Interview dialogue type information pushing method based on multi-agent cooperation and related device

Abstract

The invention provides an interview dialogue type information pushing method based on multi-agent cooperation and a related device, and relates to the technical field of artificial intelligence such as generative large language models and agents. The method comprises the following steps: determining a target theme interested by a user; determining a plurality of target virtual human agents in the field to which the target theme belongs, issuing the target theme to each target virtual human agent, and constructing different virtual human agents for different fields in advance; controlling different target virtual human agents to publish different viewpoint information for the target theme, wherein each viewpoint information is acquired from an information acquisition channel associated with the target theme; and summarizing different viewpoint information published by each target virtual human agent to obtain an information push result around the target theme. According to the scheme, through cooperation between the main agent and each virtual human agent, an implementation mode of quickly and effectively transmitting related public viewpoints of a target subject interested by a user to the user in an interview and dialogue mode is realized, and the virtual human agents are human-like objects; the viewpoint information transmitted by the method can be quickly accepted and understood by the user in a more friendly manner, and an excellent effect can be achieved in an information pushing scene.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, specifically to the field of artificial intelligence technology such as generative large language models and intelligent agents, and more particularly to a method, device, electronic device, computer-readable storage medium and computer program product for pushing information in an interview dialogue-style manner based on multi-agent collaboration. Background Art

[0002] With the rapid development and iteration of generative large language models, they have a better understanding of user input requirements and the ability to give corresponding results.

[0003] In order to make the output of the generative large language model more in line with specific needs, an intelligent agent was constructed by using the generative large language model as the base model and combining it with pre-set role parameters.

[0004] How to use intelligent agents to solve complex problems raised by users is still a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0005] The embodiments of the present disclosure propose an interview dialogue information push method, device, electronic device, computer-readable storage medium and computer program product based on multi-agent collaboration.

[0006] In the first aspect, the disclosed embodiment proposes an interview dialogue-style information push method based on multi-agent collaboration, including: determining a target topic of interest to a user; determining multiple target virtual human agents under the field to which the target topic belongs, and sending the target topic to each target virtual human agent; wherein different virtual human agents are pre-constructed for different fields; controlling different target virtual human agents to publish different opinion information on the target topic; wherein each opinion information is obtained from an information acquisition channel associated with the target topic; and summarizing the different opinion information published by each target virtual human agent to obtain information push results around the target topic.

[0007] On the second aspect, the disclosed embodiment proposes an interview dialogue-style information push device based on multi-agent collaboration, comprising: a target topic determination unit, configured to determine a target topic of interest to a user; a target virtual human agent determination and topic distribution unit, configured to determine multiple target virtual human agents under the field to which the target topic belongs, and distribute the target topic to each target virtual human agent; wherein different virtual human agents are pre-constructed for different fields; an opinion information control and publication unit, configured to control different target virtual human agents to publish different opinion information for the target topic; wherein each opinion information is obtained from an information acquisition channel associated with the target topic; an opinion information aggregation unit, configured to aggregate different opinion information published by each target virtual human agent, and obtain information push results around the target topic.

[0008] In a third aspect, an embodiment of the present disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that when the at least one processor executes, it is able to implement the interview conversational information push method based on multi-agent collaboration as described in the first aspect.

[0009] In a fourth aspect, an embodiment of the present disclosure provides a non-transitory computer-readable storage medium storing computer instructions, which are used to enable a computer to implement the interview conversational information push method based on multi-agent collaboration as described in the first aspect when executed.

[0010] In a fifth aspect, an embodiment of the present disclosure provides a computer program product comprising a computer program, which, when executed by a processor, can implement the steps of the interview conversational information push method based on multi-agent collaboration as described in the first aspect.

[0011] The interview dialogue information push scheme based on multi-agent collaboration provided by the present disclosure is that the main agent first determines the target topic of interest to the user, then determines multiple target virtual human agents in the field to which the target topic belongs, and sends the target topic to each target virtual human agent. Then, different target virtual human agents will publish different opinion information on the target topic under the control of the main agent. These opinion information are obtained from information acquisition channels associated with the target topic. Finally, the main agent summarizes the different opinion information published by each target virtual human agent to obtain information push results around the target topic.

[0012] That is, the present disclosure uses an agent cluster formed by a pre-built main agent and multiple virtual human agents to handle the task requirements proposed by users, wherein the main agent is responsible for determining the target topic that the user is interested in and which virtual human agents in the corresponding field should be selected as target virtual human agents according to the target topic, and each selected target virtual human agent publishes opinion information around the target topic according to the mobilization of the main agent. That is, through the collaboration between the main agent and each virtual human agent, a method of quickly and effectively conveying relevant public opinions on the target topic of interest to users through interviews and dialogues is achieved, and the virtual human agent is a human-like object, so that the opinion information conveyed by it can be more friendly and quickly accepted and understood by users, and excellent results can be achieved in the information push scenario.

[0013] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Other features, objects and advantages of the present disclosure will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0015] Figure 1 is an exemplary system architecture in which the present disclosure may be applied;

[0016] Figure 2 A flowchart of an interview dialogue information push method based on multi-agent collaboration provided by an embodiment of the present disclosure;

[0017] Figure 3 A flowchart of a method for determining a target topic according to natural language input provided by an embodiment of the present disclosure;

[0018] Figure 4 A flowchart of a method for creating a virtual human agent provided by an embodiment of the present disclosure;

[0019] Figure 5 A flowchart of a method for determining a self-selected virtual human agent provided by an embodiment of the present disclosure;

[0020] Figure 6 A two-branch schematic diagram of a method of controlling a target virtual human agent to output corresponding viewpoint information provided by an embodiment of the present disclosure;

[0021] Figure 7 A flowchart of a method for initiating an inquiry of interest in a related topic provided by an embodiment of the present disclosure;

[0022] Figure 8-1 to Figure 8-10 All of them are example diagrams of an interview dialogue information push method based on multi-agent collaboration in an application scenario provided by an embodiment of the present disclosure;

[0023] Fig. 9 A structural block diagram of an interview dialogue information push device based on multi-agent collaboration provided by an embodiment of the present disclosure;

[0024] Fig.10 A structural schematic diagram of an electronic device suitable for executing an interview dialogue information push method based on multi-agent collaboration provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0025] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the embodiments in the present disclosure and the features in the embodiments may be combined with each other without conflict.

