Information recommendation method and device, intelligent agent and electronic device

Abstract

The present disclosure provides an information recommendation method and device, an intelligent agent and an electronic device, and relates to the technical field of artificial intelligence such as intelligent agents. The method comprises: analyzing consumption data of a target user according to a preset role library to determine role information of the target user; controlling a thinking exploration intelligent agent to analyze the consumption data of the target user according to the role information of the target user to obtain an exploration thinking tree comprising at least one causal relationship link; and exploring the interest of the target user according to the role information of the target user, the exploration thinking tree and environment information of the target user to obtain target recommendation information for the target user. The application of the method can reduce recommendation bias and improve the user experience.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, specifically to the field of artificial intelligence technology such as intelligent agents, and particularly to an information recommendation method, as well as an information recommendation device, intelligent agent, and electronic device. Background Technology

[0002] In consumer behavior analysis, traditional methods (such as questionnaires and statistical analysis) often fail to reveal the true motivations behind users' consumption behavior. Recommendations based on these traditional methods cannot accurately meet user needs, causing users' actual needs to be ignored and reducing the user experience. Summary of the Invention

[0003] This disclosure provides an information recommendation method, apparatus, intelligent agent, and electronic device.

[0004] In a first aspect, embodiments of this disclosure propose an information recommendation method, comprising: analyzing the consumption data of a target user based on a preset role library to determine the target user's role information; controlling a thinking exploration agent to perform causal relationship analysis on the target user's consumption data based on the target user's role information to obtain an exploration thinking tree including at least one causal relationship link; and exploring the target user's interests based on the target user's role information, the exploration thinking tree, and the target user's environment information to obtain target recommendation information for the target user to use.

[0005] Secondly, embodiments of this disclosure propose an information recommendation device, comprising: a role determination unit configured to analyze the consumption data of a target user based on a preset role library to determine the target user's role information; a causal relationship exploration unit configured to control a thinking exploration agent to perform causal relationship analysis on the target user's consumption data based on the target user's role information to obtain an exploration thinking tree including at least one causal relationship link; and an information recommendation unit configured to explore the target user's interests based on the target user's role information, the exploration thinking tree, and the target user's environment information to obtain target recommendation information for the target user.

[0006] Thirdly, embodiments of this disclosure propose an intelligent agent, comprising: an input module for receiving input information; a processing module for determining a target task based on the input information received by the input module, determining an information recommendation model based on the target task, and executing an information recommendation method as described in any implementation of the first aspect by calling the information recommendation model to obtain output information; and an output module for outputting the output information obtained by the processing module.

[0007] Fourthly, embodiments of this disclosure provide 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, the instructions being executed by the at least one processor to enable the at least one processor to implement the information recommendation method as described in any implementation of the first aspect.

[0008] Fifthly, embodiments of this disclosure provide a non-transitory computer-readable storage medium storing computer instructions that enable a computer to implement the information recommendation method as described in any implementation of the first aspect.

[0009] In a sixth aspect, embodiments of this disclosure provide a computer program product including a computer program that, when executed by a processor, can implement the information recommendation method as described in any implementation of the first aspect.

[0010] The information recommendation method provided in this disclosure analyzes the consumption data of target users through a preset role database to determine the role information of the target users. It then employs a thinking exploration agent to perform causal relationship analysis on the target user's role information and consumption data to obtain an exploration mind tree. This method clearly defines the interest range of the target user's corresponding role. Compared to conventional methods that determine user interests solely based on the correlation between users and their needs, the interest range determined by this disclosure is more accurate. Furthermore, by combining the target user's environmental information, the method explores the target user's interests, realizing a causal relationship between the user's role and consumption data. This allows for the recommendation of more interesting target information to the user, thereby reducing recommendation bias and improving the user experience.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0012] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is an exemplary system architecture to which this disclosure can be applied; Figure 2 A flowchart of an information recommendation method provided in this embodiment of the disclosure; Figure 3 A flowchart illustrating a method for determining the role information of a target user, provided in an embodiment of this disclosure; Figure 4A flowchart illustrating a method for constructing an exploration mind tree for a target user, as provided in this embodiment of the disclosure; Figure 5 A flowchart illustrating a method for generating and recommending target recommendation information provided in this embodiment of the disclosure; Figure 6 A structural block diagram of an information recommendation system provided in this disclosure embodiment; Figure 7 A structural block diagram of an information recommendation device provided in an embodiment of this disclosure; Figure 8 This is a schematic diagram of the structure of an electronic device suitable for performing an information recommendation method, provided as an embodiment of this disclosure. Detailed Implementation

[0013] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding; these should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description. It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0014] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0015] Figure 1 An exemplary system architecture 100 is shown, in which embodiments of the information recommendation methods, apparatuses, smart agents and electronic devices, computer-readable storage media and computer program products of this disclosure can be applied.

[0016] like Figure 1 As shown, system architecture 100 may include terminal device 101, terminal device 102, terminal device 103, network 104, and server 105. Network 104 is used as a medium to provide communication links between terminal device 101, terminal device 102, terminal device 103, and server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0017] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various applications for enabling information communication between the terminal devices 101, 102, and 103 and server 105 can be installed. These applications include video playback applications and e-commerce shopping applications.

[0018] Terminal devices 101, 102, 103, and server 105 can be hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with displays, including but not limited to smartphones, tablets, laptops, and desktop computers. When terminal devices 101, 102, and 103 are software, they can be installed in the aforementioned electronic devices, and can be implemented as multiple software programs or software modules, or as a single software program or software module; 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 server 105 is software, it can be implemented as multiple software programs or software modules, or as a single software program or software module; no specific limitation is made here.

[0019] Server 105 can provide various services through its built-in applications. Taking an e-commerce shopping application as an example, when running this application, server 105 can achieve the following: First, it acquires the target user's consumption data and environmental information through an intelligent agent. The target user is any one of multiple consumers, and their consumption data is obtained through interactions with server 105 via terminal devices 101, 102, and 103. Then, it analyzes the target user's consumption data based on a preset role database to determine the target user's role information. Next, it uses a mind-exploration intelligent agent to perform causal relationship analysis between the target user's role information and consumption data, obtaining an exploration mind tree. This tree clarifies the target user's corresponding role's interest range, which is more accurate than conventional methods that determine user interests solely based on the correlation between users and needs. Finally, it explores the target user's interests by combining their environmental information, realizing a causal relationship between user roles and consumption data. This allows it to recommend more relevant information to the user, reducing recommendation bias and improving the user experience.

