Intelligent verbal skill recommendation method and system based on large-scale language model

By building a large-scale language model and server-side framework, the problem of inefficiency of intelligent customer service systems is solved, rapid response and customer demand identification are achieved, customer experience and system adaptability are improved, and multi-channel seamless connection and data-driven decision-making are supported.

CN119026680APending Publication Date: 2024-11-26CITIC AIBANK CORPORATION LIMITED
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411127016.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing intelligent customer service system is inefficient, unable to flexibly adapt to new scenarios, rely on manual experience, slow response speed, and unable to effectively combine historical information to meet customer needs.

Method used

By collecting and classifying customer dialogue historical corpus, a large-scale language model is introduced to build a data knowledge base, a server-side framework is established, and a quick response and update of intelligent language is achieved, customer tags are generated, and multi-channel seamless connection and data-driven decision-making is supported. .

Benefits of technology

It improves the response speed and personalization of customer experience of smart customer service, reduces costs, enhances the usability and ease of use of the system, supports multiple frequency and multi-scenario intelligent customer service sessions, and can quickly adapt to new business scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119026680A_ABST
    Figure CN119026680A_ABST
Patent Text Reader

Abstract

The invention relates to an intelligent verbal skill recommendation method based on a large-scale language model, and the method comprises the steps: S1, combing guest communication experience into an enterprise internal knowledge base, and enabling the large-scale language model to complete the training and parameter adjustment through the learning of the enterprise internal knowledge base; s2, analyzing the enterprise internal knowledge base by using the large-scale language model, generating customer tags and intelligent verbal skills, and updating the enterprise internal knowledge base according to newly added scene information; s3, establishing a server side to realize quick response and updating of the intelligent verbal skill and the client label; and S4, realizing real-time interaction with the customer through calling the customer label by the intelligent talking skill, and completing customer problem solving and / or product promotion. According to the application, the intelligent talking skill can be output by utilizing a large-scale language model according to the historical session information of the customer to help the personalized service customer.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an intelligent speech recommendation method and system based on a large-scale language model. Background Art

[0002] With the rapid development of artificial intelligence technology, more and more intelligent applications for enterprises have emerged. For example, intelligent customer service is one of the most typical examples. At present, although there are many types of intelligent customer service products on the market, most of them are single-function products that cannot meet the various needs of enterprises. Therefore, how to design a multifunctional intelligent customer service that can cope with multiple complex situations at the same time is an urgent problem to be solved. Intelligent customer service usually consists of two major components-the front-end system and the back-end system. In order to better understand the specific operation process of these two systems, we can simply divide them into two parts: the front-end customer interface part and the part using the back-end algorithm. For front-end customers, they are mainly responsible for entering questions and asking questions. When encountering some difficult to explain or unanswerable questions, they can send requests to the back-end algorithm and wait for feedback results. For the back-end, it reads a large amount of existing information and uses relevant algorithm technology to generate corresponding feedback to help solve the current problem and provide a reference for subsequent similar problems.

[0003] However, the existing technical solutions still have the following defects: (i) The existing solutions usually use manual intervention to solve these low-repeatability and cumbersome and lengthy work processes, which makes the entire intelligent customer service system inefficient and prone to the risk of wrong decision-making; (ii) Due to the lack of sufficient flexible adaptability, it is often impossible to accurately judge the cause in a timely manner when facing new scenarios or special circumstances, further reducing the actual application value and feasibility; (iii) Traditional customer service script replies mainly rely on manual experience and training, which are inefficient, slow to respond, and cannot combine historical information to effectively clarify customer needs. In summary, there is still much room for improvement in the existing technology. Summary of the invention

[0004] In order to address the deficiencies of the prior art, the present invention realizes the needs of multi-party reconciliation, specifically: after collecting and classifying the customer's conversation history corpus, a large-scale language model is introduced to build a relevant data knowledge base as the basis for subsequent business processing. At the same time, the intelligent speech can be flexibly adjusted to meet the needs of use in different scenarios through analysis of the existing knowledge base and supplementation of new data. In addition, a corresponding server-side framework is established to improve the efficiency of intelligent speech in extracting customer needs and the ability to effectively identify customer characteristics, etc., thereby improving the application effect of intelligent customer service in this field and providing technical support for the development of intelligent customer service platforms. To achieve the above purposes, the technical solutions adopted by the present invention include:

[0005] A method for intelligent speech recommendation based on a large-scale language model, characterized by comprising:

[0006] S1. The customer communication experience is sorted into an internal knowledge base of the enterprise. The large-scale language model completes training and parameter adjustment by learning the internal knowledge base of the enterprise.