[0026] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0027] Figure 1 An exemplary system architecture 100 is shown to which an embodiment of the method, apparatus, electronic device and computer-readable storage medium for interview-based conversational information push based on multi-agent collaboration of the present disclosure can be applied.

[0028] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0029] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Terminal devices 101, 102, 103 and server 105 can be installed with various applications for realizing information communication between the two, such as complex task processing applications, browser applications, instant messaging applications, etc.

[0030] Terminal devices 101, 102, 103 and server 105 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices with display screens, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, etc.; when terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above, which can be implemented as multiple software or software modules, or as a single software or software module, and no specific limitation is made here. When server 105 is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server; when the server is software, it can be implemented as multiple software or software modules, or as a single software or software module, and no specific limitation is made here.

[0031] The server 105 can provide various services through various built-in applications. Taking the complex task processing application that can provide one-click processing service for complex tasks as an example, the server 105 can achieve the following effects when running the complex task processing application: first, it can receive the information of interest transmitted by the user using the terminal devices 101, 102, and 103 through the network 104, and then determine the target topic of interest to the user based on the information of interest; then, determine multiple target virtual human agents under the field to which the target topic belongs, and send the target topic to each target virtual human agent, and pre-build different virtual human agents for different fields; next, control different target virtual human agents to publish different opinion information on the target topic, and each opinion information is obtained from the information acquisition channel associated with the target topic; finally, summarize the different opinion information published by each target virtual human agent to obtain information push results around the target topic.

[0032] Furthermore, the server 105 may also transmit the target processing result back to the terminal devices 101 , 102 , 103 via the network 104 , so that the terminal devices 101 , 102 , 103 display the received target processing result to the user.

[0033] It should be noted that, in addition to being obtained from the terminal devices 101, 102, 103 through the network 104, the information of interest can also be pre-stored locally in the server 105 in various ways. Therefore, when the server 105 detects that the data has been stored locally (for example, before starting to process the tasks to be processed that were previously reserved), it can choose to directly obtain the data locally. In this case, the exemplary system architecture 100 may also not include the terminal devices 101, 102, 103 and the network 104.

[0034] Since processing complex tasks requires more computing resources and stronger computing power, the interview dialogue information push method based on multi-agent collaboration provided in the subsequent embodiments of the present disclosure is generally executed by the server 105 with stronger computing power and more computing resources. Correspondingly, the interview dialogue information push device based on multi-agent collaboration is generally also set in the server 105. However, it should also be pointed out that when the terminal devices 101, 102, and 103 also have computing power and computing resources that meet the requirements, the terminal devices 101, 102, and 103 can also complete the various operations assigned to the server 105 by the complex task processing applications installed thereon, and then output the same results as the server 105. Especially when there are multiple terminal devices with different computing capabilities, but when the terminal device where the complex task processing application is located has strong computing capabilities and more remaining computing resources, the terminal device can be allowed to perform the above-mentioned operations, thereby appropriately reducing the computing pressure of the server 105. Accordingly, the interview dialogue information push device based on multi-agent collaboration can also be set in the terminal devices 101, 102, and 103. In this case, the exemplary system architecture 100 may also not include the server 105 and the network 104.

[0035] It should be noted that the main agent and each sub-target agent can be installed on the server 105 at the same time, and each sub-target agent called and controlled by the main agent can also be installed on other servers or terminal devices different from the server 105, which is not specifically limited here.

[0036] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0037] Please refer to Figure 2 , Figure 2 A flowchart of a method for pushing information in an interview dialogue based on multi-agent collaboration provided by an embodiment of the present disclosure, wherein process 200 includes the following steps:

[0038] Step 201: Determine the target topic that the user is interested in;

[0039] This step is intended to be performed by the execution subject (e.g. Figure 1 The server 105 shown hosting the master agent determines target topics of interest to the user.

[0040] Specifically, one possible implementation method may be: determining a target topic of interest to the user based on the user's natural language input; another possible implementation method may be: determining messages of interest based on the user's message browsing history, and determining a target topic of interest based on the information of interest.

[0041] It can be seen that no matter which implementation method is used, the purpose is to obtain relevant information that can be extracted or determined about the target subject that the user is interested in, such as the natural language description of the information of interest input by the user himself, or the information of interest analyzed from the browsing history of the message. Of course, it is also possible to obtain the information of interest or other relevant information similar to the information of interest from other channels or methods.

[0042] Step 202: Determine multiple target virtual human agents in the field to which the target subject belongs, and send the target subject to each target virtual human agent;

[0043] On the basis of step 201, this step aims to determine multiple target virtual human agents in the field to which the target topic belongs by the above-mentioned execution subject, and send the target topic to each target virtual human agent. Among them, different virtual human agents are pre-built for different fields. The virtual human agent is an anthropomorphic agent, which is used in the present disclosure to provide an anthropomorphic dialogue method to have a friendly dialogue with the user, and it is expected that this dialogue method can be used to push various different viewpoints on the target topic that the user is interested in, thereby facilitating the user to accept and quickly obtain information.

[0044] Step 203: Control different target virtual human agents to publish different opinion information on the target topic;

[0045] Based on step 202, this step aims to control different target virtual human agents to publish different opinion information on the target topic respectively by the above-mentioned execution subject, so as to realize that it is as if multiple real characters are talking to the user, and push opinion information to the user through such a dialogue.