[0020] Since processing consumption data from multiple users requires significant computing resources and power, the information recommendation methods provided in subsequent embodiments of this application are generally executed by a server 105 with strong computing power and abundant computing resources. Correspondingly, the information recommendation device is also generally located within the server 105. However, it should also be noted that when terminal devices 101, 102, and 103 also possess sufficient computing power and resources, they can also perform the aforementioned calculations performed by the server 105 through e-commerce shopping applications installed on them, thereby outputting the same results as the server 105. Accordingly, the information recommendation device can also be located within terminal devices 101, 102, and 103. In this case, the exemplary system architecture 100 may not include the server 105 and the network 104.

[0021] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0022] Please refer to Figure 2 , Figure 2 A flowchart of an information recommendation method provided in this disclosure embodiment, wherein process 200 includes the following steps: Step 201: Analyze the target user's consumption data based on the preset role database to determine the target user's role information.

[0023] The preset role library stores various types of preset roles. Each preset role represents a type of user's role in the information interaction process, and different preset roles correspond to different resource types.

[0024] Consumer data refers to the data obtained by target users through subscription, virtual consumption, or behavioral consumption of corresponding resources.

[0025] For example, subscription-based consumption involves target users subscribing to information related to a specific movie (A) on a video platform to obtain short video playback data for that movie. Virtual consumption involves target users using points / coupons / tokens to obtain virtual resource data; alternatively, target users may consume data through counted actions to obtain text resource data. Behavioral consumption involves target users obtaining the resource data they need after reaching certain thresholds in terms of page views, comments, or favorites.

[0026] This step is intended for the implementer of the information recommendation method (e.g., Figure 1The server 105 shown, which carries the agent, obtains the target user's consumption data (e.g., through...). Figure 1 The network 104 shown receives the consumption data of the user interacting with the server 105 through the terminal device 101, terminal device 102 or terminal device 103. Then, it analyzes the target user's consumption data through the information of various types of preset roles stored in the preset role library to determine the target user's role information.

[0027] In some embodiments, the target user's consumption data refers to the target user's consumption data within a historical period. For example, data on the target user's shopping activities on an e-commerce platform over the past year.

[0028] Step 202: Control the thinking exploration agent to perform causal relationship analysis on the target user's consumption data based on the target user's role information, and obtain an exploration thinking tree that includes at least one causal relationship link.

[0029] Among them, the thinking exploration agent is an intelligent agent that is based on a large model and can explore the causal relationship between user consumption behavior and user needs.

[0030] Building upon step 201, this step aims to process the target user's role information and consumption data into the mind-exploration agent, thereby analyzing and obtaining at least one causal relationship link. This causal relationship link will then reflect the causal relationship between the user's consumption behavior and user needs. Furthermore, an exploration mind tree will be constructed based on at least one causal relationship link to analyze the target user's interests.

[0031] For example, the obtained causal relationship chain is "target user consumption behavior - target user role information - target user interest information". Based on this causal relationship chain, the correlation between the target user's consumption behavior and role information can be accurately reflected. Compared with the traditional user interest scheme based on correlation statistics, it can more accurately obtain interest information that matches the target user's role information.

[0032] Step 203: Based on the target user's role information, exploration mind tree, and target user's environment information, explore the target user's interests to obtain target recommendation information for the target user.

[0033] Among these, the target user's environmental information refers to the scenario in which the target user makes certain consumption decisions in their current role. Exploring the target user's interests means that the aforementioned implementing entity queries or accesses information about the target user's interests that have not been discovered or understood.

[0034] The aforementioned execution entity formulates relevant scheduling strategies based on the target user's role information, exploration mind tree, and environmental information, attempting to query the target user's interests. These scheduling strategies may include: randomly selecting and executing certain actions; calling certain functions to search for the target user's interests; or performing operations such as query_sensor() / send_request() on the target user's unknown state to obtain the target user's interest information, thereby obtaining target recommendation information corresponding to that interest information.

[0035] For example, based on the brand of diapers a target user chooses during a particular purchase, it can be inferred that the target user is a "new dad" or "new mom," and thus determine the target user's environment as "the scenario of purchasing baby diapers."

[0036] Then, by combining at least one causal link in the mind tree, we can more accurately analyze that the target user was previously interested in basketball jerseys and shoes, and thus infer that the target user is a "new dad". We can further determine that the target recommendation information is product information needed by men, such as other basketball-related products (e.g., alcoholic beverages with a basketball star's signature).

[0037] Building upon step 202, this step aims to have the aforementioned executing entity combine information from multiple dimensions, such as the target user's role information, exploration mind tree, and the target user's environment, to explore the target user's interests, find information that the target user is more interested in, determine the target recommendation information, and recommend the target recommendation information to the target user. This will make the information pushed to the target user more in line with their usage needs and improve the user experience.

[0038] The information recommendation method provided in this disclosure analyzes the consumption data of target users based on a preset role library to determine the role information of the target users. It then employs a thinking exploration agent to perform causal relationship analysis on the target user's role information and consumption data to obtain an exploration mind tree. This method clearly defines the interest range of the target user's corresponding role. Compared to conventional methods that determine user interests solely based on the correlation between users and their needs, the interest range determined by this disclosure is more accurate. Furthermore, by combining the target user's environmental information, the method explores the target user's interests, realizing a causal relationship between the user's role and consumption data. This allows for the recommendation of more interesting target information to the user, thereby reducing recommendation bias and improving the user experience.

[0039] Please refer to Figure 3 , Figure 3 A flowchart illustrating a method for determining the role information of a target user, as provided in this embodiment of the disclosure, is shown below. Figure 2 Step 201 in process 200 provided a specific implementation. Other steps in process 200 are not adjusted; a new complete embodiment is obtained by replacing step 201 with the specific implementation provided in this embodiment. Process 300 includes the following steps: Step 301: Based on the role determination specification information and target user consumption data from the preset role library, label the target user's role and generate the target user's initial label information.

[0040] The preset role library stores multiple preset roles, each adhering to a unified role determination specification. This specification is information that can be consistently parsed by the system.

[0041] For example, the role determination specifications include, but are not limited to: identity and hierarchy specifications, permission specifications, resource characteristic specifications, and interest domain specifications.

[0042] The identity and hierarchy specification information refers to the user's identity and the corresponding interest hierarchy information. For example, if a user's identity is a viewer of a video file, and that video file is an anime-themed video, then the user's initial annotation information can be determined as the user's interest in anime.