[0007] S2. Analyze the internal knowledge base of the enterprise using the large-scale language model, generate customer tags and intelligent speech, and update the internal knowledge base of the enterprise according to the newly added scenario information;

[0008] S3. The server implements rapid response and update of the intelligent speech and customer tags;

[0009] S4. Realize real-time interaction with customers and solve customer problems and / or promote products by calling the customer tags through the intelligent speech.

[0010] In some embodiments, the S3 establishes a server to implement rapid response and update of the intelligent speech and customer tags, including a request receiving module, a message processing module, and a data calculation module. The request receiving module creates a Facade class and encapsulates internal call logic. After receiving the scene information, the request receiving module forwards the scene information to the message processing module.

[0011] Specifically, because the Facade class provides a stable interface, internal components can be easily maintained and expanded without changing external callers. In addition, it can flexibly handle requests for intelligent speech and customer tags while maintaining transparency to external callers, thereby improving the availability and ease of use of the system.

[0012] In some embodiments, the message processing module defines the subject and message format through the message middleware, receives the message scenario field sent to the subject, and selects the correct policy processing class according to the received message scenario field.

[0013] In some embodiments, the data calculation module receives the intelligent speech data sent by the message processing module, calculates the priority of the intelligent speech and / or customer tags, obtains the recommended intelligent speech and / or customer tags returned by the large-scale language model based on the configuration information of the intelligent speech and / or the customer tags, assembles and feeds back.

[0014] In some embodiments, the policy processing class implements the corresponding logic between processing the message scenario field and the intelligent speech and / or customer tag.

[0015] In some embodiments, the server also includes a configuration information loading module, which loads the configuration information of the intelligent speech into the Redis cache. When the configuration information is modified, a delayed double deletion mechanism is used to ensure the accuracy of the configuration information in the Redis cache.

[0016] The present invention also provides an intelligent speech recommendation system based on a large-scale language model, which is characterized by comprising:

[0017] The knowledge base building module is used to organize customer communication experience into an internal knowledge base of the enterprise. The large-scale language model completes training and parameter adjustment by learning the internal knowledge base of the enterprise.

[0018] A customer tag generation module, used to analyze the enterprise internal knowledge base using the large-scale language model, generate customer tags and intelligent speech, and update the enterprise internal knowledge base according to the newly added scenario information;

[0019] A server module, used to realize rapid response and update of the intelligent speech and customer tags;

[0020] The client module is used to achieve real-time interaction with customers and solve customer problems and / or promote products by calling the customer tags through the intelligent speech.

[0021] The present invention also provides an electronic device, characterized in that it includes a processor and a memory;

[0022] The memory is used to store computer programs;

[0023] The processor is configured to execute the method according to any one of claims 1 to 6 by calling a stored computer program.

[0024] The present invention also relates to a computer program product, comprising a computer program and / or instructions, characterized in that the computer program and / or instructions implement the steps of the above method when executed by a processor.

[0025] The beneficial effects of the present invention are: based on this solution, it is possible to achieve efficient and fast response to customers, personalized customer experience, seamless connection of multiple channels, data-driven decision-making, cost reduction and efficiency improvement, continuous learning and optimization, without over-reliance on the personal abilities of customer service personnel and consuming too much energy. In addition, based on the internal knowledge base of the enterprise and the newly added scenario information, the large model intelligently constructs customer tags in real time, without relying on customer service personnel to manually identify customer characteristics, and can accurately mine the potential demands of different customers, and support intelligent customer service conversation demands with multiple frequencies, multiple scenarios, custom tags and words, and linkage marketing. It has strong scalability and can quickly support the access of new business scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1The figure is a flow chart of the intelligent speech recommendation method based on a large-scale language model of the present invention.

[0027] Figure 2 It is a structural diagram of the intelligent speech recommendation system based on a large-scale language model of the present invention.