[0046] It should be understood that each target virtual human agent should actually output the opinion information under the call or arrangement of the main agent. Specifically, the main agent can simultaneously inform the execution order or execution trigger signal when sending the target topic to each target virtual human agent, so that each target virtual human agent can publish the corresponding opinion information in the informed publication order or publish the corresponding opinion information when judging that the current state meets the requirements of the execution trigger signal, that is, each target virtual human agent actively publishes the corresponding opinion information at the appropriate time; the main agent can also actively call each target virtual human agent to publish the corresponding opinion information as required when determining that different target virtual human agents meet the publication time of opinion information, that is, each target virtual human agent only passively publishes opinion information according to the call instruction.

[0047] Both the active execution mechanism and the passive execution mechanism mentioned above belong to different forms of collaboration between the main intelligent agent and multiple virtual human intelligent agents to jointly process tasks. The specific choice can be based on the actual needs of the actual application scenario, and no specific limitation is made here.

[0048] Step 204: Summarize the different viewpoints published by each target virtual human agent to obtain information push results around the target topic;

[0049] Based on step 203, this step aims to collect different viewpoints published by each target virtual human agent by the above-mentioned execution subject to obtain information push results around the target topic. The collection process may include various processing operations such as deduplication, arrangement, simplification, and adjustment of expressions, so that the final information push results are more in line with the user's expectations, preferences or needs.

[0050] The interview-conversation information push method based on multi-agent collaboration provided by the disclosed embodiment adopts an agent cluster formed by a pre-built main agent and multiple virtual human agents to process the task requirements proposed by the user, wherein the main agent is responsible for determining the target topic that the user is interested in and which virtual human agents in the corresponding field should be selected as target virtual human agents according to the target topic, and each selected target virtual human agent publishes opinion information around the target topic according to the mobilization of the main agent. That is, through the collaboration between the main agent and each virtual human agent, a method of quickly and effectively conveying the relevant public opinions of the target topic of interest to the user through an interview and conversation style is realized, and the virtual human agent is a human-like object, so that the opinion information conveyed by it can be more friendly and quickly accepted and understood by the user, and excellent results can be achieved in the information push scenario.

[0051] To deepen your understanding of how to identify target topics, please also refer to Figure 3 , Figure 3A flowchart of a method for determining a target topic according to a natural language input provided by an embodiment of the present disclosure, wherein process 300 includes the following steps:

[0052] Step 301: Convert the user's natural language input into natural language text;

[0053] This step aims to convert the natural language input in various forms into natural language text that is easy to recognize and output by the above-mentioned execution subject, that is, to uniformly convert various forms into text form. For example, speech-to-text technology can be used to convert natural speech signals in speech form into natural language text.

[0054] Step 302: Perform semantic understanding on the natural language text to obtain a semantic understanding result;

[0055] On the basis of step 301, this step aims to enable the above-mentioned execution subject to perform semantic understanding on the natural language text, and then obtain a semantic understanding result.

[0056] Step 303: Determine the suspected object of interest based on the semantic understanding result, and perform personalized modification on the suspected object of interest using the user's personalized preference to obtain the target topic of interest to the user.

[0057] On the basis of step 302, this step aims to first determine the suspected object of interest according to the semantic understanding result by the above-mentioned execution subject, and then further use the personalized preference of the user to make personalized corrections to the preliminary intention in order to improve the accuracy, and make personalized corrections to obtain the target topic that the user is interested in. The personalized preference may include at least one of age, education level, and historical object of interest.

[0058] That is, this implementation method first determines the suspected object of interest that directly corresponds to the natural language text by using semantic understanding, and then uses the user's personalized preferences to personalize the suspected object of interest, such as screening, excluding, and supplementing implicit or missing information, so that the final determined target topic is more in line with the user's actual situation.

[0059] This embodiment provides a more specific implementation method for determining the target topic based on natural language input through steps 301 to 303, which includes the key processing steps of conversion of expression form, semantic understanding, and personalized correction based on personalized preferences to finally determine the target topic, so as to provide an implementation scheme with high feasibility and more accurate determination of the target topic as much as possible.

[0060] For a deeper understanding of how to pre-create different virtual human agents, see Figure 4 , Figure 4A flowchart of a method for creating a sub-agent provided by an embodiment of the present disclosure, wherein the process 400 includes the following steps:

[0061] Step 401: Determine a domain set for which a virtual human agent needs to be created;

[0062] Step 402: For each domain in the domain set, construct different virtual human agents representing different viewpoint information that can be obtained from different information acquisition channels.

[0063] Wherein, different information acquisition channels may include: different information disclosure websites or different opinion publishing objects. When the different information acquisition channels are specifically different information disclosure websites, step 203 will be adaptively changed to: control different target virtual human agents to publish opinion information on the target topic that is disclosed on the corresponding information disclosure websites; when the different information acquisition channels are specifically different opinion publishing objects, step 203 will be adaptively changed to: control different target virtual human agents to publish opinion information on the target topic that is published by the corresponding opinion publishing objects.

[0064] Furthermore, when the object of the opinion is an expert, scholar or academic representative, the target virtual human agent may be a virtual human agent having the same object characteristics as the expert, scholar or academic representative. The same object characteristics may include at least one of expression habits, dialogue and communication habits, and name codes. Taking the name code as an example, the target virtual human agent may be a virtual human agent having the same name as the real expert or scholar.

[0065] This embodiment provides a method of constructing different virtual human agents representing different viewpoints that can be obtained from different information acquisition channels for each of the determined multiple fields through steps 401 and 402.