[0043] Interest domain specification information indicates the scope of a user's interest domains. For example, if a user's role is a forum moderator, their corresponding interest domains include posting and following / unfollowing interest tags. Therefore, it can be determined that the user's initial tagging information indicates interest in posting.

[0044] Permission rules specify the set of actions a user is allowed (or prohibited from) performing. For example, a user can post or delete posts.

[0045] Resource characteristic specification information indicates the upper limit of resources that a user can obtain. For example, a user role has an exploration budget of 200, and a user has a search quota of 500 per day.

[0046] This step aims to have the aforementioned implementing entity determine the normative information and target user consumption data based on roles, analyze the scope of the target user's interests during the consumption process, and then, in conjunction with evaluation norms of different dimensions, initially label the target user's role so that the generated initial labeling information can reflect the target user's interests, as well as the resources and processing permissions that the target user can access within the scope of their corresponding role.

[0047] Step 302: The control role agent analyzes the correlation between the initial annotation information of the target user and the resource features corresponding to multiple preset roles in the preset role library, and determines the role information of the target user based on the initial annotation information of the target user.

[0048] Among them, the role-based intelligent agent is an intelligent agent implemented by a large model that is capable of performing role analysis on the target user.

[0049] Based on step 301, this step aims to have the aforementioned executing entity control the role intelligence agent to analyze the initial annotation information of the input target user, determine which type of preset role in the preset role library the initial annotation information of the target user specifically corresponds to, and then, after determining the corresponding preset role, specifically analyze whether the resource characteristics corresponding to the preset role match the initial annotation information. If the matching degree is determined to be high, the role information of the target user is determined according to the preset role.

[0050] For example, the initial labeling information of the target user is that the user is interested in anime. The character agent queries multiple preset characters in the preset character library to obtain the preset character "anime fan". Then, the resource characteristics corresponding to "anime fan" (such as posting a maximum of 50 bullet comments or short reviews related to anime topics per day, following a maximum of 30 anime character accounts, and collecting a maximum of 500 anime themes, etc.) are matched with the target user. If the target user posts nearly 50 bullet comments related to anime topics, and / or the target user follows 29 anime character accounts, then the target user's character information can be determined to be "anime fan".

[0051] By assigning roles to target users and controlling the role-based intelligent agent to analyze the correlation between the initial annotation information of the target users and the resource features corresponding to multiple preset roles, the role information of the target users can be enriched, and the accuracy of the role positioning of the target users can be improved.

[0052] In some embodiments, the preset role library includes multiple preset roles, and each preset role includes at least two preset role levels.

[0053] Step 302 can be implemented as follows: For each preset role, the initial annotation information of the target user is matched with each preset role level to obtain the role information of the target user.

[0054] Each preset role level corresponds to a different role degree. The role degree represents the degree of user classification. For example, a preset role includes three preset role levels (i.e., the first preset role level, the second preset role level, and the third preset role level). The first preset role level represents the major category to which the user belongs, the second level is a sub-level of the first preset role level, and the third preset role level is a sub-level of the second preset role level.

[0055] For example, if the first preset role level of a certain preset role is sports, the second preset role level is basketball, and the third level is information about NBA stars, then the preset role is "sports-basketball-NBA fan".

[0056] It should be noted that the second preset role level can also be other categories, such as football, volleyball, etc. The corresponding third preset role level is also divided into more detailed categories.

[0057] Based on step 301, this step aims to have the aforementioned executing entity match the initial annotation information of the target user with each preset role level of each preset role; if it is determined that the initial annotation information matches some preset role levels of a preset role, then the preset role level of that preset role is taken as the target role level of the target user.

[0058] The target user's role information includes at least two target role levels.

[0059] By employing multiple target role levels to represent the target user's role information, we can more precisely express the characteristics of the target user, thus preparing for subsequent interest exploration based on the target user's role information, and ensuring that more accurate information can be found.

[0060] In some embodiments, each preset role level corresponds to a resource feature, which represents the maximum total amount of resources that the preset role level can control; each target role level corresponds to an interest tag, which represents the proportion of resources that the target user is allowed to use at their target role level.

[0061] The resource characteristics corresponding to a preset role level represent the total amount of resources that the preset role level can obtain and access (or control). For example, a user at a certain preset role level can conduct a maximum of 500 searches per day on a search website, or a user at a certain preset role level can post a maximum of 10 long posts per day on a community forum (e.g., posts with more than 500 characters are considered long posts).

[0062] Once the target role level is determined, an interest tag is assigned to the target role level based on the resource characteristics of the preset role level. This interest tag is used to define a resource range within the resource characteristics of the preset role level that is less than or equal to the maximum total amount of resources that the preset role level can control. In other words, it defines the proportion of resources that are allowed to be used at the target role level.

[0063] For example, if a target user's role information is determined to include three target role levels (e.g., audience - audience of movie A - audience commenting on theme song B of movie A), where the resource feature corresponding to the third target role level is the number of comments made on theme song B of movie A, and since the resource feature corresponding to the third preset role level (community forum - forum of movie A - forum of theme song B of movie A) in the preset role library is a maximum of 20 posts, it can be determined that the target user can post a maximum of 20 posts, or less, in the forum of theme song B of movie A in the community forum.

[0064] By defining an interest tag for each target role level based on the resource characteristics of the preset role levels in the preset role library, we can more precisely identify the resource characteristics of each target user's target role level. This allows us to explore the target user's interests more accurately through these resource characteristics, thereby obtaining more accurate information about the target user's interests.

[0065] It should be noted that each preset role level in the preset role library corresponds to a resource feature, which can be updated in real time based on feedback from target users to meet the evaluation needs of different target users.

[0066] Please refer to Figure 4 , Figure 4 A flowchart illustrating a method for constructing an exploration mind tree for a target user, as provided in this embodiment of the disclosure, is shown below. Figure 2 Step 202 in process 200 provided a specific implementation. Other steps in process 200 are not adjusted; a new complete embodiment is obtained by replacing step 202 with the specific implementation provided in this embodiment. Process 400 includes the following steps: Step 401: Controlling the thinking exploration agent to break down the target user's consumption data based on the target user's role information, and obtain at least one consumption behavior of the target user.

[0067] This step aims to have the aforementioned executing entity break down the target user's consumption data based on the target user's role information, such as the number of target role levels corresponding to the target user and the interest tag information corresponding to the target role levels. This consumption data will then include at least one consumption behavior, such as the target user purchasing a blind box featuring a specific anime character. This achieves the matching of the target user's interest tag information with their consumption behavior, enabling a more refined analysis of the target user's interests.

