Metaverse Intelligent Interaction Training System

Abstract

The present invention belongs to the technical field of the metaverse, and provides a metaverse intelligent interaction training system, including: a first training module, which constructs a first data set according to historical data related to cross-cultural experiences, and uses the first data set to fine-tune a BERT model to obtain a cross-cultural experience analysis model; a first providing module, which uses this model to provide first cross-cultural experience data for a first customer in a virtual space constructed based on the metaverse, and collects the experience feedback data of the first customer; a second training module, which constructs a second data set according to the experience feedback data and the first cross-cultural experience data, and iteratively trains the above model based on this; a second providing module, which uses the cross-cultural experience analysis model after iterative training to generate second cross-cultural experience data for a second customer and displays it to the second customer in the virtual space. The present invention combines cross-cultural experiences with metaverse technology, enabling customers to obtain a variety of cross-cultural immersive experiences online.

Description

Technical Field

[0001] The present invention relates to the technical field of the metaverse, and more particularly, to a metaverse intelligent interaction training system. Background Art

[0002] Cross-cultural experience refers to the learning and adaptation process that individuals or groups undergo when interacting with people from different cultural backgrounds. Such experience can enhance the understanding, respect, and appreciation of different cultures, and contribute to cultivating a global perspective and cross-cultural communication skills. Existing cross-cultural experiences are mainly achieved through offline-organized cross-cultural exchange activities or by viewing cross-cultural materials such as books and videos. The offline approach has a high implementation cost on the one hand, and on the other hand, it is difficult to cover the experiences of various different types of cultures, making it difficult to meet the needs of more people for cross-cultural experiences.

[0003] The present invention is designed to combine cross-cultural experience with the metaverse, so as to more conveniently provide diverse cross-cultural experiences for more people. Summary of the Invention

[0004] To solve at least one of the above technical problems, the present invention specifically provides a metaverse intelligent interaction training system, method, electronic device, computer storage medium, and computer program product.

[0005] The present invention provides a metaverse intelligent interaction training system, which is applied to an intelligent agent. The system includes a first training module, a first providing module, a second training module, and a second providing module;

[0006] The first training module is configured to collect historical data related to cross-cultural experience, construct a first data set based on the historical data, and use the first data set to fine-tune and train a BERT model to obtain a cross-cultural experience analysis model;

[0007] The first providing module is configured to use the cross-cultural experience analysis model to provide first cross-cultural experience data for a first customer in a virtual space constructed based on the metaverse, and collect experience feedback data of the first customer;

[0008] The second training module is configured to construct a second data set based on the experience feedback data and the first cross-cultural experience data, and use the second data set to iteratively train the cross-cultural experience analysis model;

[0009] The second providing module is configured to use the cross-cultural experience analysis model after iterative training to generate second cross-cultural experience data for a second customer and display it to the second customer in the virtual space.

[0010] The present invention also provides a metaverse intelligent interaction training method based on the foregoing system, which is applied to an intelligent agent and includes the following steps:

[0011] Collect historical data related to cross-cultural experiences, construct a first data set based on the historical data, and use the first data set to fine-tune and train a BERT model to obtain a cross-cultural experience analysis model;

[0012] Use the cross-cultural experience analysis model to provide first cross-cultural experience data for a first customer in a virtual space constructed based on the metaverse, and collect the experience feedback data of the first customer;

[0013] Construct a second data set based on the experience feedback data and the first cross-cultural experience data, and use the second data set to iteratively train the cross-cultural experience analysis model;

[0014] Use the iteratively trained cross-cultural experience analysis model to generate second cross-cultural experience data for a second customer and display it to the second customer in the virtual space.

[0015] Further, the collecting historical data related to cross-cultural experiences and constructing a first data set based on the historical data includes:

[0016] Determine several data source themes related to cross-cultural experiences, and obtain the historical data related to cross-cultural experiences through web crawling and manual collection;

[0017] Classify the historical data according to population types to obtain several groups of associated data. Each group of associated data includes cross-cultural data and population type labels, and construct all the associated data into the first data set.