[0066] Based on the above embodiment showing the solution of determining the target virtual human agent according to the target subject, in order to meet the special needs of users as much as possible, please refer to Figure 5 , Figure 5 A flowchart of a method for determining a self-selected virtual human agent provided by an embodiment of the present disclosure, wherein the process 500 includes the following steps:

[0067] Step 501: In response to receiving the user's selection information on the target preferred object, determining a virtual human agent simulating the target preferred character as a self-selected virtual human agent;

[0068] Step 502: Use the self-selected virtual human agent as a special target virtual human agent.

[0069] That is, when being able to receive the user's selection information on the target preference object, the above-mentioned execution subject can also determine the virtual human agent simulated by the target preference character as the self-selected virtual human agent, and use the self-selected virtual human agent as a special target virtual human agent. At this time, the self-selected virtual human agent as a special target virtual human agent has a higher priority for publishing opinion information than the target virtual human agent determined based on the topic. Therefore, the special target virtual human agent will publish the opinion information first, and can also interpret certain opinions at the request of the user.

[0070] Based on any of the above embodiments, see also Figure 6 , Figure 6 A two-branch schematic diagram of controlling a target virtual human agent to output corresponding viewpoint information provided by an embodiment of the present disclosure, wherein a process 600 includes the following steps:

[0071] Step 601: In response to the instruction input box being in a selected state, controlling the currently executed target virtual human agent to pause outputting corresponding viewpoint information;

[0072] The command input box is used for the user to input commands, and the command input box is in an unselected state during the process of the target virtual human agent outputting the corresponding viewpoint information. That is, in the schemes of the embodiments provided in the disclosure, once the command input box is in a selected state, it may represent that the user needs to input some new commands and needs to interrupt the original viewpoint information output process.

[0073] Step 602: In response to no new instruction being generated during the process of the instruction input box being restored from the selected state to the unselected state, controlling the currently executed target virtual human agent to continue to output corresponding viewpoint information;

[0074] This step corresponds to a branch situation in which no new instructions are generated during the process of the instruction input box being restored from a selected state to an unselected state, that is, the user does not enter a new instruction, so the opinion information publishing behavior that was originally interrupted by the selection of the instruction input box will continue.

[0075] Step 603: in response to a new instruction being generated in the process of the instruction input box being restored from the selected state to the unselected state, extracting a further inquiry initiated for the target viewpoint from the new instruction;

[0076] This step corresponds to another branch situation in which a new instruction is generated during the process of the instruction input box being restored from a selected state to an unselected state. In this case, the above-mentioned execution subject is required to extract further inquiries initiated for the target viewpoint from the new instruction.

[0077] Step 604: Control the target virtual human agent that outputs the target viewpoint to output a matching query explanation for further queries.

[0078] Step 604 is that the target virtual human agent that outputs the target viewpoint is controlled by the above-mentioned execution subject to output a matching query explanation for further inquiries, so as to eliminate the user's doubts as soon as possible in this way.

[0079] See also Figure 7 , Figure 7 A flowchart of a method for initiating an inquiry of interest in a related topic provided by an embodiment of the present disclosure, wherein the process 700 includes the following steps:

[0080] Step 701: Control each target virtual human agent to initiate an interest inquiry corresponding to a topic related to the target topic to the user;

[0081] Step 702: determining the actual interest level according to the received inquiry feedback;

[0082] Step 703: In response to the actual interest level exceeding the preset interest level, the target virtual human agent corresponding to the associated topic is arranged as the virtual human agent that subsequently publishes the corresponding opinion information.

[0083] That is, this embodiment provides a method of controlling each target virtual human agent to initiate an interest inquiry corresponding to the related topics of the target topic to the user through steps 701 to 703, so as to judge whether the user also has a high interest in the related topics based on the inquiry feedback, and then when a high interest is found, the target virtual human agent corresponding to the related topic is arranged as the virtual human agent that subsequently publishes the corresponding opinion information, thereby extending the publication of the opinion information to the related topics.

[0084] To deepen understanding, the present disclosure also provides a specific implementation scheme that attempts to eliminate the defects of the prior art and overcome the problems of the prior art in combination with the actual defects of the prior art in specific application scenarios:

[0085] The current content consumption satisfaction is to recommend and distribute the content that the creator has already produced, match the user's interests and content through personalized recommendation technology, and screen out the final distribution content through a multi-layer sorting funnel. The advantage of this static content is that it can use ecological creators to create content in multiple fields, collect content that meets the quality through content evaluation, separate production and consumption, and can be optimized separately. However, the disadvantages are also obvious. The content cannot be flexibly changed after production, the content cannot be interactive, and the matching of content and user interests is distributed in series through the recommendation system modules. The error accumulation is large, the accuracy requirements for each module are high, and the ceiling is limited.

[0086] That is, there are several difficulties in the relevant technology for content consumption scenarios:

[0087] 1) Recommendation cannot integrate content generation and content creation;

[0088] 2) Traditional content is solidified after creation and cannot be flexibly adjusted according to consumption scenarios;

[0089] 3) Users cannot participate in the content and it is not interesting.

[0090] To solve the above difficulties, this embodiment proposes an immersive consumption satisfaction technology based on multi-agent interviews. The characteristics of this solution are as follows:

[0091] 1) It is consumable content that is built purely automatically and does not need to rely on the content ecosystem and creator operations. It has the ability to quickly generate content and iterate on effects;

[0092] 2) It is based on a thorough multi-agent solution. It builds a main agent and then schedules multiple sub-task agents to conduct interviews. The generation of consumer interview content is completed through collaboration between the main agent and multiple sub-task agents, and between multiple sub-task agents. It is a new paradigm for generating consumer content.

[0093] 3) Production based on the user's real-time status can greatly capture the dynamic changes in the user's status, including changes in the user's short-term consumption interests and consumption emotions, and can better meet the user's current consumption needs.