[0068] Step 402: Controlling the thinking exploration agent to match at least one consumption behavior with the target user's role information to obtain the target user's pending interest tags.

[0069] This step aims to have the aforementioned executing entity input at least one consumption behavior into the mind exploration agent, so that the mind exploration agent can analyze the correspondence between the input at least one consumption behavior and the target user's role information, determine the target user's deeper interests, and output them in the form of interest tags to be confirmed.

[0070] Since the target user has a lot of consumption data, including multiple consumption behaviors (such as the target user purchasing a badge of anime character C on an e-commerce platform, purchasing a video membership on a video website, and watching multiple anime videos featuring anime character C on that website), these multiple consumption behaviors can be input into the mind exploration agent. This allows the mind exploration agent to perform multi-dimensional analysis of the target user's consumption data and determine the target user's interests (e.g., liking anime and being willing to spend money on anime character C). Then, it can be further determined that the target user's pending interest tag is an interest in anime character D. Here, anime character C and anime character D are two related characters in an anime work.

[0071] Step 403: Control the thinking exploration agent to construct the target user's exploration mind tree based on the target user's consumption behavior, target user's role information, and interest tags to be confirmed.

[0072] This step aims to have the aforementioned executing entity input the target user's consumption behavior, target user role information, and interest tags to be confirmed into the mind exploration agent. This allows the mind exploration agent to construct a mind exploration tree for the target user by combining information from multiple dimensions. This mind exploration tree can represent that the target user has multiple different interests, with corresponding branches for each interest. Based on this mind exploration tree, the target user's interests and hobbies can be more accurately identified, improving the accuracy of the analysis of the target user's needs.

[0073] In some embodiments, the target user's role information includes at least two target role levels, and each target role level corresponds to an interest tag information.

[0074] Step 402 can be implemented in the following way: For each consumption behavior, the control thinking exploration agent matches the consumption behavior with the interest tag information corresponding to each target role level to determine the interest tag to be confirmed.

[0075] This step is a further refinement of step 402. It aims to have the aforementioned executing entity control the thinking exploration agent to analyze each consumption behavior of the target user and further refine it to match the interest tag information corresponding to each target role level, thereby achieving more granular interest exploration and obtaining more accurate interest tags to be confirmed.

[0076] For example, if the target user's role information includes two target role levels (e.g., anime fan - fan of anime character C), then through the above analysis, it can be known that the target user's interest tag to be confirmed at the "anime fan" level is "interested in anime-related products"; and the interest tag to be confirmed at the "fan of anime character C" level is "interested in products related to anime character C" (and / or, interested in products related to anime character D)

[0077] By comparing the two levels of unconfirmed interest tags mentioned above, we can further refine whether the target user is interested in products related to "anime character C or D", thus achieving a more granular exploration of interests.

[0078] Please refer to Figure 5 , Figure 5 A flowchart illustrating a method for generating and recommending target recommendation information, provided in this embodiment of the disclosure, is shown. Figure 2 Step 203 in process 200 provided a specific implementation. Other steps in process 200 are not adjusted; a new complete embodiment is obtained by replacing step 203 with the specific implementation provided in this embodiment. Process 500 includes the following steps: Step 501: The control task planning agent matches the target user's role information and the target user's environment information with the various interest tags to be confirmed in the exploration mind tree to generate an interest exploration plan.

[0079] This step aims to have the aforementioned executing entity input the target user's role information, the target user's environment information, and the target user's exploration mind tree into the task planning agent. This allows the task planning agent to combine the information from these three dimensions to analyze and confirm each interest tag to be confirmed in the exploration mind tree, thereby obtaining an interest exploration plan to explore the target user's interests.

[0080] The interest exploration plan involves sorting the various interest tags to be confirmed according to their level of interest, selecting a preset number (e.g., 5) of interest tags that the target user is most interested in, and then formulating an interest exploration plan based on these preset number of interest tags to prepare for the subsequent execution of the interest exploration.

[0081] Step 502: Control the task execution agent to execute the interest exploration plan, explore the interests of the target user, and obtain target recommendation information.

[0082] This step aims to have the aforementioned executing agent input the interest exploration plan output by the task planning agent into the task execution agent, so that the task execution agent can execute the interest exploration plan, clarify which fine-grained information the target user is more interested in, and thus obtain target recommendation information.

[0083] Step 503: Push target recommendation information to target users.

[0084] This step aims to have the aforementioned implementing entity send the target recommendation information to the target user for their viewing.

[0085] When a target user is interested in the information provided in the target recommendation, they will take further action based on that information. For example, following the suggestions in the target recommendation, they might search for products related to the anime character "D" on an e-commerce website and then purchase those products.

[0086] In some embodiments, the method further includes: controlling a task planning agent to explore plans based on interests and determine target tools.

[0087] The target tools are used by the task-execution agent to carry out the interest exploration plan. For example, the target tools include the Model Context Protocol (MCP) tool, which is a standardized encapsulation of the actions that the task planning agent can execute.

[0088] For example, the target tool includes at least one application programming interface (API) used in the target environment.

[0089] The target tools include, but are not limited to: in the scenario of exploring web resources, the API called is: web page content grabbing function (web_browse(url)); in the scenario of querying enterprise common knowledge, the API called is: query enterprise knowledge base (query_knowledge(topic, top_k)), etc.

[0090] This step aims to enable interest exploration in different application scenarios through API calls.

[0091] In some embodiments, the method further includes: obtaining feedback information; and updating the interest exploration plan using the feedback information based on a logic graph modeling approach.

[0092] Among them, the logic graph-based modeling method constructs a graph with logical relationships by combining entities, attribute information, relationship information, and corresponding constraints to guide the updating of the interest exploration plan.

[0093] For example, entities can include, but are not limited to: roles, target users, interests, resources, etc.; attribute information includes, but is not limited to: role identifiers, weights, resource quotas, etc.; relationship information includes, but is not limited to: belonging relationships, ownership relationships, permitted relationships, etc.; constraint relationships include, but are not limited to: inheritance relationships, upper limit relationships, mutual exclusion relationships, etc. Feedback information refers to the consumption information of target users after obtaining target recommendation information. For example, feedback information may include consumption information of target users purchasing products related to anime character D on an e-commerce platform after viewing target recommendation information that includes related products.

[0094] This step aims to update the interest exploration plan based on feedback information, thereby enabling the task-execution agent to self-update.