[0018] Further, the using the cross-cultural experience analysis model to provide first cross-cultural experience data for a first customer in a virtual space constructed based on the metaverse and collecting the experience feedback data of the first customer includes:

[0019] The cross-cultural experience analysis model receives the target cultural type and customer associated data of the first customer, calls a customer analysis model to deeply analyze the customer associated data, and obtains the population type label of the first customer;

[0020] Search for cross-cultural data according to the population type label and the target cultural type of the first customer, perform format conversion processing on the cross-cultural data to obtain the first cross-cultural experience data, and provide the first cross-cultural experience data for the first customer in a virtual space constructed based on the metaverse;

[0021] During the process of providing the first cross-cultural experience data, the first experience feedback data of the first customer is obtained through speech recognition, motion, and expression capture technologies. After the first cross-cultural experience data is provided, the research questionnaire data of the first customer is received and semantically analyzed to obtain the second experience feedback data. The first experience feedback data and the second experience feedback data are used as the experience feedback data.

[0022] Further, the customer analysis model is called to deeply analyze the customer association data to obtain the population type label of the first customer, including:

[0023] The customer association data and the population type classification rules are input into the BERT model, and the BERT model outputs the population type label of the first customer obtained after semantic recognition and classification prediction. Among them, the customer association data includes the identity information and network publishing information of the first customer, and the identity information includes the education level and work type.

[0024] Further, the second data set is constructed according to the experience feedback data and the first cross-cultural experience data, and the second data set is used to iteratively train the cross-cultural experience analysis model, including:

[0025] Determine the cross-cultural experience type applicable to the cross-cultural experience analysis model, and determine the training quantity according to the cross-cultural experience type. Among them, the cross-cultural experience type includes cross-country cultural experience and cross-region cultural experience.

[0026] The experience feedback data and the first cross-cultural experience data are constructed into a training data and input into the second data set. When the amount of training data in the second data set reaches the training quantity, the second data set is used to iteratively train the cross-cultural experience analysis model.

[0027] Further, determining the training quantity according to the cross-cultural experience type includes:

[0028] If the cross-cultural experience type applicable to the cross-cultural experience analysis model is cross-country cultural experience, set the training quantity to the first value; if the cross-cultural experience type applicable to the cross-cultural experience analysis model is cross-region cultural experience, set the training quantity to the second value.

[0029] Among them, the first value is greater than the second value.

[0030] The present invention also provides an electronic device, including: a memory storing executable program code; a processor coupled to the memory; the processor invoking the executable program code stored in the memory to execute the method described in any one of the preceding items.

[0031] The present invention also provides a computer storage medium, on which a computer program is stored, and when the computer program is run by a processor, it executes the method described in any one of the above items.

[0032] The present invention also provides a computer program product, which contains computer program instructions, and when the computer program instructions are executed by a processor of an electronic device, the method described in any one of the above items is implemented.

[0033] The beneficial effects of the present invention are as follows:

[0034] 1) The present invention organically combines cross-cultural experiences with metaverse technologies, enabling customers to obtain immersive cross-cultural experiences online and effectively reducing the costs of cross-cultural experiences.

[0035] 2) The present invention uses the existing BERT large language model to build a cross-cultural experience analysis model, thereby obtaining an intelligent agent that can accurately provide cross-cultural experience data for customers. Compared with the existing methods that can only provide fixed cross-cultural experience resources, it allows customers to obtain more personalized experiences.

[0036] 3) The present invention also sets up iterative upgrade training for the cross-cultural experience analysis model to ensure that the cross-cultural experience analysis model can always analyze and obtain the most accurate cross-cultural experience data, thereby providing the best cross-cultural experience for customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0038] Figure 1 It is a schematic structural diagram of a metaverse intelligent interaction training system disclosed in an embodiment of the present invention;

[0039] Figure 2 It is a schematic flowchart of a metaverse intelligent interaction training method disclosed in an embodiment of the present invention;

[0040] Figure 3 It is a schematic structural diagram of an intelligent agent disclosed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The following describes exemplary embodiments of the present disclosure in conjunction with the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, descriptions of well-known functions and structures are omitted below for clarity and conciseness.

[0042] Referring to Figure 1 As shown, a metaverse intelligent interaction training system according to an embodiment of the present invention is applied to an intelligent agent. The system includes a first training module, a first providing module, a second training module, and a second providing module;

[0043] The first training module is configured to collect historical data related to cross-cultural experiences, construct a first data set based on the historical data, and use the first data set to fine-tune and train a BERT model to obtain a cross-cultural experience analysis model;

[0044] The first providing module is configured to use the cross-cultural experience analysis model to provide first cross-cultural experience data for a first customer in a virtual space constructed based on the metaverse and collect experience feedback data of the first customer;

[0045] The second training module is configured to construct a second data set based on the experience feedback data and the first cross-cultural experience data, and use the second data set to iteratively train the cross-cultural experience analysis model;

[0046] The second providing module is configured to use the iteratively trained cross-cultural experience analysis model to generate second cross-cultural experience data for a second customer and display it to the second customer in the virtual space.