[0028] Figure 3 This is a schematic diagram corresponding to the structure and method of the intelligent speech recommendation system based on a large-scale language model of the present invention. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0030] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0031] like Figure 1 The figure shows a flow chart of an intelligent speech recommendation method based on a large-scale language model provided by the present invention, which mainly includes the following steps:

[0032] S1. The customer communication experience is sorted into an internal knowledge base of the enterprise. The large-scale language model completes training and parameter adjustment by learning the internal knowledge base of the enterprise.

[0033] S2. Analyze the internal knowledge base of the enterprise using the large-scale language model, generate customer tags and intelligent speech, and update the internal knowledge base of the enterprise according to the newly added scenario information;

[0034] S3. The server implements rapid response and update of the intelligent speech and customer tags;

[0035] S4. Realize real-time interaction with customers and solve customer problems and / or promote products by calling the customer tags through the intelligent speech.

[0036] Specifically, if Figure 3 As shown, the first step is to organize the key words and corresponding answers in the conversation into documents based on the historical experience of customer communication and store them in the company's database to form the initial knowledge base within the company.

[0037] The second step is to use a large-scale language model to analyze the semantic relationships in the company's original knowledge base, and continuously learn from the company's internal knowledge base, so as to continuously optimize the parameter settings of the large-scale language model. After reaching the set effect threshold, stop further adjustments and finally obtain a large-scale language model that can meet application needs.

[0038] The third step is to start formally analyzing the intelligent speech and corresponding customer tags required for existing business scenarios after the initial construction of the large-scale language model is completed, and to continuously improve the content of the company's internal knowledge base to adapt to changes in new scenarios.

[0039] The fourth step is to analyze the existing large amount of high-quality data, and then use a deep neural network-based method to build a server-side model framework, which enables rapid response and intelligent speech output for application scenarios, as well as identity recognition and label judgment for application customers, thereby improving customer experience and increasing product sales.

[0040] The fifth step is to develop APP and web-based applications suitable for different platforms and environments for customers to download and use. At the same time, it can also support web interface access to facilitate customers' use anytime and anywhere.

[0041] The sixth step is to use the delayed double deletion mechanism to ensure the reliability of data in the redis cache. At the same time, the redis cache will be regularly checked to see if there are any expiration issues caused by unavailable keys or values, and these issues will be repaired in a timely manner.

[0042] Step 7. After using the API interface provided by the redis server, record the status information in the log file in a timely manner so that subsequent users can view it.

[0043] Step 8: Use a scheduled task to scan whether the submit button on the form submission page has been clicked. If yes, delete all data in the post form to avoid leakage.

[0044] In the ninth step, each attribute submitted in each form will be checked separately. If there are illegal characters, it will prompt that the check failed and the submission will not be allowed to continue, thus ensuring the security of the form.

[0045] The tenth step is to use a variety of filters, such as regular expressions, SQL injection protection, XSS defense and other security measures to protect and improve the robustness and security of the entire app application.

[0046] like Figure 2 The present invention provides an intelligent speech recommendation system based on a large-scale language model, which specifically includes:

[0047] Knowledge base building module: used to organize customer communication experience into an internal knowledge base of the enterprise. The large-scale language model completes training and parameter adjustment by learning the internal knowledge base of the enterprise.

[0048] A customer tag generation module, used to analyze the enterprise internal knowledge base using the large-scale language model, generate customer tags and intelligent speech, and update the enterprise internal knowledge base according to the newly added scenario information;

[0049] A server module, used to realize rapid response and update of the intelligent speech and customer tags;

[0050] The client module is used to achieve real-time interaction with customers and solve customer problems and / or promote products by calling the customer tags through the intelligent speech.

[0051] By using this system, the above-mentioned operation processing method can be executed and the corresponding technical effects can be achieved.

[0052] Specifically, the system and the method can be transformed and applied to each other. The internal knowledge base of the enterprise is a summary based on historical conversation experience, and historical information is used as data for training large-scale language models. By adjusting parameters, the large-scale language model can analyze real-time customer communication information. In the process of information analysis in specific scenarios, customer labels are fed back and output to facilitate the priority determination of intelligent speech recommendations.