[0094] This embodiment proposes a recommendation technology based on multi-agent collaboration, which can be applied to information flow consumer products. By using multiple agents to collaboratively generate real-time interview content on topics of interest to users, it is equivalent to completing the following multiple functions in one task:

[0095] 1) Generate recommended topics based on user status. Combine the user's interest profile, real-time consumption dynamics, and real-time consumption emotions to generate real-time recommended topics that users can consume.

[0096] 2) Subtask agent generation: Based on the understanding of key steps and agent capabilities, the main agent is responsible for generating multiple subtask agent candidates required to complete the current interview task from the candidate agent distribution library.

[0097] 3) Interview dialogue generation for real-time consumption and interaction. Based on the topic and subtask agent candidates, multi-agent collaborative dialogue group chats of subtasks generate content that is in line with the user's status, highly consumable, and user-participatory.

[0098] This is a great innovation in the information flow recommendation method, which upgrades the original keyword interest content matching recommendation to a new recommendation paradigm that generates content based on user status dialogue. In terms of content satisfaction, through anthropomorphic dialogue, user needs are interactively met, realizing the organic combination of content and interaction.

[0099] The implementation block diagram of this solution is as follows Figure 8-1 As shown in the figure, there are two types of inputs to the system: 1) user portraits and short-term user status, as well as short-term user consumption interests, consumption satisfaction moods and other emotional states; 2) consumption content (pictures, texts, videos, etc.) centered around a certain topic. The output is the generated interview topic, as well as multiple interview agents under the corresponding topic and the interview content generated between them. The entire system is built based on a multi-agent collaborative approach and mainly includes two types of agents: 1) Main Agent: performs overall understanding, decomposition and concatenation of tasks, and gives the final result. 2) Interview subtask agent (TaskAgent): The main agent summons the interview subtask agent to collaborate to complete the task issued by the main agent.

[0100] The solution provided in this embodiment can be widely applied to various information-satisfying applications or independent products. The following takes the scenario where the BaiX application is used as an example to illustrate the specific application form of this embodiment:

[0101] In the user's information flow recommendation list, add the real-time generated immersive content flow of the multi-agent interview in the invention, as follows Figure 8-2 The generated topic shown is: "Are large models suitable for making games?" and "Sam Altman and other intelligent agents in the field of science and technology are hotly discussed...". Add the generated results in the invention to the landing page scene of the content, as follows Figure 8-3 shown.

[0102] After the user clicks, he enters the consumption page of the multi-agent immersive interview, as shown in 8-4 below. The page presents multiple agent dialogues, each agent expresses his or her own views on the topic, and conducts a multi-person discussion guided and commented by the host. The host summarizes the dialogue process in series, such as Figure 8-5 and Figure 8-6 Finally, the host summarized the interview as follows Figure 8-7 shown.

[0103] Users can directly discuss a certain point in the interview content through a natural conversation method similar to other social applications, such as Figure 8-8 to Figure 8-10 .

[0104] It can be more specifically seen from the above examples that this embodiment has the following improvements and technical effects compared to the prior art:

[0105] 1) The user's consumption content is matched and generated in real time with the user's interests and status, and the consumption experience can be greatly improved.

[0106] Traditional content recommendation algorithms are often based on users' historical interests and feedback, but they are unable to capture users' real-time consumption status, and are unable to make good use of users' real-time status. It is impossible to establish an accurate connection between users' consumption status and content production. The present invention can take the user's current real-time status, including interests, consumption content, consumption mood and other highly real-time signals as input, and generate content that matches the immediate status in real time, which can greatly improve users' acceptance of content and truly understand users and satisfy them with consumption content.

[0107] 2) Based on the big model technology, the generative method of recommendation and the generative method of content are organically combined. Since the content of the current mainstream content flow consumption products is created by the ecosystem, the content consumption technology can only achieve generative recommendation in the recommendation algorithm, or real-time generation of content, but not immersive consumption. The present invention models the two as a whole, and the integrated generation system can greatly improve the ceiling of immersive consumption experience optimization.

[0108] 3) Use multi-agent interview-style content production and recommendation to enhance user participation in content and improve the organic combination of conversation and content consumption.

[0109] The interview-style content produced allows users to participate in the production and creation of content. Production and consumption are carried out simultaneously, and users can lead the direction of content production. User needs can be expressed in the production process, which enhances the user's sense of participation. At the same time, the experience of talking with people is more in line with the way humans obtain information through communication with others.

[0110] Further references Fig. 9 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of an interview dialogue information push device based on multi-agent collaboration. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0111] like Fig. 9As shown, the interview dialogue information push device 900 based on multi-agent collaboration of this embodiment may include: a target topic determination unit 901, a target virtual human agent determination and topic delivery unit 902, a viewpoint information control and delivery unit 903, and a viewpoint information aggregation unit 904. The target topic determination unit 901 is configured to determine the target topic that the user is interested in; the target virtual human agent determination and topic delivery unit 902 is configured to determine multiple target virtual human agents under the field to which the target topic belongs, and deliver the target topic to each target virtual human agent; wherein different virtual human agents are pre-constructed for different fields; the viewpoint information control and delivery unit 903 is configured to control different target virtual human agents to publish different viewpoint information for the target topic; wherein each viewpoint information is obtained from an information acquisition channel associated with the target topic; and the viewpoint information aggregation unit 904 is configured to aggregate the different viewpoint information published by each target virtual human agent to obtain the information push result around the target topic.

[0112] In this embodiment, in the interview dialogue information push device 900 based on multi-agent collaboration: the specific processing of the target topic determination unit 901, the target virtual human agent determination and topic delivery unit 902, the opinion information control publishing unit 903, and the opinion information aggregation unit 904 and the technical effects thereof can be referred to respectively. Figure 2 The relevant descriptions of steps 201 - 204 in the corresponding embodiment are not repeated here.

[0113] In some optional implementations of this embodiment, the target theme determination unit 901 may include:

[0114] The target subject determination subunit is configured to determine the target subject that the user is interested in based on the user's natural language input.