[0047] Referring to Figure 2 As shown, an embodiment of the present invention provides a metaverse intelligent interaction training method applied to an intelligent agent, including the following steps:

[0048] Collect historical data related to cross-cultural experiences, construct a first data set based on the historical data, and use the first data set to fine-tune and train a BERT model to obtain a cross-cultural experience analysis model;

[0049] Use the cross-cultural experience analysis model to provide first cross-cultural experience data for a first customer in a virtual space constructed based on the metaverse and collect experience feedback data of the first customer;

[0050] constructing a second data set according to the experience feedback data and the first cross-cultural experience data, and iteratively training the cross-cultural experience analysis model using the second data set;

[0051] The iteratively trained cross-cultural experience analysis model is used to generate second cross-cultural experience data for the second customer, and the data is presented to the second customer in the virtual space.

[0052] A virtual space for cross-cultural experience is constructed based on the metaverse, and customers can enter the virtual space through virtual reality equipment. The present invention configures an intelligent agent for providing cross-cultural experience data for entering customers in the virtual space, and the data providing function is realized by the cross-cultural experience analysis model preset in the intelligent agent. Before the intelligent agent is arranged, historical data related to the cross-cultural experience is first collected and a first data set is constructed based on it, and the first data set is used to fine-tune the existing BERT model to obtain a preliminary cross-cultural experience analysis model. The intelligent agent is arranged, and the preliminary cross-cultural experience analysis model is used to provide the first cross-cultural experience data for the customer entering the virtual space, and the experience feedback data of the first customer is collected at the same time. According to the experience feedback data and the first cross-cultural experience data, a second data set can be constructed, and the cross-cultural experience analysis model is trained for multiple iterations using the second data set, that is, the cross-cultural experience analysis model is gradually upgraded, so that more suitable cross-cultural experience data can be provided for the second customer, and a better cross-cultural experience can be obtained.

[0053] It should be noted that the cross-cultural experience in the present invention includes both cross-national cultural experience, such as cultural experience of African culture, Southeast Asian culture, etc.; and cross-regional cultural experience, such as cultural experience of Jiangnan culture and Yunnan-Guizhou culture in China.

[0054] Furthermore, the collecting of historical data related to the cross-cultural experience and constructing a first data set based on the historical data include:

[0055] Determine several data source topics related to cross-cultural experience, and obtain the historical data related to cross-cultural experience by web crawling and manual collection;

[0056] The historical data is classified and processed according to the population type to obtain several groups of related data, each group of related data includes cross-cultural data and population type labels, and all the related data are constructed as the first data set.

[0057] In this embodiment, several data source topics related to cross-cultural experience are manually specified in advance, such as the Internet, newspapers, video data on various online platforms during holidays (such as the celebration activity videos released on the Douyin platform during the Chinese New Year), etc. Taking the above data source topics as the entry point, web crawler technology is used to crawl historical data related to cross-cultural experience from these channels. At the same time, data that is difficult to obtain through web crawling can also be obtained through manual collection, such as a large amount of paper data stored in libraries and archives. Integrating the data obtained through the above two channels can obtain historical data.

[0058] Next, the above-mentioned historical data obtained is classified by population. There can be multiple criteria for population classification, such as classification according to educational level (below university level, university level, postgraduate level, etc.), classification according to job type (technical work, cultural work, farming / breeding work, etc.), etc. After classification processing, multiple sets of associated data can be obtained. Each set of associated data includes cross-cultural data and population type labels. Each set of associated data represents the cross-cultural data that a specific type of population is interested in. The population type labels therein are set by an automatic annotation device or manually. After processing such as deduplication and deletion of error data on the sorted associated data, it is integrated into the first data set.

[0059] Furthermore, using the cross-cultural experience analysis model to provide the first cross-cultural experience data for the first customer in the virtual space constructed based on the metaverse, and collecting the experience feedback data of the first customer, including:

[0060] The cross-cultural experience analysis model receives the target cultural type and customer associated data of the first customer, and calls the customer analysis model to deeply analyze the customer associated data to obtain the population type label of the first customer;

[0061] Search for cross-cultural data according to the population type label and the target cultural type of the first customer, perform format conversion processing on the cross-cultural data to obtain the first cross-cultural experience data, and provide the first cross-cultural experience data for the first customer in the virtual space constructed based on the metaverse;

[0062] During the process of providing the first cross-cultural experience data, obtain the first experience feedback data of the first customer through voice recognition, motion and expression capture technologies, and after the first cross-cultural experience data is provided, receive and perform semantic analysis on the survey questionnaire data of the first customer to obtain the second experience feedback data, and use the first experience feedback data and the second experience feedback data as the experience feedback data.