[0095] After the interest exploration plan is updated, the aforementioned execution entity can execute step 502 again, controlling the task execution agent to execute the updated interest exploration plan, further explore the interests of the target user, and obtain more accurate target recommendation information.

[0096] In some embodiments, the interest exploration plan is updated using feedback information based on logical graph modeling, including: updating the logical graph model based on the causal logical relationship between the target user's role information and the feedback information; and updating the interest exploration plan according to at least one node in the updated logical graph model and at least one edge representing the logical relationship between two adjacent nodes.

[0097] The logic graph model is constructed based on the target user's role information and the exploration mind tree. Each node represents at least one of the following: the target user's order, the event that occurred, or the device used.

[0098] Edges represent the logical relationships between two adjacent nodes, such as: belonging, dependency, cause, call, restriction, flow, etc.

[0099] This step aims to update the weights of nodes and / or edges between nodes in the logical graph model based on the causal relationship between the target user's role information and feedback information, so that the updated logical graph model is more in line with the current target user's usage scenario.

[0100] Furthermore, the interest exploration plan is updated using an updated logic graph model to make the updated plan more logical and suitable for exploring the interests of target users, thereby improving the accuracy of exploring the interests of target users.

[0101] In some embodiments, the method further includes updating the exploration mind tree based on feedback information.

[0102] In particular, after receiving feedback, it can be clearly seen that the needs of the target users have changed. The exploration mind tree of the target users can be updated based on the change information carried in the feedback information, so that the updated exploration mind tree can better meet the current usage needs of the target users, so as to obtain a more accurate interest exploration plan based on the updated exploration mind data.

[0103] For example, if the feedback information indicates that the target user has further purchased merchandise related to comic character D, the analysis shows that the target user is also interested in comic character D. Therefore, the relevant information of comic character D can be used as a level at the same level as the target role level "fan of anime character C", thus obtaining the target role level of "fan of anime character D". Then, this new target role level is added to the target user's exploration mind tree as part of the target user's role information, thereby obtaining an updated exploration mind tree and refining the branches of the exploration mind tree.

[0104] In some embodiments, the method further includes updating a preset role library based on feedback information and / or the role information of the target user.

[0105] Updating the preset role library based on feedback information and / or updating the preset role library based on the target user's role information are both aimed at enriching the preset roles in the preset role library and meeting the role positioning needs of different users.

[0106] For example, if the feedback information indicates that the target user purchased merchandise related to comic character D, the analysis shows that the target user is also interested in comic character D. Therefore, it can be determined that the target user's role information also includes the level of "fan of anime character C". Then, the preset roles in the preset role library are updated according to the updated target user's role information.

[0107] In some embodiments, based on the updated target user's role information, any one of the following operations can be performed on the preset role level of the preset role in the preset role library: add operation, delete operation, or modify operation.

[0108] To enhance understanding, this disclosure also provides a specific implementation scheme in conjunction with a particular application scenario. Figure 6 This is a structural block diagram of an information recommendation system provided in an embodiment of the present disclosure.

[0109] like Figure 6 As shown, the system recommends the following devices, including but not limited to: role agent 610, thinking exploration agent 620, task planning agent 630, and task execution agent 640.

[0110] The role-based intelligent agent 610 includes a labeling unit 611, a first analysis unit 612, a first architecture management unit 613, and a preset role library 614.

[0111] The annotation unit 611 is used to annotate the roles of the target user according to the role determination specification information of the preset role library 614 and the consumption data of the target user, generate the initial annotation information of the target user, and send the initial annotation information of the target user to the first analysis unit 612.

[0112] The preset role library 614 includes multiple preset roles, each preset role includes at least two preset role levels, and each preset role level corresponds to a resource feature, which represents the maximum total amount of resources that the preset role level can control.

[0113] In some embodiments, the annotation unit 611 is further configured to generate improvement suggestions based on information that cannot be annotated, and send the improvement suggestions to the first analysis unit 612, so that the first analysis unit 612 can forward the improvement suggestions to the first architecture management unit 613, and the first architecture management unit 613 can then update the preset role library 614 based on the improvement suggestions.

[0114] For example, extract information that cannot be marked in the improvement suggestions, and create new preset roles based on that information, and / or create new preset role levels.

[0115] The annotation unit 611 can perform preliminary annotation of the target user's role based on the preset role library 614, thereby achieving a deeper understanding of the target user's role and preparing for subsequent exploration of the target user's interests.

[0116] The first analysis unit 612 is used to analyze the correlation between the initial annotation information received from the target user and the resource features corresponding to multiple preset roles in the preset role library 614, and to determine the role information of the target user.

[0117] The target user's role information includes at least two target role levels, each target role level corresponds to an interest tag, and the interest tag indicates the proportion of resources that the target user is allowed to use at their target role level.

[0118] It should be noted that, during the analysis of the initial annotation information of the target user, the first analysis unit 612 can query multiple preset roles that match the initial annotation information. Then, for each preset role level of each preset role, the interest tag information corresponding to the target role level of the target user is obtained by matching.

[0119] In other words, the interest tag information corresponding to the target role level is the proportion (e.g., percentage) of resources that the target user can be allowed to use within the range of resource characteristics corresponding to the preset role level.

[0120] For example, if the initial labeling information of the target user indicates that the target user is interested in sports and has purchased basketball jerseys and related products from the National Basketball Association (NBA), and the first analysis unit 612 determines that the target user's initial labeling information corresponds to the third-level preset role in the preset role library 614 (e.g., sports-basketball-NBA fan), then it can be determined that the target user is highly likely to be an NBA fan and may be interested in NBA-related merchandise (e.g., NBA jerseys, soccer balls, autographed merchandise, etc.). Therefore, it can be determined that the target user's interest tag information represents 80% of the resources allowed at the "sports-basketball-NBA fan" target role level (e.g., accessing basketball-NBA related product recommendation information in the system).

[0121] The first architecture management unit 613 is used to maintain and update the preset role library 614, receive the target user's role information sent by the first analysis unit 612, and perform any of the following operations on the preset roles in the preset role library 614 through reflection verification: add operation, delete operation, modify operation.

[0122] The thinking exploration intelligent agent 620 includes a second analysis unit 621 and a second architecture management unit 623.

[0123] During the training process, the second analysis unit 621 can use positive sample data as the basis for exploration. Positive sample data refers to the consumption data of sample users who have consistently made stable purchases. For example, if a sample user has consistently purchased anime products throughout a historical period (e.g., in the past year, the sample user has consistently purchased blind boxes, souvenirs, etc. related to anime characters), then the consumption data of that sample user can be preprocessed to obtain positive sample data.

