Resource recommendation processing method and device

Abstract

The embodiment of the invention provides a resource recommendation processing method and device.The resource recommendation processing method comprises the steps that in the process that an access user accesses a resource service, recommendation decision data is generated according to access data of the access user and user data, and inputting the recommendation decision data into a large language model to perform recommendation action decision to obtain a recommendation action, further calling an agent corresponding to the recommendation action in the multiple agents to perform resource recommendation processing to obtain resource recommendation data, and displaying the resource recommendation data to the access user. Therefore, resource recommendation is carried out on the access user in the resource access process of the access user through cooperation of the large language model and the multiple agents.

Description

Technical Field

[0001] This document relates to the field of data processing technology, and in particular to a resource recommendation processing method and device. Background Art

[0002] With the continuous development and promotion of the Internet and artificial intelligence, many services can be automated through online large language models. For example, in the scenario where users access online services, online services can make targeted recommendations to users based on the deployed large language models. However, as more and more online services make access recommendations based on large language models, the competition faced by all parties is becoming increasingly fierce. In this case, higher requirements are placed on access recommendations made through large language models. Summary of the invention

[0003] One or more embodiments of this specification provide a resource recommendation processing method, including: obtaining access data obtained after the access user of the resource service triggers the service identifier, and generating recommendation decision data based on the access data and user data. The recommendation decision data is input into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action. The agent corresponding to the recommended action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0004] One or more embodiments of the present specification provide a resource recommendation processing device, including: a data generation module, configured to obtain access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data. An action decision module, configured to input the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action. A recommendation processing module, configured to call the agent corresponding to the recommended action in the multi-agent to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0005] One or more embodiments of the present specification provide a resource recommendation processing device, including: a processor; and a memory configured to store computer executable instructions, wherein when the computer executable instructions are executed, the processor: obtains access data obtained after the access user of the resource service triggers the service identifier, and generates recommendation decision data based on the access data and user data. The recommendation decision data is input into a large language model for recommendation action evaluation and recommendation action decision to obtain a recommended action. The agent corresponding to the recommended action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0006] One or more embodiments of this specification provide a computer-readable storage medium for storing computer-executable instructions, which implement the following process when executed: obtain access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data. Input the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action. Call the agent corresponding to the recommended action in the multi-agent to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate one or more embodiments of the present specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present specification. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor. Figure 1 A schematic diagram of an implementation environment of a resource recommendation processing method provided in one or more embodiments of this specification; Figure 2 A processing flow chart of a resource recommendation processing method provided by one or more embodiments of this specification; Figure 3 A processing flow chart of a resource recommendation processing method applied to a resource service scenario provided in one or more embodiments of this specification; Figure 4 A schematic diagram of an embodiment of a resource recommendation processing device provided in one or more embodiments of this specification; Figure 5 A schematic diagram of the structure of a resource recommendation processing device provided in one or more embodiments of this specification. DETAILED DESCRIPTION

[0008] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0009] The resource recommendation processing method provided in one or more embodiments of this specification can be applied to the implementation environment of the resource service recommendation system. Figure 1 , the implementation environment at least includes: User terminal 101 and server 102; The user terminal 101 is used to access the resource service and perform corresponding recommendation interaction by triggering the service identifier during the service access process, including submitting questions to the server 102, receiving and displaying the answers or questions returned by the server 102; the user terminal 101 can be a mobile phone, a personal computer, a tablet computer, an e-book reader, a device for information interaction based on VR (Virtual Reality, virtual reality technology), a vehicle terminal, an IoT device, a wearable smart device, a laptop portable computer and a desktop computer, etc.; Server 102 is used to generate recommendation data for input into a large language model during the process of a visiting user accessing a resource service, and to recommend resource services to the visiting user through the large language model and multiple agents. Server 102 can be a single server, or a server cluster consisting of several servers, or one or more cloud servers in a cloud computing platform.

[0010] In this implementation environment, during the process of a visiting user accessing a resource service through the user terminal 101, when the visiting user triggers the service identification displayed on the user terminal 101, the server 102 generates recommendation decision data based on the access data and user data of the visiting user, and inputs the recommendation decision data into the large language model for recommendation action evaluation and recommendation action decision, and further based on the recommended action obtained by the decision, calls the agent corresponding to the recommended action in the multi-agent to perform resource recommendation processing to obtain resource recommendation data, and sends the resource recommendation data to the user terminal 101, and displays the resource recommendation data in the interactive area after the service identification is triggered on the user terminal 101, so that in the process of the visiting user accessing the resource service, resource recommendation is realized through the cooperation of the large language model and the multi-agent, and more flexible recommendation interaction is performed with the visiting user by triggering the service identification and displaying the resource recommendation data in the interactive area after the service identification is triggered.