[0115] In some optional implementations of this embodiment, the target subject determination subunit is further configured to:

[0116] Convert user's natural language input into natural language text;

[0117] Perform semantic understanding on natural language text to obtain semantic understanding results;

[0118] The suspected object of interest is determined based on the semantic understanding results, and the suspected object of interest is personalized modified using the user's personalized preferences to obtain the target topic of interest to the user.

[0119] In some optional implementations of this embodiment, the personalized preference includes at least one of the following:

[0120] Age, education level, history of interest.

[0121] In some optional implementations of this embodiment, the target theme determination unit 901 may be further configured to:

[0122] Determine messages of interest based on the user's message browsing history;

[0123] Determine target topics of interest based on the information of interest.

[0124] In some optional implementations of this embodiment, the multi-agent collaborative interview dialog information push device 900 may further include: a virtual human agent construction unit configured to construct different virtual human agents, and the virtual human agent construction unit may be further configured to:

[0125] Determine the set of domains in which a virtual human agent needs to be created;

[0126] For each domain in the domain set, different virtual human agents are constructed to represent different viewpoints of information that can be obtained from different information acquisition channels.

[0127] In some optional implementations of this embodiment, different information acquisition channels include: different information disclosure websites or different opinion expression objects.

[0128] In some optional implementations of this embodiment, the opinion information control publishing unit 903 may be further configured to:

[0129] In response to different information acquisition channels including different information disclosure websites, different target virtual human agents are controlled to publish opinion information on target topics disclosed on corresponding information disclosure websites.

[0130] In some optional implementations of this embodiment, the opinion information control publishing unit 903 may be further configured to:

[0131] In response to different information acquisition channels including different opinion-posting objects, different target virtual human agents are controlled to respectively post opinion information on a target topic posted by corresponding opinion-posting objects.

[0132] In some optional implementations of this embodiment, in response to the subject of the opinion being an expert scholar or academic representative, the target virtual human agent is a virtual human agent having the same object characteristics as the expert scholar or academic representative.

[0133] In some optional implementations of this embodiment, the same object feature includes at least one of the following:

[0134] Expression habits, dialogue and communication habits, and name codes.

[0135] In some optional implementations of this embodiment, the multi-agent collaborative interview dialogue information push device 900 may further include:

[0136] A self-selected virtual human agent determining unit is configured to determine a virtual human agent simulating a target preferred character as a self-selected virtual human agent in response to receiving selection information of a target preferred object by a user;

[0137] The special target virtual human agent determination unit is configured to use the self-selected virtual human agent as a special target virtual human agent; wherein the self-selected virtual human agent has a higher priority for expressing opinion information than the target virtual human agent.

[0138] In some optional implementations of this embodiment, the multi-agent collaborative interview dialogue information push device 900 may further include:

[0139] The previous and next opinion information association control unit is configured to control the opinion information published by the target virtual human agent that publishes the opinion information later to be associated with the opinion information published by the target virtual human agent that published the opinion information earlier in response to the number of target virtual human agents being greater than 1.

[0140] In some optional implementations of this embodiment, the opinion information control publishing unit 903 may be further configured to:

[0141] Control different target virtual human agents to publish different opinion information on the target topic in turn.

[0142] In some optional implementations of this embodiment, the multi-agent collaborative interview dialogue information push device 900 may further include:

[0143] The switching transition prompt output unit is configured to output a switching transition prompt for switching to a different target virtual human agent between different target virtual human agents successively publishing different opinion information on the target topic.

[0144] In some optional implementations of this embodiment, the multi-agent collaborative interview dialogue information push device 900 may further include:

[0145] The pause output control unit is configured to control the currently executed sub-agent to pause outputting the corresponding sub-processing result in response to the instruction input box being in a selected state; wherein the instruction input box is used for the user to input instructions, and the instruction input box is in an unselected state during the process of the sub-agent outputting the corresponding sub-processing result;

[0146] a further inquiry extraction unit, configured to generate a new instruction in response to the instruction input box being restored from a selected state to an unselected state, and extract a further inquiry initiated for the target viewpoint from the new instruction;

[0147] The query interpretation control output unit is configured to control the target virtual human agent that outputs the target viewpoint to output a matching query interpretation for further queries.

[0148] In some optional implementations of this embodiment, the multi-agent collaborative interview dialogue information push device 900 may further include:

[0149] A related subject inquiry control unit is configured to control each target virtual human agent to initiate an interest level inquiry corresponding to a related subject of the target subject to the user;

[0150] an interest level determination unit configured to determine an actual interest level based on the received query feedback;

[0151] The other virtual human agent speech arrangement unit is configured to arrange the target virtual human agent corresponding to the associated topic as the virtual human agent that subsequently expresses the corresponding opinion information in response to the actual interest level exceeding the preset interest level.

[0152] This embodiment exists as a device embodiment corresponding to the above method embodiment. The interview-conversation information push device based on multi-agent collaboration provided by this embodiment adopts an agent cluster formed by a pre-built main agent and multiple virtual human agents to process the task requirements proposed by the user, wherein the main agent is responsible for determining the target topic of interest to the user and which virtual human agents in the field to which the target topic belongs should be selected as the target virtual human agents, and each selected target virtual human agent publishes opinion information around the target topic according to the mobilization of the main agent. That is, through the collaboration between the main agent and each virtual human agent, a method of quickly and effectively conveying the relevant public opinions of the target topic of interest to the user through an interview and conversation style is realized, and the virtual human agent is a human-like object, so that the opinion information conveyed by it can be more friendly and quickly accepted and understood by the user, and excellent results can be achieved in the information push scenario.

[0153] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can implement the interview conversational information push method based on multi-agent collaboration described in any of the above embodiments when executing.

[0154] According to an embodiment of the present disclosure, the present disclosure also provides a readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to implement the interview conversational information push method based on multi-agent collaboration described in any of the above embodiments when executed.