[0124] The second analysis unit 621 analyzes and decomposes the positive sample data to obtain at least one causal relationship link. This causal relationship link can be represented as "sample user consumption behavior - sample user role information - sample user interest information". For example, consumption behavior of anime products related to a certain anime character - sample user is an anime fan - sample user's interest in offline anime conventions. Offline anime conventions can exhibit or sell souvenirs such as commemorative badges related to anime characters. The information that sample users are interested in at offline anime conventions can also be behind-the-scenes videos of anime character filming, which can be uploaded to video applications for sample users to watch.

[0125] Then, the second analysis unit 621 sends at least one causal link obtained from the above analysis to the second architecture management unit 623.

[0126] The second architecture management unit 623 constructs an exploration mind tree based on at least one causal relationship link, and updates and maintains the exploration mind tree. For example, based on the causal relationship link sent in real time by the second analysis unit 621, it performs any of the following operations on the branches in the exploration mind tree: modification operation, deletion operation, addition operation, and query operation.

[0127] In the application scenario, based on the target user's role information input by the role agent 610, the thinking exploration agent 620 performs causal relationship analysis on the target user's consumption data to obtain an exploration mind tree for the target user, including at least one causal relationship link. Then, this exploration mind tree is sent to the task planning agent 630.

[0128] The task planning agent 630 includes a central control unit 631, which is used to receive the target user's role information input by the role agent 610, receive the target user's exploration thought tree including at least one causal relationship link input by the thinking exploration agent 620, receive the target user's environmental information fed back by the task execution agent 640, and explore the target user's interests based on the above multiple information to obtain target recommendation information.

[0129] Based on the environmental information of the target user, the task planning agent 630 can formulate different exploration strategies for different application scenarios.

[0130] For example, different application scenarios include: community scenarios, local scenarios, and TV series scenarios. Community scenarios include: comment communities, open-source communities, interest groups, etc.; local scenarios refer to situations where the target user's device is not connected to a communication network and only uses local resources, such as data read / write operations on a local device. TV series scenarios include: streaming media platforms, anime character fan communities, TV series knowledge bases, etc.

[0131] In some embodiments, the task planning agent 630 can also obtain feedback information from other agents and further refine the interest exploration strategy for the target user based on this feedback information. For example, the feedback information includes information from the content agent (not shown in the figure) about the target user's video viewing, and information about the target user's consumption after obtaining the previous round of target recommendation information.

[0132] Specifically, the task planning agent 630 matches the target user's role information and the target user's environment information with the exploration mind tree to generate an interest exploration plan; then, based on the interest exploration plan, it determines the target tool, which is used by the task execution agent 640 to call.

[0133] In some embodiments, the target tool includes at least one API used in the target environment. For example, the target tool includes an MCP tool, which is a standardized encapsulation of the actionable actions of the task planning agent 630. For example, the target tool includes, but is not limited to: webpage content capture functionality (web_browse(url)), vector retrieval (vector_search(index,query)), returning role resource features (list_role_resources(role_id)), and querying the enterprise knowledge base (query_knowledge(topic, top_k)), etc.

[0134] The exploration mind tree includes multiple interest tags to be confirmed. Each interest tag is determined by the mind exploration agent 620 by matching the consumption behavior of the target user with the interest tag information corresponding to each target role level.

[0135] The task execution intelligent agent 640 includes a central control unit 641, a logic diagram model 642, and an environmental information acquisition unit 643.

[0136] The central control unit 641 is used to invoke the target tools provided by the task planning agent 630, execute the interest exploration plan, explore the interests of the target user, obtain target recommendation information, and push the target recommendation information to the target user.

[0137] In the process of executing the interest exploration plan, the central control unit 641 can also combine the multiple nodes provided by the logic graph model 642 and the edges between each node, as well as the feedback information input by other intelligent agents (such as content intelligent agents), to explore the interests of the target user and obtain more accurate target recommendation information.

[0138] The logic graph model 642 is initially constructed based on the target user's role information and exploration mind tree. As the exploration of the target user's interests deepens, the logic graph model 642 is updated based on the causal logical relationship between the target user's role information and feedback information to obtain an updated logic graph model, thereby strengthening the relationship between the nodes in the updated logic graph model.

[0139] The logical graph model 642 includes multiple nodes and edges between them. Each node represents at least one of the following: the target user's order, an event that occurred, or a device used.

[0140] An edge between two adjacent nodes represents the logical relationship between the two nodes. This logical relationship includes, but is not limited to: belonging relationship, dependency relationship, cause relationship, call relationship, restriction relationship, flow relationship, etc.

[0141] The environmental information acquisition unit 643 is used to collect environmental information during the process of a target user performing consumption actions or other actions (such as commenting, watching videos, etc.) in different application scenarios. For example, the collected environmental information of the target user includes: user sentiment information in the scenario of the target user commenting on a popular topic, the intelligent agent information used by the target user when performing data read and write operations in the local scenario, the duration of the target user watching video files on the streaming media platform, and the evaluation information of other users on the video file, etc.

[0142] In this embodiment, the environmental information collection unit in the task execution agent collects the environmental information of the target user, enabling perception and contextual understanding of the environment in which the target user consumes information, quickly responding to the target user's dynamic needs, and making the target recommendation information recommended to the target user more practical. Furthermore, by combining the logic graph model to update the causal relationship between the feedback information and the target user's role information, iterative updates of the target user's role information can be achieved, thereby breaking the information cocoon, focusing on the target user's long-term needs, and improving the target user's user experience.

[0143] In this embodiment, a role-based intelligent agent 610 analyzes the target user's consumption data to determine the target user's role information. A pre-defined role library, constructed through multi-agent reflection, significantly improves the accuracy of demand identification and adapts to the behavioral logic of different roles. Then, a thinking exploration intelligent agent 620, combined with the target user's role information input by the role-based intelligent agent 610, further analyzes the target user's consumption data to obtain an exploration thinking tree containing at least one causal link, enabling precise insight into the true motivations behind the target user's behavior. Furthermore, a task planning intelligent agent 630 and a task execution intelligent agent 640, based on the target user's role information, the exploration thinking tree, and the target user's environment, explore the target user's interests to obtain target recommendation information for the target user, avoiding rigid recommendation logic. Moreover, by employing a multi-agent architecture to achieve efficient collaboration between role construction, causal mining, task planning, and execution, it is applicable to various application scenarios, recommending target information that is more of the target user's interest, thus improving the user experience.

