Information recommendation method, device, electronic device and computer readable storage medium

Abstract

The present application discloses an information recommendation method, device, electronic device and computer-readable storage medium, the method comprising: extracting initial attribute information associated with a target object from the acquired target information to be processed; updating the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; determining the target recommendation information according to the action attribute in the target attribute information, and sending the target recommendation information to the target object end. The above scheme first updates the current attribute information of the target object according to the initial attribute information, which can improve the timeliness of the target object information update and ensure the accuracy of the target attribute information; then recommends information based on the action attribute in the updated target attribute information, so that the recommended information is more in line with the intention of the target object, and improves the accuracy of the information recommendation.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an information recommendation method, an information recommendation device, an electronic device and a computer-readable storage medium. Background Art

[0002] With the rapid development of big data processing technology, integrating big data to obtain the preferences of target objects and providing information recommendations to target objects based on their preferences is a very effective personalized service method that can significantly improve the experience and satisfaction of target objects.

[0003] At present, traditional information recommendation services mainly rely on analyzing active search records to determine the preferences of target objects. The recommended content generated by this method is too broad, which can easily lead to information overload and affect the target object's experience. At the same time, the triggering of information recommendation is subject to set conditions, such as setting a time threshold to recommend information under the set time threshold. Summary of the invention

[0004] The main technical problem solved by the present application is to provide an information recommendation method, an information recommendation device, an electronic device and a computer-readable storage medium, which can improve the accuracy of information recommendation.

[0005] To solve the above technical problems, a technical solution adopted in the present application is: to provide an information recommendation method, the method comprising: extracting initial attribute information associated with a target object from the acquired target information to be processed; updating the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; determining the target recommendation information according to the action attributes in the target attribute information, and sending the target recommendation information to the target object end.

[0006] In some embodiments, the step of extracting initial attribute information associated with the target object from the acquired target information to be processed includes: inputting the target information to be processed and first preset prompt information into a large model to obtain attribute information associated with the target object output by the large model; and determining the attribute information associated with the target object output by the large model as the initial attribute information.

[0007] In some embodiments, the step of updating the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object includes: inputting the initial attribute information, the current attribute information and the second preset prompt information into a large model to obtain updated attribute information output by the large model; and determining the updated attribute information output by the large model as the target attribute information of the target object.

[0008] In some embodiments, the method further includes: inputting the initial attribute information, the first preset prompt information and the preset evaluation prompt information into the large model for evaluation optimization processing to obtain the optimization information of the first preset prompt information; optimizing the first preset prompt information according to the optimization information of the first preset prompt information to obtain the first preset prompt information after optimization processing, and the first preset prompt information after optimization processing is used by the large model to extract attribute information of the next target information to be processed.

[0009] In some embodiments, the action attributes include at least one action intention and a corresponding intention intensity. The step of determining the target recommendation information based on the action attributes in the target attribute information and sending the target recommendation information to the target object end includes: determining the action attributes of the target object from the target attribute information; determining the target recommendation information based on the intention intensity of each action intention in the action attributes of the target object, and sending the target recommendation information to the target object end.

[0010] In some embodiments, the step of determining the target recommendation information based on the intention intensity of each action intention in the action attributes of the target object, and sending the target recommendation information to the target object end, includes: judging whether the intention intensity of each action intention in the action attributes of the target object is greater than a preset intensity threshold; if so, determining the target recommendation information based on the corresponding action intention, and sending the target recommendation information to the target object end.

[0011] In some embodiments, the step of extracting initial attribute information associated with the target object from the acquired target information to be processed includes: acquiring initial information to be processed, wherein the initial information to be processed includes information to be processed of the target object and information to be processed of other objects;

[0012] In response to receiving a first mode instruction, the initial information to be processed is determined as the target information to be processed, and the initial attribute information associated with the target object is extracted from the target information to be processed; in response to receiving a second mode instruction, the information to be processed of the target object is extracted from the initial information to be processed, and the information to be processed of the target object is determined as the target information to be processed, and the initial attribute information associated with the target object is extracted from the target information to be processed.

[0013] To solve the above technical problems, another technical solution adopted by the present application is: an information recommendation device is provided, the device comprising: an extraction module, used to extract initial attribute information associated with a target object from the acquired target information to be processed; an update module, used to update the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; a recommendation module, used to determine the target recommendation information according to the action attributes in the target attribute information, and send the target recommendation information to the target object end.

[0014] To solve the above technical problem, another technical solution adopted in the present application is: to provide an electronic device, including a memory and a processor, wherein the memory stores program instructions, and the processor retrieves the program instructions from the memory to execute the above information recommendation method.

[0015] In order to solve the above technical problem, another technical solution adopted by the present application is: providing a computer-readable storage medium including program data stored therein, wherein the program data is used to implement the above information recommendation method when executed by a processor.