[0155] According to an embodiment of the present disclosure, the present disclosure also provides a computer program product, which, when executed by a processor, can implement the interview conversational information push method based on multi-agent collaboration described in any of the above embodiments.

[0156] Fig.10 A schematic block diagram of an example electronic device 1000 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0157] like Fig.10 As shown, the device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0158] A number of components in the device 1000 are connected to the I / O interface 1005, including: an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a disk, an optical disk, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the device 1000 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0159] The computing unit 1001 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 1001 performs the various methods and processes described above, such as an interview conversational information push method based on multi-agent collaboration. For example, in some embodiments, the interview conversational information push method based on multi-agent collaboration may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 1008. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 1000 via the ROM 1002 and / or the communication unit 1009. When the computer program is loaded into the RAM 1003 and executed by the computing unit 1001, one or more steps of the interview conversational information push method based on multi-agent collaboration described above may be executed. Alternatively, in other embodiments, the computing unit 1001 may be configured in any other appropriate manner (e.g., by means of firmware) to execute an interview conversational information push method based on multi-agent collaboration.

[0160] Various implementations of the systems and techniques described above herein 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 chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0161] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0162] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0163] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0164] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0165] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and virtual private servers (VPS) services.

[0166] The technical solution of the disclosed embodiment adopts an agent cluster formed by a pre-built main agent and multiple virtual human agents to process the task requirements proposed by the user, wherein the main agent is responsible for determining the target topic that the user is interested in and which virtual human agents in the corresponding field should be selected as the target virtual human agents according to the target topic, and each selected target virtual human agent publishes opinion information around the target topic according to the mobilization of the main agent. That is, through the collaboration between the main agent and each virtual human agent, a method of quickly and effectively conveying the relevant public opinions of the target topic that the user is interested in to the user through interviews and dialogues is realized, and the virtual human agent is a human-like object, so that the opinion information conveyed by it can be more friendly and quickly accepted and understood by the user, and excellent results can be achieved in the information push scenario.

[0167] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0168] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. An interview dialogue information push method based on multi-agent collaboration, applied to a main agent, comprising: Identify target topics of interest to users; Determine a plurality of target virtual human agents in the field to which the target subject belongs, and send the target subject to each of the target virtual human agents; wherein different virtual human agents are pre-constructed for different fields; Controlling different target virtual human agents to publish different opinion information on the target topic; wherein each opinion information is obtained from an information acquisition channel associated with the target topic; The different viewpoints expressed by the target virtual human agents are summarized to obtain information push results around the target topic.

2. The method according to claim 1, wherein: Determining the target topic that the user is interested in includes: A target topic of interest to the user is determined based on the user's natural language input.

3. The method according to claim 2, wherein: Determining the target topic that the user is interested in based on the natural language input of the user includes: Converting the user's natural language input into natural language text; Performing semantic understanding on the natural language text to obtain a semantic understanding result; A suspected object of interest is determined according to the semantic understanding result, and the suspected object of interest is personalized modified using the personalized preference of the user to obtain a target topic of interest to the user.

4. The method according to claim 3, wherein: The personalized preference includes at least one of the following: Age, education level, history of interest.

5. The method according to claim 1, wherein: Determining the target topic that the user is interested in includes: Determining messages of interest based on the message browsing history of the user; A target topic of interest is determined according to the information of interest.

6. The method according to claim 1, wherein: The process of building different virtual human agents includes: Determine the set of domains in which a virtual human agent needs to be created; For each domain in the domain set, different virtual human agents are constructed to represent different viewpoint information that can be obtained from different information acquisition channels.

7. The method according to claim 6, wherein: Different channels for obtaining information include: different information disclosure websites or different objects for expressing opinions.

8. The method according to claim 7, wherein: The controlling different target virtual human agents to publish different opinion information on the target topic includes: In response to different information acquisition channels including different information disclosure websites, different target virtual human agents are controlled to publish opinion information on the target topic disclosed on the corresponding information disclosure websites.

9. The method according to claim 7, wherein: The controlling different target virtual human agents to publish different opinion information on the target topic includes: In response to different information acquisition channels including different opinion-posting objects, different target virtual human agents are controlled to respectively post opinion information on the target topic posted by the corresponding opinion-posting objects.

10. The method according to claim 9, wherein: In response to the object of the opinion being an expert scholar or an academic representative, the target virtual human agent is a virtual human agent having the same object characteristics as the expert scholar or the academic representative.

11. The method according to claim 10, wherein: The same object features include at least one of the following: Expression habits, dialogue and communication habits, and name codes.

12. The method according to claim 1, further comprising: In response to receiving the user's selection information of the target preferred object, determining a virtual human agent simulating the target preferred character as a self-selected virtual human agent; The self-selected virtual human agent is used as a special target virtual human agent; wherein the self-selected virtual human agent has a higher priority for expressing opinion information than the target virtual human agent.

13. The method according to claim 1, further comprising: In response to the number of the target virtual human agents being greater than 1, the opinion information published by the target virtual human agents that publish opinion information later is controlled to be associated with the opinion information published by the target virtual human agents that publish opinion information earlier.

14. The method according to claim 1, wherein: The controlling different target virtual human agents to publish different opinion information on the target topic includes: Control different target virtual human agents to publish different opinion information on the target topic in turn.

15. The method according to claim 14, further comprising: When different target virtual human agents successively publish different opinion information on the target topic, a switching transition prompt for switching to a different target virtual human agent is output.

16. The method according to any one of claims 1 to 15, further comprising: In response to the instruction input box being in a selected state, controlling the currently executed sub-agent to pause outputting the corresponding sub-processing result; wherein the instruction input box is used for the user to input instructions, and the instruction input box is in an unselected state during the process of the sub-agent outputting the corresponding sub-processing result; In response to a new instruction being generated in the process of the instruction input box being restored from the selected state to the unselected state, extracting a further inquiry initiated for the target viewpoint from the new instruction; The target virtual human agent that outputs the target viewpoint is controlled to output a matching query interpretation for the further query.