[0144] Further reference Figure 7 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of an information recommendation device, which is similar to... Figure 2 The information recommendation method embodiments shown correspond to those described. The above-described apparatus can be specifically applied to various electronic devices.

[0145] like Figure 7 As shown, the information recommendation device 700 of this embodiment may include: a role determination unit 710, a causal relationship exploration unit 720, and an information recommendation unit 730.

[0146] The role determination unit 710 is configured to analyze the consumption data of the target user based on a preset role library to determine the role information of the target user.

[0147] The causal relationship exploration unit 720 is configured to control the thinking exploration agent to perform causal relationship analysis on the target user's consumption data based on the target user's role information, and obtain an exploration thinking tree that includes at least one causal relationship link.

[0148] The information recommendation unit 730 is configured to explore the target user's interests based on the target user's role information, exploration mind tree, and target user's environment information, and obtain target recommendation information for the target user to use.

[0149] In this embodiment, the specific processing of the role determination unit 710, the causal relationship exploration unit 720, and the information recommendation unit 730 in the information recommendation device 700, and the resulting technical effects, can be referred to respectively. Figure 2The relevant descriptions of steps 201-203 in the corresponding embodiments will not be repeated here.

[0150] In some optional implementations of this embodiment, the role determination unit 710 may include: a labeling module, configured to label the roles of the target user according to the role determination specification information of the preset role library and the consumption data of the target user, and generate the initial labeling information of the target user; and an analysis module, configured to control the role agent to analyze the correlation between the initial labeling information and the resource features corresponding to multiple preset roles in the preset role library according to the initial labeling information of the target user, and determine the role information of the target user.

[0151] In some optional implementations of this embodiment, the preset role library includes multiple preset roles, and each preset role includes at least two preset role levels; The analysis module can be further configured to: for each preset role, match the initial annotation information of the target user with each preset role level to obtain the role information of the target user; wherein the role information of the target user includes at least two target role levels.

[0152] In some optional implementations of this embodiment, each preset role level corresponds to a resource feature, which represents the maximum total amount of resources that the preset role level can control; each target role level corresponds to an interest tag, which represents the proportion of resources that the target user is allowed to use at their target role level.

[0153] In some optional implementations of this embodiment, the causal relationship exploration unit 720 may include: a decomposition module, configured to control the thinking exploration agent to decompose the target user's consumption data according to the target user's role information, and obtain at least one consumption behavior of the target user; The matching module is configured to control the thinking exploration agent to match at least one consumption behavior with the target user's role information to obtain the target user's interest tags to be confirmed. The module is configured to control the thinking exploration agent to construct the target user's exploration mind tree based on the target user's consumption behavior, the target user's role information, and interest tags to be confirmed.

[0154] In some optional implementations of this embodiment, the target user's role information includes at least two target role levels, and each target role level corresponds to an interest tag information; The matching module can be further configured to: for each consumption behavior, control the thinking exploration agent to match the consumption behavior with the interest tag information corresponding to each target role level, and determine the interest tags to be confirmed.

[0155] In some optional implementations of this embodiment, the information recommendation unit 730 may include: a plan generation module, configured to control the task planning agent to match the target user's role information and the target user's environment information with each interest tag to be confirmed in the exploration mind tree to generate an interest exploration plan; The exploration module is configured to control the task-execution agent to carry out an interest exploration plan, explore the interests of the target user, and obtain target recommendation information; The push module is configured to push targeted recommendation information to target users.

[0156] In some optional implementations of this embodiment, the information recommendation device 700 further includes: a tool determination unit, configured to control the task planning agent to determine a target tool according to the interest exploration plan; wherein the target tool is used for the task execution agent to call to execute the interest exploration plan.

[0157] In some optional implementations of this embodiment, the target tool includes at least one application programming interface (API) used in the target environment.

[0158] In some optional implementations of this embodiment, the information recommendation device 700 further includes: an acquisition unit configured to acquire feedback information; the feedback information is information about the target user consuming the target recommendation information after acquiring it; The first update unit is configured to update the interest exploration plan using a logic graph modeling approach and feedback information.

[0159] In some optional implementations of this embodiment, the first updating unit is configured to: update the logic graph model based on the causal logical relationship between the target user's role information and feedback information; the logic graph model is a model constructed based on the target user's role information and exploration mind tree; and update the interest exploration plan based on at least one node in the updated logic graph model and at least one edge representing the logical relationship between two adjacent nodes. In this context, a node represents at least one of the following: a target user's order, an event that occurred, or a device used.

[0160] In some optional implementations of this embodiment, the information recommendation device 700 further includes a second updating unit configured to update the exploration mind tree based on feedback information.

[0161] In some optional implementations of this embodiment, the information recommendation device 700 further includes a third update unit configured to update the preset role library based on feedback information and / or the role information of the target user.

[0162] This embodiment exists as a device embodiment corresponding to the above method embodiment. The information recommendation device provided in this embodiment analyzes the consumption data of the target user according to a preset role library to determine the target user's role information, and uses a thinking exploration agent to perform causal relationship analysis on the target user's role information and consumption data to obtain an exploration mind tree. This can clearly define the interest range of the target user's corresponding role. Compared with the conventional method of determining user interests based only on the correlation between users and needs, the interest range of the target user determined by this disclosure is more accurate. Then, combined with the target user's environmental information, the target user's interests are explored, realizing the causal relationship between the user's role and the user's consumption data, and recommending target recommendation information that the user is more interested in, thereby reducing recommendation bias and improving the user experience.

[0163] According to embodiments of this disclosure, this disclosure also provides an intelligent agent, including: an input module for receiving input information; a processing module for determining a target task based on the input information received by the input module, determining an information recommendation model based on the target task, and executing the information recommendation method described in any of the above embodiments by calling the information recommendation model to obtain output information; and an output module for outputting the output information obtained by the processing module.

[0164] According to embodiments of this disclosure, this disclosure also provides an electronic device, the 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, the instructions being executed by the at least one processor to enable the at least one processor to implement the information recommendation method described in any of the above embodiments when executed.

[0165] According to embodiments of this disclosure, this disclosure also provides a readable storage medium storing computer instructions that, when executed by a computer, enable the information recommendation method described in any of the above embodiments.

[0166] This disclosure provides a computer program product that, when executed by a processor, can implement the information recommendation method described in any of the above embodiments.