[0016] Beneficial effects of the present application: The information recommendation method of the embodiment of the present application extracts the initial attribute information associated with the target object from the acquired target information to be processed; updates the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; determines the target recommendation information according to the action attribute in the target attribute information, and sends the target recommendation information to the target object end. The above scheme first updates the current attribute information of the target object according to the initial attribute information, which can improve the timeliness of the target object information update and ensure the accuracy of the target attribute information; then recommends information based on the action attribute in the updated target attribute information, so that the recommended information is more in line with the intention of the target object, and improves the accuracy of the information recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0018] Figure 1 is a flowchart of an exemplary embodiment of the information recommendation method shown in the present application;

[0019] Figure 2 yes Figure 1 A flowchart of an exemplary embodiment of step S110 in the information recommendation method is shown;

[0020] Figure 3 It is a schematic diagram of a framework of an exemplary embodiment of the attribute information update processing process shown in the present application;

[0021] Figure 4 It is a schematic diagram of a framework of an exemplary embodiment of the prompt information evaluation optimization processing process shown in the present application;

[0022] Figure 5 It is a schematic diagram of a framework of an exemplary embodiment of the information recommendation method shown in the present application;

[0023] Figure 6 is a structural diagram of an exemplary embodiment of an information recommendation device shown in the present application;

[0024] Figure 7 It is a structural schematic diagram of an embodiment of an electronic device provided by the present application;

[0025] Figure 8 It is a structural schematic diagram of an embodiment of a computer-readable storage medium provided by the present application. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It will be appreciated that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only some but not all structures related to the present application are shown in the drawings. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the art without making creative work are within the scope of protection of the present application.

[0027] This application proposes an information recommendation method, an information recommendation device, an electronic device, and a computer-readable storage medium, which can improve the accuracy of information recommendation. Figure 1 , Figure 1 It is a flowchart of an exemplary embodiment of the information recommendation method shown in the present application.

[0028] The execution subject of the information recommendation method may be a terminal device or a server or other processing device, wherein the terminal device may be a user equipment (UE), a computer, a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. The execution subject of the information recommendation method may also be an information recommendation device. In some possible implementations, the information recommendation method may be implemented by a processor calling a computer-readable instruction stored in a memory.

[0029] Specifically, the information recommendation method of this embodiment includes the following steps:

[0030] S110: extracting initial attribute information associated with the target object from the acquired target information to be processed.

[0031] The target information to be processed refers to the information of the initial attribute information associated with the target object to be extracted. Exemplarily, it can be obtained through user submission, system logs and other channels. In some embodiments, the target information to be processed can be dialogue information authorized by the user, such as human-computer dialogue, or it can be description information of the target object, such as a text, a picture, etc. The target information to be processed can include text, images, videos, etc., which are unstructured information.

[0032] The target object refers to the service object of the information recommendation. For example, the target object can be a person, an animal, etc.

[0033] Initial attribute information refers to information associated with the target object that can be obtained from the target information to be processed. Exemplarily, the initial attribute information can be extracted from the target information to be processed by preset rules, including keyword extraction, model analysis, and statistical analysis. When the target object is a person, the initial attribute information can also be called a user portrait, which is actively submitted by the user after authorization, including but not limited to the following attributes:

[0034] Basic information: gender, age, place of origin, education level, place of residence, occupation, income level, birthday, etc.;

[0035] Status information: health status, marital / relationship status, emotional status, etc.;

[0036] Action attributes: consumption level, action intention, intensity of intention, etc.

[0037] Preference attributes: hobbies, living habits, pets, etc.;

[0038] Personality attributes: character, values, etc.;

[0039] Other attributes: Other clear attribute labels besides the above attributes, such as: vehicle, family population, etc.

[0040] The information recommendation device obtains target information to be processed, extracts attribute information from the target information to be processed, and obtains initial attribute information associated with the target object.

[0041] S120: updating the current attribute information of the target object according to the initial attribute information to obtain target attribute information of the target object.

[0042] Current attribute information refers to the attribute information known to the target object. Exemplarily, the current attribute information can be extracted based on the historical processing information of the target object, or it can be information actively submitted by the target object. In practical applications, the current attribute information of the target object may contain errors or lack information. Therefore, when the information recommendation device receives new target information to be processed, it is necessary to update the current attribute information with the new target information to be processed to ensure the real-time and accuracy of the updated target attribute information.

[0043] The target attribute information is the attribute information of the target object after the update processing. Exemplarily, after the information recommendation device extracts the initial attribute information of the target object from the acquired target information to be processed, the current attribute information of the target object is updated according to the initial attribute information to obtain the target attribute information. Specifically, the update processing includes deduplication and merging. Deduplication is based on the same attributes, and the attributes of the initial attribute information are overwritten with the corresponding attributes of the current attribute information; merging is to add the newly added attributes in the initial attribute information to the current attribute information to obtain the updated target attribute information. In some embodiments, the target information to be processed can be periodically obtained to update the current attribute information of the target object to obtain the target attribute information of the target object. The regular time can be weekly, monthly, etc.