[0063] In this embodiment, the cross-cultural experience analysis model can receive the target cultural type of the first customer and customer-related data input, where the target cultural type can be manually input by the first customer, such as African culture, Jiangnan culture, etc., and the customer-related data can be obtained by retrieving according to the identity information of the first customer. Refer to Figure 3 As shown, a customer analysis model is also set in the intelligent agent. The cross-cultural experience analysis model can call this customer analysis model, and this model can deeply analyze the above customer-related data to obtain the population type label of the first customer. Based on the determined population type label of the first customer and the specified target cultural type, the corresponding cross-cultural data can be searched. This search can be from the Internet or from a database connected to the intelligent agent. It should be noted that the cross-cultural data obtained from the Internet also needs to be converted into an adaptation format in the virtual space. Mainly, cross-cultural data is extracted from the obtained data and then converted into a virtual character image. The virtual character image outputs the cross-cultural data to the customer. For example, if the cross-cultural data obtained from the search is an opera, an opera virtual character is obtained after conversion, and the customer interacts with the opera virtual character, and the opera virtual character sings the opera; and what is stored in the database is the cross-cultural experience data that has completed the format conversion process.

[0064] In the process of providing the first cross-cultural experience data to the first customer, the satisfaction degree of the first customer with the provided first cross-cultural experience data, that is, the first experience feedback data, can be obtained by extracting from the voice, body movements, and facial expressions of the first customer; and the second experience feedback data can be obtained by performing semantic analysis on the research questionnaire data of the first customer's subsequent feedback. The first experience feedback data and the second experience feedback data are integrated into the experience feedback data, and the experience feedback data is used to construct the second data set, and the second data set is used to iteratively train the cross-cultural experience analysis model.

[0065] It should be noted that the cross-cultural experience data includes data in various cultural display forms. For example, for opera culture, it includes singing scene data, costume data, opera history data, opera experience data, etc., and one or more of them are selected according to different population type labels of customers.

[0066] Further, the calling of the customer analysis model to deeply analyze the customer-related data to obtain the population type label of the first customer includes:

[0067] Input the customer association data and the population type classification rules into the BERT model, and the BERT model outputs the population type label of the first customer obtained after semantic recognition and classification prediction; wherein, the customer association data includes the identity information and network publication information of the first customer, and the identity information includes the education level and work type.

[0068] In this embodiment, the present invention re-uses the existing BERT model without the need for fine-tuning training, that is, the original BERT model is used to perform semantic recognition and classification prediction on the customer association data and the population type classification rules, so as to obtain the population type label of the first customer. Among them, the population type classification rule is a method and standard set by humans for classifying customers. For example, the aforementioned classification according to the education level and classification according to the work type, etc. The population type classification rule can guide the customer analysis model to accurately analyze the population type label of the first customer.

[0069] The customer association data refers to the identity information and network publication information of the first customer. Among them, the identity information includes the education level and work type, and the network publication information mainly refers to the cross-cultural related information published by the first customer on various network platforms. The customer analysis model can determine multiple general population type labels of the population with the same identity based on the identity information, and then screen out more appropriate population type labels considering the personalized hobbies of the first customer.

[0070] Among them, the Transformer deep learning model is based on the self-attention mechanism and is good at processing sequence data, such as text or time series. The present invention preferably constructs a customer analysis model based on the Transformer deep learning model, which can realize accurate semantic analysis and prediction of the population type label of the first customer.

[0071] Further, constructing a second data set according to the experience feedback data and the first cross-cultural experience data, and using the second data set to perform iterative training on the cross-cultural experience analysis model, includes:

[0072] Determine the cross-cultural experience type applicable to the cross-cultural experience analysis model, and determine the training quantity according to the cross-cultural experience type; wherein, the cross-cultural experience type includes cross-national cultural experience and cross-regional cultural experience;

[0073] Construct the experience feedback data and the first cross-cultural experience data into a piece of training data and input it into the second data set. When the amount of training data in the second data set reaches the training quantity, use the second data set to perform iterative training on the cross-cultural experience analysis model.