17. The method according to claim 16, further comprising: Controlling each of the target virtual human agents to initiate an interest inquiry corresponding to a topic associated with the target topic to the user; Determine the actual level of interest based on the inquiry feedback received; In response to the actual interest level exceeding the preset interest level, the target virtual human agent corresponding to the associated topic is arranged as the virtual human agent that subsequently publishes corresponding opinion information.

18. An interview dialogue information push device based on multi-agent collaboration, applied to a main agent, comprising: a target topic determination unit, configured to determine a target topic of interest to the user; The target virtual human agent determination and topic distribution unit is configured to determine multiple target virtual human agents in the field to which the target topic belongs, and distribute the target topic to each of the target virtual human agents; wherein different virtual human agents are pre-constructed for different fields; The opinion information control and publication unit is configured to control different target virtual human agents to publish different opinion information on the target topic; wherein each of the opinion information is obtained from an information acquisition channel associated with the target topic; The opinion information aggregation unit is configured to aggregate the different opinion information published by each of the target virtual human agents to obtain information push results around the target topic.

19. The device according to claim 18, wherein: The target theme determination unit comprises: The target subject determination subunit is configured to determine the target subject that the user is interested in based on the natural language input of the user.

20. The device according to claim 19, wherein The target theme determination unit is further configured to: Converting the user's natural language input into natural language text; Performing semantic understanding on the natural language text to obtain a semantic understanding result; A suspected object of interest is determined according to the semantic understanding result, and the suspected object of interest is personalized modified using the personalized preference of the user to obtain a target topic of interest to the user.

21. The device according to claim 20, wherein: The personalized preference includes at least one of the following: Age, education level, history of interest.

22. The device according to claim 18, wherein The target theme determination unit is further configured to: Determining messages of interest based on the message browsing history of the user; A target topic of interest is determined according to the information of interest.

23. The apparatus of claim 18, further comprising: A virtual human agent construction unit configured to construct different virtual human agents, wherein the virtual human agent construction unit is further configured to: Determine the set of domains in which a virtual human agent needs to be created; For each domain in the domain set, different virtual human agents are constructed to represent different viewpoint information that can be obtained from different information acquisition channels.

24. The device according to claim 23, wherein: Different channels for obtaining information include: different information disclosure websites or different objects for expressing opinions.

25. The device according to claim 24, wherein: The opinion information control publishing unit is further configured to: In response to different information acquisition channels including different information disclosure websites, different target virtual human agents are controlled to publish opinion information on the target topic disclosed on the corresponding information disclosure websites.

26. The device according to claim 24, wherein: The opinion information control publishing unit is further configured to: In response to different information acquisition channels including different opinion-posting objects, different target virtual human agents are controlled to respectively post opinion information on the target topic posted by the corresponding opinion-posting objects.

27. The device according to claim 26, wherein: In response to the object of the opinion being an expert scholar or an academic representative, the target virtual human agent is a virtual human agent having the same object characteristics as the expert scholar or the academic representative.

28. The device according to claim 27, wherein The same object features include at least one of the following: Expression habits, dialogue and communication habits, and name codes.

29. The apparatus of claim 18, further comprising: A self-selected virtual human agent determining unit is configured to determine a virtual human agent simulating the target preferred character as a self-selected virtual human agent in response to receiving the user's selection information on the target preferred object; The special target virtual human agent determination unit is configured to take the self-selected virtual human agent as a special target virtual human agent; wherein the self-selected virtual human agent has a higher priority for expressing opinion information than the target virtual human agent.

30. The apparatus of claim 18, further comprising: The previous and next opinion information association control unit is configured to control the opinion information published by the target virtual human agent that publishes the opinion information later to be associated with the opinion information published by the target virtual human agent that publishes the opinion information earlier in response to the number of the target virtual human agents being greater than 1.

31. The apparatus of claim 18, wherein: The opinion information control publishing unit is further configured to: Control different target virtual human agents to publish different opinion information on the target topic in turn.

32. The apparatus of claim 31 , further comprising: The switching transition prompt output unit is configured to output a switching transition prompt for switching to a different target virtual human agent between different target virtual human agents successively publishing different opinion information on the target topic.

33. The apparatus according to any one of claims 18 to 32, further comprising: a pause output control unit, in response to the instruction input box being in a selected state, controlling the currently executed sub-agent to pause outputting the corresponding sub-processing result; wherein the instruction input box is used by the user to input instructions, and the instruction input box is in an unselected state during the process of the sub-agent outputting the corresponding sub-processing result; a further inquiry extraction unit, configured to generate a new instruction in response to the instruction input box being restored from the selected state to the unselected state, and extract a further inquiry initiated for the target viewpoint from the new instruction; The query interpretation control output unit is configured to control the target virtual human agent that outputs the target viewpoint to output a matching query interpretation for the further query.

34. The apparatus of claim 33, further comprising: A related subject inquiry control unit configured to control each of the target virtual human agents to initiate an interest level inquiry corresponding to a related subject of the target subject to the user; an interest level determination unit configured to determine an actual interest level based on the received query feedback; The other virtual human agent speech arrangement unit is configured to arrange the target virtual human agent corresponding to the related topic as the virtual human agent that subsequently expresses the corresponding opinion information in response to the actual interest level exceeding the preset interest level.

35. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the interview conversational information push method based on multi-agent collaboration as described in any one of claims 1-17.

36. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the interview conversational information push method based on multi-agent collaboration as described in any one of claims 1-17.

37. A computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the interview conversational information push method based on multi-agent collaboration according to any one of claims 1-17.

Images