tuning material according to the preset cycle to fine-tune the large model to adapt to new business needs.

[0014] Furthermore, in step 5 of the method for using a large model to interact with government services, the extracted real certificates and licenses are used as training samples to train the YOLO model to identify the authenticity of the certificates and licenses, and perform authenticity judgment on the certificates and licenses.

[0015] The present invention also provides a system for government service interaction using a large model, including a data acquisition module, a large model fine-tuning module, a client management module, an information extraction and matching module, a process arrangement and execution module, and a license recognition module.

[0016] The data collection module directly obtains OCR data and ASR records from the existing communication records in the cloud lobby.

[0017] The large model fine-tuning module performs RAG updates and regular fine-tuning on the large model based on OCR data and ASR records.

[0018] The client management module is deployed on the PC side of the cloud hall. The client management module obtains communication information in real time. The client management module records the screen information during the government service process, including key steps and operation interfaces, and uses voice recognition methods to convert the customer's voice content into text content in real time.

[0019] The information extraction and matching module uses the fine-tuned large model to analyze communication information in real time, extract key information from screen information and text content, and match it with the knowledge base.

[0020] The process orchestration and execution module automatically orchestrates and executes government affairs processing based on the matching results, and the certificate recognition module makes authenticity judgments when certificates are involved.

[0021] Furthermore, the data acquisition module of the system for government service interaction using a large model performs data desensitization and preprocessing on the acquired OCR data and ASR records to ensure the quality and consistency of the data.

[0022] Furthermore, the large model fine-tuning module of the system using large models for government service interaction performs regular screenshots and voice transcription based on the cloud lobby, uses the acquired data as the RAG industry knowledge expansion of the large model intelligent body, acquires and adds new communication records to the knowledge base in real time, and updates the RAG content in real time.

[0023] Regularly use RAG content as fine-tuning material according to the preset cycle to fine-tune the large model to adapt to new business needs.

[0024] Furthermore, the certificate recognition module of the system for government service interaction using a large model uses the extracted real certificates as training samples to train the YOLO model to identify the authenticity of the certificates and perform authenticity judgment on the certificates.

[0025] The benefits of the present invention are:

[0026] Improve service efficiency: Through the 24-hour uninterrupted service of large-scale intelligent bodies, the efficiency of government service processing has been greatly improved.

[0027] Reduce manual dependence: Large model intelligent agents can automatically schedule and process government affairs procedures according to actual application needs, greatly reducing dependence on manual services, reducing manual workload, and making government services more efficient and convenient.

[0028] Process orchestration optimization: Automated process orchestration mechanism, the system can automatically identify and handle the dependencies between various matters, avoiding omissions and errors, making the overall business process clearer and more orderly.

[0029] User experience optimization: greatly reduces the problem of information asymmetry and improves the user experience.

[0030] Automated data processing: Through voice recognition and screenshot technology, communication information can be collected and analyzed in real time, and information extraction and matching can be completed automatically, which greatly improves the automation level of data processing and reduces human errors and operation delays.

[0031] Preliminary identification of the authenticity of certificates and licenses: It can identify and verify the authenticity of certificates and licenses, reduce rejections and misjudgments, and improve the quality and security of government services. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the application process of the method of the present invention. DETAILED DESCRIPTION

[0033] The present invention will be further described below with reference to the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it. However, the embodiments are not intended to limit the present invention.

[0034] Example 1

[0035] The present invention provides a method for government service interaction using a large model, comprising:

[0036] Step 1: Directly obtain OCR data and ASR records from the existing communication records in the cloud lobby, and perform data desensitization and preprocessing on the obtained OCR data and ASR records to ensure data quality and consistency.

[0037] Step 2: Update the RAG and regularly fine-tune the large model based on OCR data and ASR records. You can further perform regular screenshots and voice transcription based on the cloud lobby, use the acquired data as the RAG industry knowledge expansion of the large model intelligent body, acquire and add new communication records to the knowledge base in real time, and update the RAG content in real time.

[0038] Regularly use RAG content as fine-tuning material according to the preset cycle to fine-tune the large model to adapt to new business needs.

[0039] Step 3: Deploy the client on the PC side of the cloud lobby and obtain communication information in real time through the client. The screen information of the government service process is recorded through the client, including key steps and operation interface. The voice content of the customer is converted into text content in real time using the voice recognition method.

[0040] Step 4: Use the fine-tuned large model to analyze communication information in real time, extract key information from screen information and text content, and match it with the knowledge base.

[0041] Step 5: Automatically orchestrate and execute government affairs processes based on the matching results, and perform authenticity verification when documents are involved. The extracted authentic documents can be used as training samples to train a YOLO model to identify and verify the authenticity of documents.

[0042] Example 2

[0043] The present invention also provides a system for government service interaction using a large model, including a data acquisition module, a large model fine-tuning module, a client management module, an information extraction and matching module, a process arrangement and execution module, and a license recognition module.

[0044] The data collection module directly obtains OCR data and ASR records from the existing communication records in the cloud lobby.

[0045] The large model fine-tuning module performs RAG updates and regular fine-tuning on the large model based on OCR data and ASR records.

[0046] The client management module is deployed on the PC side of the cloud hall. The client management module obtains communication information in real time. The client management module records the screen information during the government service process, including key steps and operation interfaces, and uses voice recognition methods to convert the customer's voice content into text content in real time.

[0047] The information extraction and matching module uses the fine-tuned large model to analyze communication information in real time, extract key information from screen information and text content, and match it with the knowledge base.

[0048] The process orchestration and execution module automatically orchestrates and executes government affairs processing based on the matching results, and the certificate recognition module makes authenticity judgments when certificates are involved.