[0011] It should be noted that, considering that the user data, access data, conversation data and other related data of the visiting user involved in this specification may belong to the privacy of the visiting user to a certain extent, therefore, if you want to collect the user data, access data, conversation data and other related data of the visiting user, you can obtain the authorization of the visiting user before collecting the data, so that the operation of collecting data complies with relevant data management regulations. For example, data authorization can be performed when the visiting user accesses the resource service for the first time, or during any of the visiting user's access to the resource service, or when the visiting user triggers the service identification. The specific method of data authorization can be to send a user data authorization reminder to the visiting user, and the visiting user can obtain the data authorization of the visiting user after confirming the reminder through an instruction. Alternatively, the method of data authorization can also be to obtain the data authorization of the visiting user by signing a data authorization agreement.

[0012] One or more embodiments of a resource recommendation processing method provided in this specification are as follows: Reference Figure 2 The resource recommendation processing method provided in this embodiment specifically includes steps S202 to S206.

[0013] Step S202: obtaining access data obtained after the access user of the resource service triggers the service identifier, and generating recommendation decision data according to the access data and user data.

[0014] The resource service described in this embodiment refers to the resource-related services provided within the application, such as the resource services related to security projects provided by the application. The accessing users can consult, inquire, purchase and / or sell the purchased security projects by accessing the resource services, and can also apply for security projects or other processing related to the security projects through the resource services; the security projects can be resource management security projects that guarantee the user's resource management in the application, such as financial security projects.

[0015] The service identifier refers to the identifier configured by the resource service for interacting with the visiting user. When the service identifier is triggered, the interactive area will be displayed, and the visiting user can make corresponding consultations, inquiries and / or purchases in the interactive area. In addition, the resource service can also recommend corresponding resources to the visiting user in the interactive area. Optionally, the service identifier includes the interactive identifier of the virtual person or digital person configured by the resource service.

[0016] In this embodiment, the display of the service logo can be determined based on the access data and / or user data of the accessing user, and can also be determined based on other data, such as service data of the resource service and / or resource data of the resource, or it can also be determined based on at least one of the access data and user data and at least one of the service data and resource data of the resource whether to display the service logo; Specifically, the display of the service identifier may correspond to the resource service or the service page of the resource service. For example, the service identifier may be displayed on a specific service page of the resource service, and the service identifier may not be displayed on other service pages other than the specific service page of the resource service. Specifically, the specific service page for displaying the service identifier may be determined by inputting access data, user data, service data of the resource service and / or resource data of the resource into a pre-trained identifier decision model or a large language model; In addition, the display of the service identifier may also correspond to the resources in the resource service, for example, the service identifier may be displayed on the resource page of a specific resource in the resource service, and the service identifier may not be displayed on the resource page of other resources other than the specific resource. Specifically, the specific resource displaying the service identifier may be determined by inputting access data, user data, service data of the resource service and / or resource data of the resource into a pre-trained identifier decision model or a large language model, or the resource page of the specific resource displaying the service identifier may be determined by inputting access data, user data, service data of the resource service and / or resource data of the resource into a pre-trained identifier decision model or a large language model; Alternatively, the display of the service logo may also correspond to the visiting user, such as displaying the service logo during the access process of a specific visiting user. Specifically, whether to display the service logo for the visiting user may be determined by inputting the access data of the visiting user, the user data, the service data of the resource service and / or the resource data of the resource into a pre-trained logo decision model or a large language model; Alternatively, the display of the service logo may also correspond to at least two of the resource service, the service page of the resource service, the resource, and the visiting user. For example, the service logo is displayed when a specific visiting user accesses the resource page of a specific resource. For another example, the service logo is displayed when a specific visiting user accesses a specific service page of the resource service. Similarly, whether the display result for the service logo is targeted can be determined by inputting the visiting user's access data, user data, service data of the resource service, and / or resource data of the resource into a pre-trained identification decision model or a large language model.

[0017] The access data refers to the data related to the access behavior of the access user obtained or generated during the access user's access to the resource service; optionally, the access data includes at least one of the following: the conversation data of the access user in the interactive area, the resource data of the current access resource of the access user in the resource service, the recommended action of the service node where the current access resource is located, and the conversation question entered by the access user in the interactive area. User data refers to the data related to the access user himself, such as the access data including the identity data, risk data, credit data and / or salary data of the access user.

[0018] During specific implementation, when the accessing user triggers the service identifier, it is necessary to perform corresponding resource recommendation interactions with the accessing user. The recommendation interaction can be performed in the interaction area displayed after the service identifier is triggered, and the access data is the data basis for performing resource recommendation interactions to the accessing user under the premise that the accessing user triggers the service identifier. To this end, the access data can be obtained after the accessing user triggers the service identifier, that is, the access data obtained after the accessing user of the resource service triggers the service identifier; specifically, the access data can be obtained from the application, and can also be obtained from the server, or the access data can also be composed of the access data obtained from the application and the access data obtained from the server.