[0074] In this embodiment, when retraining the cross-cultural experience analysis model, it is necessary to consider whether the amount of training data is appropriate. Only a sufficient amount of training data can ensure that the cross-cultural experience analysis model accurately analyzes the cross-cultural experience data adapted to different types of customers. In this regard, the present invention sets the cross-cultural experience types to include the aforementioned cross-national cultural experience and cross-regional cultural experience. Correspondingly, cross-cultural experience analysis models are configured for the above two cross-cultural experience types respectively. The training quantity corresponding to each cross-cultural experience analysis model is determined according to the cross-cultural experience type, so that when the amount of training data in the second data set reaches this training quantity, the retraining of this cross-cultural experience analysis model can be started.

[0075] Further, the determining the training quantity according to the cross-cultural experience type includes:

[0076] If the cross-cultural experience type applicable to the cross-cultural experience analysis model is cross-national cultural experience, set the training quantity to a first value; if the cross-cultural experience type applicable to the cross-cultural experience analysis model is cross-regional cultural experience, set the training quantity to a second value;

[0077] Wherein, the first value is greater than the second value.

[0078] As mentioned above, the present invention sets the cross-cultural experience types to cross-national cultural experience or cross-regional cultural experience. For the former, customers generally have less understanding of the cultures of other countries, and the first cross-cultural experience data determined by the cross-cultural experience analysis model is more likely to be mismatched with customers, resulting in the experience feedback data of customers being prone to contain negative content. Therefore, when the cross-cultural experience analysis model is applicable to cross-national cultural experience, the present invention uses more training data to self-train the cross-cultural experience analysis model. In this way, the trained cross-cultural experience analysis model can achieve a more thorough understanding of the demand laws of customers, so as to output more accurate cross-cultural experience data. For the latter, customers generally have more understanding of the cultures of different regions in the country (mainly compared with cross-national cultures), and the first data set obtained through the aforementioned multiple data sources can already achieve a relatively sufficient training of the cross-cultural experience analysis model. Therefore, when the cross-cultural experience analysis model is applicable to cross-regional cultural experience, the present invention sets to use less training data to self-train the cross-cultural experience analysis model, which can effectively reduce the data load and training duration of self-training and improve the self-training efficiency. For example, the first value is 500 pieces and the second value is 200 pieces.

[0079] An embodiment of the present invention also discloses an electronic device, including: a memory storing executable program code; a processor coupled to the memory; the processor invoking the executable program code stored in the memory to execute the method as described in the foregoing embodiment.

[0080] An embodiment of the present invention also discloses a computer storage medium, on which a computer program is stored, and when the computer program is run by a processor, it executes the method as described in the foregoing embodiment.

[0081] An embodiment of the present invention also discloses a computer program product, which includes computer program instructions, and when the computer program instructions are executed by a processor of an electronic device, the method as described in any one of the above is implemented.

[0082] The various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs, which can be executed and / or interpreted on a programmable system including at least one programmable processor. The programmable processor can be a dedicated or general-purpose programmable processor, and can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0083] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable load balancing devices, so that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program codes can be executed entirely on the machine, partially on the machine, executed partially on the machine as an independent software package and partially on a remote machine, or executed entirely on a remote machine or server.

[0084] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0085] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in this disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution disclosed in this disclosure can be achieved, and no limitation is imposed herein.

[0086] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A Metaverse Intelligent Interactive Training System, applied to an intelligent agent, characterized in that: The system comprises a first training module, a first providing module, a second training module, and a second providing module; The first training module is used to collect historical data related to cross-cultural experience, construct a first data set based on the historical data, and use the first data set to fine-tune the BERT model to obtain a cross-cultural experience analysis model; The first providing module is used to use the cross-cultural experience analysis model to provide first cross-cultural experience data for the first customer in a virtual space built based on the metaverse, and collect experience feedback data of the first customer, including: The cross-cultural experience analysis model receives the target culture type and customer-related data of the first customer, and calls the customer analysis model to perform in-depth analysis on the customer-related data to obtain a population type label of the first customer; Searching for cross-cultural data based on the first customer's population type label and the target culture type, performing format conversion processing on the cross-cultural data to obtain the first cross-cultural experience data, and providing the first cross-cultural experience data to the first customer in a virtual space constructed based on the metaverse; obtaining first experience feedback data of the first customer through voice recognition, motion and expression capture technology during the process of providing the first cross-cultural experience data, and receiving and performing semantic analysis on the survey questionnaire data of the first customer after the first cross-cultural experience data is provided, obtaining second experience feedback data, and using the first experience feedback data and the second experience feedback data as the experience feedback data; The second training module is used to construct a second data set according to the experience feedback data and the first cross-cultural experience data, and use the second data set to iteratively train the cross-cultural experience analysis model; The second providing module is used to generate second cross-cultural experience data for a second customer using the iteratively trained cross-cultural experience analysis model, and present the second cross-cultural experience data to the second customer in a virtual space.