[0049] The information interaction, execution process and other contents between the modules in the above system are based on the same concept as the embodiment of the method of the present invention. For specific contents, please refer to the description in the embodiment of the method of the present invention and will not be repeated here.

[0050] Likewise, the system of the present invention is beneficial in that:

[0051] Improve service efficiency: Through the 24-hour uninterrupted service of large-scale intelligent bodies, the efficiency of government service processing has been greatly improved.

[0052] Reduce manual dependence: Large model intelligent agents can automatically schedule and process government affairs procedures according to actual application needs, greatly reducing dependence on manual services, reducing manual workload, and making government services more efficient and convenient.

[0053] Process orchestration optimization: Automated process orchestration mechanism, the system can automatically identify and handle the dependencies between various matters, avoiding omissions and errors, making the overall business process clearer and more orderly.

[0054] User experience optimization: greatly reduces the problem of information asymmetry and improves the user experience.

[0055] Automated data processing: Through voice recognition and screenshot technology, communication information can be collected and analyzed in real time, and information extraction and matching can be completed automatically, which greatly improves the automation level of data processing and reduces human errors and operation delays.

[0056] Preliminary identification of the authenticity of certificates and licenses: It can identify and verify the authenticity of certificates and licenses, reduce rejections and misjudgments, and improve the quality and security of government services.

[0057] Correspondingly, users can use mobile applications to handle new matters or supplement materials. For example, through the application login: users log in with their mobile phone number, ID number or other valid credentials;

[0058] Obtain service guide: Provide detailed service guide, including the handling procedures and required materials for each matter;

[0059] Instant communication: Users can communicate with the agent in real time in the application to solve problems encountered during the application process;

[0060] Handle matters: Users can submit the required materials online, and the system will automatically complete the handling process of the matter;

[0061] Upload materials: Users can upload required materials at any time in the application to avoid matters being stalled due to insufficient materials.

[0062] Task reminder: The system will remind users to complete unfinished tasks or supplement materials based on their free time.

[0063] It should be noted that not all steps and modules in the above-mentioned processes and system structures are required, and certain steps or modules can be omitted according to actual needs. The execution order of each step is not fixed and can be adjusted as needed. The system structure described in the above-mentioned embodiments can be a physical structure or a logical structure, that is, some modules may be implemented by the same physical entity, or some modules may be implemented by multiple physical entities, or may be implemented by certain components in multiple independent devices.

[0064] The above embodiments are merely preferred embodiments for the purpose of fully illustrating the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are within the scope of protection of the present invention. The scope of protection of the present invention shall be subject to the claims.

Claims

1. A method for government service interaction using a large model, characterized by include: Step 1: Directly obtain OCR data and ASR records from the existing communication records in the cloud lobby. Step 2: Update the RAG and adjust the timing of the large model based on the OCR data and ASR records. Step 3: Deploy the client on the PC side of the cloud lobby and obtain communication information in real time through the client. The screen information of the government service process is recorded through the client, including key steps and operation interface. The voice content of the customer is converted into text content in real time using the voice recognition method. Step 4: Use the fine-tuned large model to analyze communication information in real time, extract key information from screen information and text content, and match it with the knowledge base. Step 5: Automatically organize and execute government affairs processing procedures based on the matching results, and make authenticity judgments when certificates and licenses are involved.

2. The method for government service interaction using a large model according to claim 1 is characterized by: In step 1, the acquired OCR data and ASR records are desensitized and preprocessed to ensure data quality and consistency.

3. The method for government service interaction using a large model according to claim 1 is characterized by: In step 2, based on the cloud lobby, we take regular screenshots and transcribe voices, and use the acquired data as the RAG industry knowledge expansion of the large model intelligent body. We acquire and add new communication records to the knowledge base in real time, and update the RAG content in real time. Regularly use RAG content as fine-tuning material according to the preset cycle to fine-tune the large model to adapt to new business needs.

4. The method for government service interaction using a large model according to claim 1 is characterized by: In step 5, the extracted real certificates are used as training samples to train the YOLO model to identify the authenticity of the certificates and make authenticity judgments.

5. A system for government service interaction using a large model, characterized by It includes data acquisition module, large model fine-tuning module, client management module, information extraction and matching module, process arrangement and execution module and license recognition module. The data collection module directly obtains OCR data and ASR records from the existing communication records in the cloud lobby. The large model fine-tuning module performs RAG updates and regular fine-tuning on the large model based on OCR data and ASR records. The client management module is deployed on the PC side of the cloud hall. The client management module obtains communication information in real time. The client management module records the screen information during the government service process, including key steps and operation interfaces, and uses voice recognition methods to convert the customer's voice content into text content in real time. The information extraction and matching module uses the fine-tuned large model to analyze communication information in real time, extract key information from screen information and text content, and match it with the knowledge base. The process orchestration and execution module automatically orchestrates and executes government affairs processing based on the matching results, and the certificate recognition module makes authenticity judgments when certificates are involved.

6. A system for government service interaction using a large model according to claim 5, characterized in that The data acquisition module desensitizes and preprocesses the acquired OCR data and ASR records to ensure data quality and consistency.

7. The system for government service interaction using a large model according to claim 5 is characterized by: The large model fine-tuning module takes regular screenshots and voice transcriptions based on the cloud lobby, uses the acquired data as the RAG industry knowledge expansion of the large model intelligent body, acquires and adds new communication records to the knowledge base in real time, and updates the RAG content in real time. Regularly use RAG content as fine-tuning material according to the preset cycle to fine-tune the large model to adapt to new business needs.

8. The system for government service interaction using a large model according to claim 5 is characterized by: The certificate recognition module uses the extracted real certificates as training samples to train the YOLO model to identify the authenticity of certificates and make authenticity judgments.

Images