[0019] Based on the access data, recommendation decision data is generated according to the access data and user data. The recommendation decision data refers to the data required for recommendation interaction with the visiting user, specifically the data required to recommend what kind of action to take to the visiting user during the process of the visiting user accessing resource services, that is, the decision data required to decide what kind of action to take to the visiting user.

[0020] Step S204: input the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action.

[0021] In this embodiment, a large language model is coordinated with an intelligent agent to make resource recommendations to a visiting user during the process of the visiting user accessing a resource service. Alternatively, a large language model is coordinated with an intelligent agent to interact with a visiting user about resource recommendations during the process of the visiting user accessing a resource service. The resource recommendation interaction includes a conversation interaction and / or question-and-answer interaction with the visiting user regarding the recommended resources based on the resource recommendation made by the intelligent agent. The resource recommendation interaction may also include corresponding resource recommendations, question replies and / or conversation responses to the visiting user's questions or conversations.

[0022] In specific implementation, the large language model and the intelligent agent cooperate to make resource recommendations or resource recommendation interactions during the process of users accessing resource services. The recommended action is obtained by inputting the recommendation decision data into the large language model for recommendation action evaluation and recommendation action decision, so that the intelligent agent corresponding to the recommended action can be called to perform corresponding resource recommendation processing.

[0023] In this embodiment, multiple agents are deployed for performing resource recommendation and / or resource recommendation interaction, and the multiple agents constitute a multi-agent. Each agent in the multi-agent performs corresponding resource recommendation processing by executing a corresponding recommendation action. Here, the large language model performs the process of evaluating and deciding the recommendation action, that is, the process of deciding which agent or agents in the multi-agent to call to perform resource recommendation and / or resource recommendation interaction to the visiting user. Among them, in the process of evaluating recommended actions, the large language model obtains the value evaluation results of each recommended action by performing value evaluation on each pre-set recommended action. Each pre-set recommended action can correspond to an agent in the multi-agent. Specifically, the recommended action can correspond one-to-one to each agent in the multi-agent. In this case, in the process of performing value evaluation on each recommended action, that is, in the process of performing value evaluation on the agent that executes the recommended action, the value evaluation results of each recommended action obtained reflect the adaptability of recommending resources to visiting users through each recommended action, that is, the adaptability of recommending resources to visiting users by calling each agent.

[0024] Specifically, in an optional implementation manner provided by this embodiment, the recommended action evaluation and recommended action decision include: Perform value evaluation on the recommended actions corresponding to each agent in the multi-agent system to obtain the value evaluation results of each recommended action; A recommended action is selected from among the recommended actions based on the value assessment result.

[0025] For example, in the resource services related to the security project provided by the application, the recommended actions corresponding to each agent in the multi-agent include: a consultation recommendation action recommended during the consultation process of the visiting user, an inquiry recommendation action recommended during the inquiry process of the visiting user, a payment recommendation action recommended during the payment process of the visiting user for the security project, and a purchase recommendation action recommended after the visiting user confirms the payment for the security project; by inputting the recommendation decision data into the large language model to perform value evaluation on each recommended action, the value evaluation value of each recommended action is output, and the recommended action is selected from each recommended action according to the value evaluation value.

[0026] In addition, in order to improve the sales success rate of resources in resource services, the resource status of resources in resource services can be introduced in the process of visiting users accessing resource services, so as to more intuitively perceive the sales situation of resources through changes in resource status. Specifically, the resource status of resources can be corresponded to the agents in the multi-agents, so that in the process of the large language model performing value evaluation on the recommended actions corresponding to each agent in the multi-agents, the status changes of resources accessed by visiting users in resource services can be intuitively perceived, so that the resource provider or the service platform of the resource service can provide more targeted access strategies or sales strategies based on a clearer perception of the status changes of resources in the process of visiting users accessing resources, so as to promote the successful sale of resources.

[0027] Optionally, an agent in the multi-agent corresponds to a resource state of a resource in a resource service; any resource state of a resource corresponds to one or more agents.

[0028] For example, in the resource services related to security projects provided by the application, the resource status corresponding to each agent in the multi-agent includes: the consultation status for consulting on security projects, the inquiry status for inquiring for security projects, the payment status for paying for security projects, and the purchase status for purchased security projects.