[0053] For the priority confirmation of intelligent speech, the data calculation module first queries the intelligent speech data obtained by the intelligent speech message processing module, then calculates the intelligent speech / customer tag recommendation priority and other information based on the intelligent speech configuration information, and finally obtains the recommended speech / tag information returned by the large model based on the intelligent speech / customer tag configuration information, assembles it, and returns the intelligent speech / customer tag request to the client.

[0054] In the information processing on the server side, one or more topics and the message formats of the messages sent to these topics are defined in the message middleware. Different parts of the system publish intelligent speech scenario messages to the defined topics. The message processing module subscribes to the relevant topics and receives messages sent to these topics. According to the content of the message, the message is dynamically routed to the correct policy processing class.

[0055] Through message middleware (such as RocketMQ), the system can process messages asynchronously, improving performance and throughput. The use of message middleware decouples message producers and consumers, allowing each part of the system to be developed and deployed independently. The use of strategy mode allows the system to flexibly expand new processing strategies according to different intelligent speech scenarios without modifying existing code.

[0056] In addition, the client starts the judgment thread when the customer service is talking to the customer, and immediately initiates an intelligent speech / customer tag transaction request to the server, and performs actions such as intelligent speech / customer tag data assembly and intelligent speech / customer tag rendering based on the transaction request feedback.

[0057] After completing an intelligent speech / customer tag transaction request, the intelligent speech / customer tag judgment thread enters the polling mode, waits for the specified time, and then initiates the intelligent speech / customer tag transaction request again to obtain the latest intelligent speech / customer tag data and complete the real-time recommendation of intelligent speech and customer tags.

[0058] It can be seen from the technical solution of the above invention that it has the following advantages compared with the prior art:

[0059] (1) The present invention provides an intelligent speech recommendation method and system based on a large-scale language model. The method mainly includes two parts: one is how to generate a large-scale language model suitable for application from the original enterprise, and the other is how to find reasonable intelligent speech and corresponding customer tags for different application scenarios and combine and output them. Finally, relevant hardware and software resources are provided to realize these functions. This method not only improves the convenience of customer communication but also improves the efficiency of customer service staff.

[0060] (2) The present invention also provides a system for intelligent speech recommendation based on a large-scale language model, which mainly includes two parts. One part is an improvement on the original large-scale language model, and the other part is to add multiple related components on the basis of the original model to facilitate the later development and application expansion of application scenarios. This design not only saves costs and meets the needs of later development, but also has good compatibility.

[0061] (3) The system provided by the present invention can be run not only on the PC side but also on the mobile platform, which greatly broadens the scope of use and increases the convenience of use. The above is only a preferred embodiment of the present invention and is not a basis for limiting the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention shall be included in the scope of protection of the claims of the present invention.

[0062] An embodiment of the present invention also provides a computer-readable storage medium capable of implementing all the steps in the method in the above embodiment, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all the steps in the method in the above embodiment are implemented.

[0063] An embodiment of the present invention also provides an electronic device for executing the above method. As an implementation device of the method, the electronic device has at least a processor and a memory, and in particular, the memory stores data and related computer programs required for executing the method, such as storing corresponding algorithm models, etc., and the processor calls the data and programs in the memory to execute all the steps of the implementation method and obtains the corresponding technical effect.

[0064] Preferably, the electronic device may include a bus architecture, the bus may include any number of interconnected buses and bridges, and the bus will include various circuits linked together by one or more processors and memories. The bus may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are all well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the receiver and transmitter. The receiver and transmitter may be the same component, i.e., a transceiver, providing a unit for communicating with various other systems over a transmission medium. The processor is responsible for managing the bus and general processing, while the memory may be used to store data used by the processor when performing operations.

[0065] In addition, the electronic device may further include components such as a communication module, an input unit, an audio processor, a display, and a power supply. The processor (or controller, operation control) used may include a microprocessor or other processor device and / or a logic device, which receives input and controls the operation of various components of the electronic device; the memory may be one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory, or other suitable devices, which may store the above-mentioned related data information, and may also store a program for executing related information, and the processor may execute the program stored in the memory to realize information storage or processing, etc.; the input unit is used to provide input to the processor, for example, it may be a key or a touch input device; the power supply is used to provide power to the electronic device; the display is used to display display objects such as images and text, for example, it may be an LCD display. The communication module is a transmitter / receiver that sends and receives signals via an antenna. The communication module (transmitter / receiver) is coupled to the processor to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal. Based on different communication technologies, multiple communication modules may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless local area network module, etc. The communication module (transmitter / receiver) is also coupled to a speaker and a microphone via an audio processor to provide an audio output via the speaker and receive an audio input from the microphone, thereby realizing a common telecommunication function. The audio processor may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor is also coupled to a central processor, so that recording can be performed on the machine through the microphone, and the sound stored on the machine can be played through the speaker.