2. A metaverse intelligent interactive training method based on the system of claim 1, applied to an intelligent agent, characterized in that: The steps include: Collect historical data related to cross-cultural experience, construct a first data set based on the historical data, and use the first data set to fine-tune the BERT model to obtain a cross-cultural experience analysis model; Using the cross-cultural experience analysis model to provide first cross-cultural experience data to a first customer in a virtual space built based on the metaverse, and collecting experience feedback data from the first customer, including: The cross-cultural experience analysis model receives the target culture type and customer-related data of the first customer, and calls the customer analysis model to perform in-depth analysis on the customer-related data to obtain a population type label of the first customer; Searching for cross-cultural data based on the first customer's population type label and the target culture type, performing format conversion processing on the cross-cultural data to obtain the first cross-cultural experience data, and providing the first cross-cultural experience data to the first customer in a virtual space constructed based on the metaverse; obtaining first experience feedback data of the first customer through voice recognition, motion and expression capture technology during the process of providing the first cross-cultural experience data, and receiving and performing semantic analysis on the survey questionnaire data of the first customer after the first cross-cultural experience data is provided, obtaining second experience feedback data, and using the first experience feedback data and the second experience feedback data as the experience feedback data; constructing a second data set according to the experience feedback data and the first cross-cultural experience data, and iteratively training the cross-cultural experience analysis model using the second data set; The iteratively trained cross-cultural experience analysis model is used to generate second cross-cultural experience data for the second customer, and the data is presented to the second customer in the virtual space.

3. A metaverse intelligent interactive training method according to claim 2, characterized in that: Collecting historical data related to cross-cultural experience, and constructing a first data set based on the historical data, including: Determine several data source topics related to cross-cultural experience, and obtain the historical data related to cross-cultural experience by web crawling and manual collection; The historical data is classified and processed according to the population type to obtain several groups of related data, each group of related data includes cross-cultural data and population type labels, and all the related data are constructed as the first data set.

4. A metaverse intelligent interactive training method according to claim 3, characterized in that: The customer analysis model is called to perform in-depth analysis on the customer-related data to obtain a population type label of the first customer, including: The customer association data and population type classification rules are input into the BERT model, and the BERT model outputs the population type label of the first customer obtained after semantic recognition and classification prediction; wherein the customer association data includes the identity information and network publishing information of the first customer, and the identity information includes education level and job type.

5. A metaverse intelligent interactive training method according to claim 4, characterized in that: Constructing a second data set according to the experience feedback data and the first cross-cultural experience data, and iteratively training the cross-cultural experience analysis model using the second data set, including: Determine the type of cross-cultural experience to which the cross-cultural experience analysis model is applicable, and determine the amount of training according to the type of cross-cultural experience; wherein the type of cross-cultural experience includes cross-national cultural experience and cross-regional cultural experience; The experience feedback data and the first cross-cultural experience data are constructed into a piece of training data and are input into the second data set. When the amount of training data in the second data set reaches the training quantity, the cross-cultural experience analysis model is iteratively trained using the second data set.

6. A Metaverse Intelligent Interaction Training Method according to claim 5, characterized in that: The amount of training will be determined based on the type of cross-cultural experience being described, and may include: If the cross-cultural experience type to which the cross-cultural experience analysis model is applicable is a cross-country cultural experience, the training quantity is set to a first value; if the cross-cultural experience type to which the cross-cultural experience analysis model is applicable is a cross-regional cultural experience, the training quantity is set to a second value; Wherein, the first value is greater than the second value.

7. An electronic device comprising: A memory storing executable program code; A processor coupled to the memory; characterized in that: the processor calls the executable program code stored in the memory to execute the method according to any one of claims 2-6.

8. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 2 to 6 is executed.

9. A computer program product, comprising computer program instructions, characterized in that: When the computer program instructions are executed by a processor of an electronic device, the method according to any one of claims 2 to 6 is implemented.

Images