[0029] It should be pointed out that the recommended action in this embodiment can be replaced by the resource status. Specifically, the process of inputting the recommended decision data into the large language model for recommended action evaluation and recommended action decision to obtain the recommended action can be replaced by inputting the recommended decision data into the large language model for resource status evaluation and resource status decision to obtain the resource status. Correspondingly, the subsequent process of calling the agent corresponding to the recommended action in the multi-agent to perform resource recommendation processing to obtain resource recommendation data can also be replaced by calling the agent corresponding to the resource status in the multi-agent to perform resource recommendation processing to obtain resource recommendation data; the recommended action is replaced by the specific implementation method of the resource status, which is similar to the specific implementation method of the recommended action provided in this embodiment. Please refer to the implementation method of the recommended action provided in this embodiment, which will not be repeated here.

[0030] In addition, the recommendation action can also be replaced by other states related to resource sales or states related to resource access by visiting users, for example, the recommendation action can be replaced by a recommendation interaction state or a service state; the recommendation interaction state refers to the state of interacting with the visiting user in an interactive manner during the recommendation process, such as the consultation interaction state of recommending interaction with the visiting user during the consultation process of the visiting user, and the inquiry interaction state of recommending interaction with the visiting user during the inquiry process of the visiting user; the service status refers to the service status of the resource service, and the service status can also refer to the service status of the guarantee project; the specific implementation method of the recommendation action being replaced by the recommendation interaction state or the service state is similar to the specific implementation method of the recommendation action provided in this embodiment, so please refer to the implementation method of the recommendation action provided in this embodiment, which will not be repeated here.

[0031] As described above, in the scenario where resource recommendations or resource recommendation interactions are performed in the process of accessing user access to resource services by means of the cooperation between the large language model and the intelligent agent, in order to improve the accuracy of the large language model in evaluating and deciding the recommended actions, the large language model can also be trained according to the execution results of the intelligent agent, so as to continuously improve the accuracy of the large language model in evaluating and deciding the recommended actions in the process of accessing user access to resource services, thereby making the recommendation effect of the recommended actions performed by the intelligent agent more targeted and more effective; Optionally, the large language model is obtained by training the pre-trained model using a reinforcement learning method based on the training samples. Optionally, the training samples are constructed based on the state of the agent before executing the recommended action, the recommended action, the reward after the recommended action is executed, and the state after the recommended action is executed.

[0032] Here, the state before executing the recommended action may be the resource state before executing the recommended action; the state after executing the recommended action may be the resource state after the recommended action; the reward after executing the recommended action may be the reward of the resource state after the recommended action relative to the resource state before executing the recommended action. For example, the resource state before executing the recommended action is the consulting state, and the resource state before executing the recommended action is the inquiry state. In this case, the resource state evolves in the direction of expected sales success, and the reward for executing the recommended action is a positive reward. For another example, the resource state before executing the recommended action is the pending payment state, and the resource state before executing the recommended action is the consulting state. In this case, the resource state has not evolved in the direction of expected sales success, and the reward for executing the recommended action is a negative reward.

[0033] Step S206, calling the agent corresponding to the recommendation action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0034] This embodiment makes resource recommendations or interacts with resource recommendations to visiting users through the cooperation of a large language model and an intelligent agent. After obtaining the recommended action by inputting the recommendation decision data into the large language model for recommendation action evaluation and recommendation action decision, in this step, resource recommendation data is obtained by calling the intelligent agent corresponding to the recommended action in the multi-agents to perform resource recommendation processing, so as to display the resource recommendation data in the interaction area after the service logo is triggered, thereby realizing recommendations for visiting users by displaying the resource recommendation data in the interaction area after the service logo is triggered, that is, making resource recommendations to visiting users in the process of the visiting users interacting by triggering the service logo.

[0035] During specific implementation, in the process of calling the agent corresponding to the recommendation action in the multi-agents to perform resource recommendation processing, due to the differences in the agents included in the multi-agents and the differences in the recommendation actions corresponding to each agent, in the process of calling the agent to perform resource recommendation processing, in order to improve the efficiency of resource recommendation processing, thereby improving the response efficiency of recommendations to visiting users during the process of resource service access by visiting users, resource recommendation data can be obtained by deploying a corresponding resource recommendation processing model in each agent to perform resource recommendation processing, and resource recommendation data can be obtained by generating a resource recommendation text for the agent and inputting the resource recommendation text into the resource recommendation processing model to generate resource recommendation data.

[0036] In this process, by generating resource recommendation text for an intelligent agent and inputting the resource recommendation text into a resource recommendation processing model to generate resource recommendation data to obtain resource recommendation data, the resource recommendation text can be generated according to the pre-configured recommendation templates of each intelligent agent and the access data and / or user data. The resource recommendation processing model can be a large language model that is pre-trained to generate resource recommendation data. The generated resource recommendation text for each intelligent agent is input into the large language model to generate resource recommendation data to obtain resource recommendation data.