[0066] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or multiple boxes.

[0068] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction system, which is implemented in the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps of the functions specified in a box or multiple boxes. Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the attached claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0070] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent speech recommendation method based on a large-scale language model, characterized in that: include: S1. The customer communication experience is sorted into an internal knowledge base of the enterprise. The large-scale language model completes training and parameter adjustment by learning the internal knowledge base of the enterprise. S2. Analyze the internal knowledge base of the enterprise using the large-scale language model, generate customer tags and intelligent speech, and update the internal knowledge base of the enterprise according to the newly added scenario information; S3. Establish a server to realize the rapid response and update of the intelligent speech and customer tags; S4. Realize real-time interaction with customers and solve customer problems and / or promote products by calling the customer tags through the intelligent speech.

2. The method according to claim 1, characterized in that The S3 establishes a server to realize the rapid response and update of the intelligent speech and customer tags, including a request receiving module, a message processing module, and a data calculation module. The request receiving module creates a Facade class and encapsulates the internal calling logic. After receiving the scene information, the request receiving module forwards the scene information to the message processing module.

3. The method according to claim 2, characterized in that The message processing module defines the subject and message format through the message middleware, receives the message scenario field sent to the subject, and selects the correct policy processing class according to the received message scenario field.

4. The method according to claim 3, characterized in that The data calculation module receives the intelligent speech data sent by the message processing module, calculates the priority of the intelligent speech and / or customer tags, obtains the recommended intelligent speech and / or customer tags returned by the large-scale language model based on the configuration information of the intelligent speech and / or the customer tags, assembles and feeds back.

5. The method according to claim 3, characterized in that The policy processing class implements the corresponding logic between processing the message scenario field and the intelligent speech and / or customer label.

6. The method according to claim 1, characterized in that The server also includes a configuration information loading module, which loads the configuration information of the intelligent speech into the Redis cache. When the configuration information is modified, a delayed double deletion mechanism is used to ensure the accuracy of the configuration information in the Redis cache.

7. An intelligent speech recommendation system based on a large-scale language model, characterized in that: include: The knowledge base building module is used to organize customer communication experience into an internal knowledge base of the enterprise. The large-scale language model completes training and parameter adjustment by learning the internal knowledge base of the enterprise. A customer tag generation module, used to analyze the enterprise internal knowledge base using the large-scale language model, generate customer tags and intelligent speech, and update the enterprise internal knowledge base according to the newly added scenario information; A server module, used to realize rapid response and update of the intelligent speech and customer tags; The client module is used to achieve real-time interaction with customers and solve customer problems and / or promote products by calling the customer tags through the intelligent speech.

8. The server module according to claim 7, characterized in that: It includes a request receiving module, a message processing module, and a data calculation module, wherein the request receiving module creates a Facade class and encapsulates the internal call logic, and after receiving the scene information, forwards the scene information to the message processing module; The message processing module defines the subject and message format through the message middleware, receives the message scenario field sent to the subject, and selects the correct policy processing class according to the received message scenario field; The data calculation module receives the intelligent speech data sent by the message processing module, calculates the priority of the intelligent speech and / or customer tags, obtains the recommended intelligent speech and / or customer tags returned by the large-scale language model based on the configuration information of the intelligent speech and / or the customer tags, assembles and feeds back.

9. An electronic device, characterized in that: including a processor and a memory; The memory is used to store computer programs; The processor is configured to execute the method according to any one of claims 1 to 6 by calling a stored computer program.

10. A computer program product comprising a computer program and / or instructions, characterized in that: When the computer program and / or the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Citation Information

Cited By

  • Marketing verbal skill generation method and device based on sales crown data

    CN119722124A

  • A marketing script generation method and device based on a crown data

    CN119722124B

  • Method and device for replying verbal skill based on large model

    CN121255992A