[0167] Figure 8This is a schematic diagram of an electronic device suitable for performing an information recommendation method, provided as an embodiment of this disclosure. 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 may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0168] like Figure 8 As shown, the electronic device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. The RAM 803 may also store various programs and data required for the operation of the device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0169] Multiple components in device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of monitors, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0170] The computing unit 801 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as information recommendation methods. For example, in some embodiments, the information recommendation method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the information recommendation method described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform information recommendation methods by any other suitable means (e.g., by means of firmware).

[0171] Various embodiments 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-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0172] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0173] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0174] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, 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 sound input, voice input, or tactile input).

[0175] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0176] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and Virtual Private Server (VPS) services, such as high management difficulty and weak business scalability.

[0177] The technical solution of this disclosure analyzes the consumption data of target users based on a preset role library to determine the role information of the target users. It then employs a thinking exploration agent to perform causal relationship analysis on the role information and consumption data of the target users to obtain an exploration mind tree. This clearly defines the interest range of the target user's corresponding role. Compared to conventional methods that determine user interests solely based on the correlation between users and their needs, this disclosure determines the target user's interest range more accurately. Furthermore, by combining the target user's environmental information, the disclosure explores the target user's interests, realizing a causal relationship between the user's role and consumption data. This allows for the recommendation of more interesting target information to the user, thereby reducing recommendation bias and improving the user experience.

[0178] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0179] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An information recommendation method, comprising: The target user's role information is determined by analyzing the target user's consumption data based on a preset role database. The control thinking exploration agent performs causal relationship analysis on the target user's consumption data based on the target user's role information, and obtains an exploration thinking tree that includes at least one causal relationship link; Based on the target user's role information, the exploration mind tree, and the target user's environment information, the target user's interests are explored to obtain target recommendation information for the target user.

2. The method of claim 1, wherein, The step of analyzing the target user's consumption data based on a preset role database to determine the target user's role information includes: Based on the role determination specification information of the preset role library and the consumption data of the target user, the roles of the target user are labeled, and the initial labeling information of the target user is generated; The control role agent analyzes the correlation between the initial annotation information of the target user and the resource features corresponding to multiple preset roles in the preset role library, and determines the role information of the target user based on the initial annotation information of the target user.

3. The method of claim 2, wherein, The preset role library includes multiple preset roles, and each preset role includes at least two preset role levels; The control role agent analyzes the correlation between the initial annotation information of the target user and the resource features corresponding to multiple preset roles in the preset role library, based on the initial annotation information of the target user, to determine the role information of the target user, including: For each preset role, the initial annotation information of the target user is matched with each preset role level to obtain the role information of the target user; The target user's role information includes at least two target role levels.

4. The method according to claim 3, wherein, Each of the preset role levels corresponds to a resource feature, and the resource feature represents the maximum total amount of resources that the preset role level can control. Each target role level corresponds to an interest tag, which represents the proportion of resources that the target user is allowed to use at their target role level.

5. The method according to claim 1, wherein, The control-oriented thinking exploration agent performs causal relationship analysis on the target user's consumption data based on the target user's role information, obtaining an exploration thinking tree that includes at least one causal relationship link, including: The intelligent agent controlling the thinking exploration is to break down the target user's consumption data based on the target user's role information, and obtain at least one consumption behavior of the target user; The intelligent agent is controlled to match the at least one consumption behavior with the role information of the target user to obtain the target user's interest tags to be confirmed; The intelligent agent controlling the thinking exploration constructs the target user's exploration mind tree based on the target user's consumption behavior, the target user's role information, and the interest tags to be confirmed.

6. The method according to claim 5, wherein, The target user's role information includes at least two target role levels, and each target role level corresponds to an interest tag information; The control mechanism of the thought exploration agent matches at least one consumption behavior with the target user's role information to obtain the target user's pending interest tags, including: For each of the aforementioned consumption behaviors, the intelligent thinking agent is controlled to match the consumption behavior with the interest tag information corresponding to each of the target role levels to determine the interest tag to be confirmed.

7. The method according to claim 5, wherein, The step of exploring the target user's interests based on the target user's role information, the exploration mind tree, and the target user's environment information to obtain target recommendation information for the target user includes: The control task planning agent matches the target user's role information and the target user's environment information with each interest tag to be confirmed in the exploration mind tree to generate an interest exploration plan; The control task execution agent executes the interest exploration plan, explores the interests of the target user, and obtains the target recommendation information; The target recommendation information is pushed to the target user.

8. The method according to claim 7, wherein, The method further includes: The task planning agent is controlled to explore the plan based on the interests and determine the target tool; The target tool is used by the task execution agent to execute the interest exploration plan.

9. The method according to claim 8, wherein, The target tool includes at least one application programming interface (API) used in the target environment.

10. The method according to any one of claims 7 to 9, wherein, The method further includes: Obtain feedback information; the feedback information is the information about the target user's consumption after obtaining the target recommendation information; The interest exploration plan is updated using the feedback information based on a logic graph modeling approach.

11. The method according to claim 10, wherein, The logic graph-based modeling method, which uses the feedback information to update the interest exploration plan, includes: The logic graph model is updated based on the causal relationship between the target user's role information and the feedback information; the logic graph model is a model constructed based on the target user's role information and the exploration mind tree. The interest exploration plan is updated based on at least one node in the updated logical graph model and at least one edge representing the logical relationship between two adjacent nodes. The node represents at least one of the following: the target user's order, the event that occurred, or the device used.

12. The method according to claim 10, wherein, The method further includes: The exploration mind tree is updated based on the feedback information.

13. The method according to claim 10, wherein, The method further includes: Update the preset role library based on the feedback information and / or the role information of the target user.

14. An information recommendation device, comprising: The role determination unit is configured to analyze the consumption data of the target user based on a preset role library to determine the role information of the target user; The causal relationship exploration unit is configured to control the thinking exploration agent to perform causal relationship analysis on the target user's consumption data based on the target user's role information, and obtain an exploration thinking tree including at least one causal relationship link; The information recommendation unit is configured to explore the target user's interests based on the target user's role information, the exploration mind tree, and the target user's environment information, and obtain target recommendation information for the target user to use.

15. An intelligent agent, comprising: The input module is used to receive input information; The processing module is configured to determine a target task based on the input information received by the input module, determine an information recommendation model based on the target task, and execute the information recommendation method according to any one of claims 1 to 13 by calling the information recommendation model to obtain output information; An output module is used to output the output information obtained by the processing module.

16. 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 to enable the at least one processor to perform the information recommendation method according to any one of claims 1 to 13.

17. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the information recommendation method according to any one of claims 1 to 13.

18. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the information recommendation method according to any one of claims 1 to 13.

Images