[0037] In addition, during specific implementation, when the visiting user inputs a dialogue question during the process of accessing the resource service, in the process of calling the agent corresponding to the recommended action in the multi-agent for resource recommendation processing, the agent can also be called to generate an answer to the dialogue question input by the visiting user, and recommend to the visiting user through dialogue interaction. Specifically, an optional implementation method provided by this embodiment is that if the access data contains the dialogue question input by the visiting user in the interactive area; accordingly, calling the agent corresponding to the recommended action in the multi-agent for resource recommendation processing to obtain resource recommendation data includes: calling the agent to generate an answer to the dialogue question; wherein the answer to the question is obtained by inputting the dialogue question and the access data into a large language model to generate the answer, or the answer to the question is obtained by inputting the dialogue question and the resource detail data of the accessed resource into the large language model to generate the answer, For example, after a visiting user triggers a service logo, the dialogue question entered in the interactive area displayed after the service logo is triggered is "What good insurance projects are recommended this year?"; when the access data containing the dialogue question is input into the large language model for answer generation, the generated answer to the question is "Recommend xx insurance project to you."

[0038] For another example, when a visiting user is visiting the xx security project, the dialogue question inputted is "If I invest in this security project for 3 years, how much cash can I receive in the 4th year?"; after inputting this dialogue question and the project details data of the security project currently visited by the visiting user into the big language model for answer generation, the generated answer to the question is "According to calculation, the cumulative investment in this security project for 3 years is xxx1 yuan, and the cash that can be received in the 4th year is xxx2 yuan".

[0039] Furthermore, in the process of displaying resource recommendation data in the interactive area after the service logo is triggered, specifically on the basis of displaying the answer to the question in the interactive area after the service logo is triggered, in an optional implementation provided by this embodiment, if the access control of the recommended resource carried in the answer to the question is triggered, the access page of the resource service is jumped to the resource access page corresponding to the access control of the recommended resource.

[0040] In actual applications, in the process of calling an agent to perform resource recommendation processing, there may be a situation where the visiting user is not satisfied with the currently visited or recommended resources. In view of this, in order to make the resources recommended to the visiting user more compatible with the needs of the visiting user, the resources recommended to the visiting user can also be determined by a model decision method. Specifically, in an optional implementation manner provided by this embodiment, calling an agent corresponding to a recommendation action in multiple agents to perform resource recommendation processing to obtain resource recommendation data includes: The recommended resources are obtained by inputting the dialogue questions, user data and / or access data input by the access user into the large language model for resource evaluation and resource decision; Based on the recommended resource updated access data, updated access data is obtained, and dialogue is generated based on the user data and the updated access data, and the generated dialogue questions are used as resource recommendation data; or, based on the recommended resource updated access data, updated access data is obtained, and dialogue is generated based on the resource detail data and updated access data of the recommended resource, and the generated dialogue questions are used as resource recommendation data.

[0041] In addition, in actual applications, in the process of calling the agent corresponding to the recommended action to perform resource recommendation processing, one or more sub-agents can be deployed inside the agent, and more fine-grained resource recommendations can be made to the accessing user through the deployment and execution of the sub-agents. Optionally, the agent corresponding to the recommended action is deployed with at least one sub-agent, and each sub-agent corresponds to one or more sub-recommended actions; Accordingly, in an optional implementation provided by this embodiment, an agent corresponding to a recommended action in multiple agents is called to perform resource recommendation processing to obtain resource recommendation data, including: obtaining sub-recommended actions by inputting node decision data into a large language model for sub-action evaluation and sub-action decision, and performing resource recommendation processing by calling the sub-agent corresponding to the sub-recommended action deployed by the agent.

[0042] Among them, the node decision data can be composed of at least one of the access data and the user data, and can also be composed of at least one of the data contained in the access data and the user data; in addition, the node decision data input by the latter sub-agent can also be the output data of the previous sub-agent, or the data of the previous sub-agent and at least one of the access data and the user data, or at least one of the data contained in the access data and the user data.

[0043] It should be noted that the above-mentioned several implementation methods for calling the agent corresponding to the recommended action in the multi-agent to perform resource recommendation processing to obtain resource recommendation data, in the specific implementation process, any one implementation method can be selected for implementation according to actual needs, it should be noted that the above-mentioned several implementation methods can also be combined as a whole of any two or three according to actual needs, and corresponding technical means can be selected from several implementation methods according to actual needs to form a new implementation method. For example, calling the agent corresponding to the recommended action in the multi-agent to perform resource recommendation processing to obtain resource recommendation data includes: inputting the dialogue question input by the access user, user data and / or access data into the large language model for resource evaluation and resource decision-making to obtain recommended resources; inputting the dialogue question and the access data and / or recommended resources into the large language model for answer generation to obtain the answer to the question, and using the recommended resources and the answer to the question as the resource recommendation data; For another example, calling an agent corresponding to a recommendation action in a multi-agent to perform resource recommendation processing to obtain resource recommendation data includes: calling a first sub-agent of the agent corresponding to the recommendation action to perform resource evaluation and resource decision-making to obtain recommended resources; calling a second sub-agent of the agent to generate answers to obtain answers to questions, and using the recommended resources and the answers to questions as resource recommendation data; Here, the input for calling the first sub-agent may be a dialogue question, user data and / or access data input by the access user, and the first sub-agent may obtain recommended resources by inputting the dialogue question, user data and / or access data into a large language model for resource evaluation and resource decision-making; the input for calling the second sub-agent may be a dialogue question and access data and / or recommended resources, and the second sub-agent may obtain the answer to the question by inputting the dialogue question and access data and / or recommended resources into a large language model for answer generation.

[0044] In this embodiment, in order to improve the conversion rate of resource recommendations and thus improve the success rate of resource sales, corresponding multi-agents can be configured with resources in the resource service as the core. In this case, the multi-agents configured for different resources may be different. When the visiting user accesses different resources, the resource recommendation processing performed by the multi-agent in cooperation with the large language model to the visiting user may also be different. Optionally, the resources in the resource service are configured with their own multi-agents; the agents in the multi-agent of the resource are determined according to the configuration of the resource provider, and / or the sub-agents deployed by the agents in the multi-agent of the resource are determined according to the configuration of the resource provider.

[0045] It should be pointed out that the above implementation method provides multiple large language models, specifically including: a large language model that inputs recommendation decision data to perform recommendation action evaluation and recommendation action decision, a large language model that inputs access data, user data, service data of resource services and / or resource data of resources to determine a specific service page for displaying a service identifier, a large language model that inputs resource recommendation text to generate resource recommendation data, and a large language model that inputs dialogue questions and access data to generate answers. These four large language models can be the same large language model or different large language models. In the case where these four are different large language models, the large language model that performs recommendation action evaluation and recommendation action decision can be referred to as the first large language model. The big language model or the decision big language model, and so on, determine the big language model of the specific service page that displays the service logo, the big language model that generates resource recommendation data, and the big language model that generates answers, which can be named the second large language model, the third large language model, or the fourth large language model according to actual needs; further, it should be pointed out that, in addition to the input data of the four large language models provided above, the input data of any one or more of the four large language models can also be replaced by other corresponding data according to actual needs, such as being replaced by: recommendation decision data, access data, user data, service data of resource services, resource data of resources, dialogue questions, and at least one of resource recommendation texts.

[0046] To summarize, the resource recommendation processing method provided in this embodiment generates recommendation decision data based on the access data and user data of the access user when the accessing user triggers the service identifier in the process of accessing the resource service, and makes resource recommendations or resource recommendation interactions to the accessing user in the process of the accessing user accessing the resource service by cooperating with the large language model and the intelligent agent. Specifically, the recommendation decision data is input into the large language model for recommendation action evaluation and recommendation action decision to obtain the recommendation action, and the resource recommendation data is obtained by calling the intelligent agent corresponding to the recommendation action in the multi-intelligent agent for resource recommendation processing, so as to display the resource recommendation data in the interaction area after the service identifier is triggered, thereby realizing recommendations for the accessing user by displaying the resource recommendation data in the interaction area after the service identifier is triggered, that is, making resource recommendations to the accessing user in the process of the accessing user interacting by triggering the service identifier, which helps to improve the conversion rate of resource recommendations, and thereby improve the success rate of resource sales.

[0047] The following takes the application of a resource recommendation processing method provided in this embodiment in a resource service scenario as an example. Figure 3 , for further explanation of the resource recommendation processing method provided in this embodiment, see Figure 3 , a resource recommendation processing method applied to a resource service scenario specifically includes the following steps.

[0048] Step S302: obtaining access data of users accessing the resource service, and generating recommendation decision data according to the access data and user data.

[0049] Step S304: input the recommendation decision data into the large language model to perform recommendation action evaluation and recommendation action decision to obtain the first recommended action.

[0050] Step S306, calling the first agent corresponding to the first recommendation action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0051] The resource recommendation process includes obtaining recommended resources by inputting the recommendation decision data into a large language model for resource evaluation and resource decision making.

[0052] Step S308: If the dialogue question submitted by the visiting user in the interactive area is obtained, the dialogue question is updated to the visiting data to obtain updated visiting data.

[0053] Step S310: input the recommendation decision data generated by the user data and the updated access data into the large language model to perform recommendation action evaluation and recommendation action decision, and obtain a second recommended action.

[0054] Step S312, calling the second agent corresponding to the second recommended action in the multi-agent to generate an answer to the dialogue question, so as to display the answer in the interactive area after the service logo is triggered.

[0055] It should be noted that any one of steps S302 to S312 or any combination of multiple steps can be combined with any one of steps S202 to S206 to form a new implementation method according to the needs of implementation deployment; in addition, according to the needs of actual deployment, any one or multiple technical features can be selected from steps S302 to S312 and combined with any one or multiple technical features provided by steps S202 to S206 to form a new implementation method; or, any one or multiple technical features in steps S302 to S312 can be replaced with any one or multiple technical features provided by steps S202 to S206 to form a new implementation method according to the needs of actual deployment, which will not be repeated here.

[0056] An embodiment of a resource recommendation processing device provided in this specification is as follows: In the above-mentioned embodiment, a resource recommendation processing method is provided, and correspondingly, a resource recommendation processing device is also provided, which is described below with reference to the accompanying drawings.

[0057] Reference Figure 4 , which shows a schematic diagram of an embodiment of a resource recommendation processing device provided in this embodiment.

[0058] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the relevant parts can refer to the corresponding description of the method embodiment provided above. The device embodiment described below is only illustrative.

[0059] This embodiment provides a resource recommendation processing device, the device comprising: The data generation module 402 is configured to obtain access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data according to the access data and user data; The action decision module 404 is configured to input the recommendation decision data into the large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action; The recommendation processing module 406 is configured to call the agent corresponding to the recommendation action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0060] An embodiment of a resource recommendation processing device provided in this specification is as follows: Corresponding to the resource recommendation processing method described above, based on the same technical concept, one or more embodiments of this specification also provide a resource recommendation processing device, which is used to execute the resource recommendation processing method provided above. Figure 5 A schematic diagram of the structure of a resource recommendation processing device provided in one or more embodiments of this specification.

[0061] This embodiment provides a resource recommendation processing device, including: like Figure 5 As shown, the resource recommendation processing device may have relatively large differences due to different configurations or performances, and may include one or more processors 501 and memory 502, and the memory 502 may store one or more storage applications or data. Among them, the memory 502 may be a short-term storage or a persistent storage. The application stored in the memory 502 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the resource recommendation processing device. Furthermore, the processor 501 may be configured to communicate with the memory 502, and execute a series of computer executable instructions in the memory 502 on the resource recommendation processing device. The resource recommendation processing device may also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input / output interfaces 505, one or more keyboards 506, etc.

[0062] In a specific embodiment, the resource recommendation processing device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the resource recommendation processing device, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following: Acquire access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data; Inputting the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action; The agent corresponding to the recommendation action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0063] An embodiment of a computer-readable storage medium provided in this specification is as follows: Corresponding to the resource recommendation processing method described above, based on the same technical concept, one or more embodiments of this specification also provide a computer-readable storage medium.

[0064] The computer-readable storage medium provided in this embodiment is used to store computer-executable instructions, and the computer-executable instructions implement the following process when executed: Acquire access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data; Inputting the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action; The agent corresponding to the recommendation action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0065] It should be noted that an embodiment of a computer-readable storage medium in this specification and an embodiment of a resource recommendation processing method in this specification are based on the same inventive concept, so the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0066] An embodiment of a computer program product provided in this specification is as follows: Corresponding to the resource recommendation processing method described above, based on the same technical concept, one or more embodiments of this specification also provide a computer program product.

[0067] A computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the following steps: Acquire access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data; Inputting the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action; The agent corresponding to the recommendation action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

[0068] It should be noted that an embodiment of a computer program product in this specification and an embodiment of a resource recommendation processing method in this specification are based on the same inventive concept, so the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0069] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. For example, the device embodiment, equipment embodiment, computer-readable storage medium embodiment, and computer program product embodiment are similar to the method embodiment, so the description is relatively simple. To read the relevant contents of the device embodiment, equipment embodiment, computer-readable storage medium embodiment, and computer program product embodiment, please refer to the partial description of the method embodiment.

[0070] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0071] In the 1930s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to circuit structures such as diodes, transistors, switches, etc.) or software improvements (improvements to method flows). However, with the development of technology, many of today's method flow improvements can be regarded as direct improvements to hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using hardware entity modules. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask chip manufacturers to design and produce dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one kind of HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0072] The controller may be implemented in any suitable manner, for example, the controller may take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (e.g., software or firmware) executable by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, and the memory controller may also be implemented as part of the control logic of the memory. It is also known to those skilled in the art that, in addition to implementing the controller in a purely computer-readable program code manner, the controller may be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, such a controller may be considered as a hardware component, and the devices for implementing various functions included therein may also be considered as structures within the hardware component. Or even, the devices for implementing various functions may be considered as both software modules for implementing the method and structures within the hardware component.

[0073] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0074] For the convenience of description, the above devices are described in terms of functions and are divided into various units. Of course, when implementing the embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0075] It should be understood by those skilled in the art that one or more embodiments of the present specification may be provided as a method, system or computer program product. Therefore, one or more embodiments of the present specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present specification may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0076] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. 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 resource recommended processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable resource recommended 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 device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable resource processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0078] These computer program instructions may also be loaded onto a computer or other programmable resource 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 executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0079] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0080] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0081] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0082] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "includes at least one ..." do not exclude the presence of other identical elements in the process, method, commodity or device including the elements.

[0083] One or more embodiments of the present specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0084] The above description is only an embodiment of this document and is not intended to limit this document. For those skilled in the art, this document may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this document should be included in the scope of the claims of this document.

Claims

1. A resource recommendation processing method, comprising: Acquire access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data; Inputting the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action; The agent corresponding to the recommendation action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

2. The resource recommendation processing method according to claim 1, wherein the recommended action evaluation and recommended action decision-making include: Performing value evaluation on the recommended actions corresponding to each agent in the multi-agents to obtain a value evaluation result of each recommended action; The recommended action is selected from among the recommended actions according to the value evaluation result.

3. The resource recommendation processing method according to claim 1, wherein the access data comprises at least one of the following: The conversation data of the visiting user in the interactive area, the resource data of the currently visited resource of the resource service by the visiting user, the recommended action of the service node where the currently visited resource is located, and the conversation questions input by the visiting user in the interactive area.

4. The resource recommendation processing method according to claim 1, if the access data includes a dialogue question input by the accessing user in the interactive area; The calling of the agent corresponding to the recommended action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data includes: Invoking the intelligent agent to generate an answer to the dialogue question; The answer to the question is obtained by inputting the dialogue question and the access data into a large language model to generate the answer.

5. The resource recommendation processing method according to claim 4, further comprising: If the access control of the recommended resource carried in the answer to the question is triggered, the access page of the resource service is jumped to the resource access page corresponding to the access control of the recommended resource.

6. The resource recommendation processing method according to claim 1, wherein the step of calling the agent corresponding to the recommendation action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data comprises: Obtaining recommended resources by inputting the user data and the access data into a large language model to perform resource evaluation and resource decision making; The access data is updated based on the recommended resource to obtain updated access data, a dialogue is generated based on the user data and the updated access data, and the generated dialogue question is used as the resource recommendation data.

7. The resource recommendation processing method according to claim 1, wherein the agents in the multi-agents correspond to the resource states of the resources in the resource service; Any resource state of the resource corresponds to one or more agents.

8. The resource recommendation processing method according to claim 1, wherein the agent corresponding to the recommended action is deployed with at least one sub-agent, and each sub-agent corresponds to one or more sub-recommended actions; The calling of the agent corresponding to the recommended action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data includes: The sub-recommended action is obtained by inputting the node decision data into the large language model for sub-action evaluation and sub-action decision, and the resource recommendation processing is performed by calling the sub-agent corresponding to the sub-recommended action deployed by the agent.

9. The resource recommendation processing method according to claim 1, wherein the resources in the resource service are configured with respective multi-agents; The agents in the multi-agent of the resource are determined according to the configuration of the resource provider, and / or the sub-agents deployed by the agents in the multi-agent of the resource are determined according to the configuration of the resource provider.

10. The resource recommendation processing method according to claim 1, wherein the large language model is obtained by training a pre-trained model using a reinforcement learning method according to training samples; The training samples are constructed based on the state of an agent in the multi-agent before executing the recommended action, the recommended action, the reward after executing the recommended action, and the state after executing the recommended action.

11. A resource recommendation processing device, comprising: A data generation module is configured to obtain access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data; An action decision module is configured to input the recommended decision data into a large language model to perform recommended action evaluation and recommended action decision to obtain a recommended action; The recommendation processing module is configured to call the agent corresponding to the recommendation action in the multi-agents to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

12. A resource recommendation processing device, comprising: processor; and a memory configured to store computer executable instructions that, when executed, cause the processor to: Acquire access data obtained after the access user of the resource service triggers the service identifier, and generate recommendation decision data based on the access data and user data; Inputting the recommendation decision data into a large language model to perform recommendation action evaluation and recommendation action decision to obtain a recommended action; The agent corresponding to the recommendation action in the multi-agent is called to perform resource recommendation processing to obtain resource recommendation data, so as to display the resource recommendation data in the interactive area after the service identifier is triggered.

13. A computer-readable storage medium for storing computer-executable instructions, wherein the computer-executable instructions implement the steps of the method of claim 1 when executed.

Images