[0044] S130: Determine target recommendation information according to the action attribute in the target attribute information, and send the target recommendation information to the target object end.

[0045] The action attribute may include the action that the target object has performed or wants to perform. Exemplarily, the action attribute of the target object can be extracted from the target attribute information. The action attribute includes the action intention of the target object and the corresponding intention intensity, which reflects the current behavior state of the target object and provides guidance for information recommendation.

[0046] After the information recommendation device determines the corresponding target recommendation information according to the action attribute of the target object, the target recommendation information is actively sent to the target object end and displayed on the target object end. The target object can receive the target recommendation information through the interactive interface.

[0047] It can be seen that the information recommendation method of the embodiment of the present application extracts the initial attribute information associated with the target object from the acquired target information to be processed; updates the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; determines the target recommendation information according to the action attribute in the target attribute information, and sends the target recommendation information to the target object end. The above scheme first updates the current attribute information of the target object according to the initial attribute information, which can improve the timeliness of the target object information update and ensure the accuracy of the target attribute information; then recommends information based on the action attribute in the updated target attribute information, so that the recommended information is more in line with the intention of the target object, thereby improving the accuracy of the information recommendation.

[0048] Based on the above embodiments, the embodiments of the present application adopt Figure 2 The flowchart details how to extract the initial attribute information of the target object from the target information to be processed. Figure 2 , Figure 2 yes Figure 1 The flowchart of an exemplary embodiment of step S110 in the information recommendation method is shown. Specifically, the process of extracting the initial attribute information associated with the target object from the acquired target information to be processed in step S110 may include the following steps:

[0049] S210: Input the target information to be processed and the first preset prompt information into the large model to obtain the attribute information associated with the target object output by the large model.

[0050] The first preset prompt information is used to instruct the large model to determine the attribute information associated with the target object according to the target information to be processed. Exemplarily, the first preset prompt information may include the attribute extraction rules of the target object, the attribute extraction template, and the task description. When extracting the user's attribute information, the user's attribute information is extracted according to the User portrait prompt.

[0051] The large model can be a large language model (LLM), which is a large-scale artificial intelligence model used to understand and answer questions raised by users. The large model may include but is not limited to: chat GPT, Tongyi Qianwen, etc. The target information to be processed and the first preset prompt information are input into the large model, and the large model understands and analyzes the target information to be processed according to the first preset prompt information to obtain attribute information associated with the target object.

[0052] S220: Determine the attribute information associated with the target object output by the large model as initial attribute information.

[0053] The information recommendation device determines the attribute information associated with the target object output by the large model as the initial attribute information of the target object. In other embodiments, the information recommendation device can also extract the initial attribute information associated with the target object from the target information to be processed by means of keyword extraction or pre-trained neural network model.

[0054] After the information recommendation device determines the initial attribute information of the target object according to the acquired target information to be processed, it updates the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object. In some embodiments, the information recommendation device can update the current attribute information through a large model. Specifically, the initial attribute information, the current attribute information, and the second preset prompt information are input into the large model to obtain the updated attribute information output by the large model; the updated attribute information output by the large model is determined as the target attribute information of the target object. Compared with the traditional scheme that relies on manually set rules or simple data analysis, the above scheme is difficult to process a large amount of unstructured data, and cannot reflect the changes in information in real time, resulting in low comprehensiveness and accuracy of the attribute information. The embodiment of the present application determines the initial attribute information and the target attribute information through a large model. The large model has the ability to understand and can quickly integrate the provided target information to be processed, and efficiently and accurately construct and update the attribute information of the target object.

[0055] Among them, the second preset prompt information is used to prompt the big model to update the current attribute information in combination with the initial attribute information to obtain the target attribute information. Exemplarily, the second preset prompt information may include attribute update rules, attribute update templates, and task descriptions. When updating the user's current attribute information, the user's current attribute information is updated according to the User portrait update prompt. For example: "Please rebuild the target attribute information of the target object p2 based on the current attribute information and initial attribute information of the target object p2." In addition, the big model can be the same big model as the initial attribute information, or it can be a different big model. When it is the same big model, the big model can extract contextual information through CoT (Chain of Thought) reasoning to improve the ability of the big model to handle complex problems.

[0056] In other embodiments, the information recommendation device may also compare the initial attribute information with the current attribute information to determine the incremental information and the change information of the initial attribute information relative to the current attribute information; update the current attribute information according to the incremental information and the change information to obtain the target attribute information.

[0057] In order to elaborate on the process of updating attribute information in this application, Figure 3The framework diagram shown further illustrates it, as follows:

[0058] First of all, it should be noted that the target information to be processed in this example is dialogue information, and the update process is a dynamic update process, which can effectively avoid problems such as not updating attribute information, discarding determined features, and not adding new features.

[0059] Firstly, the dialogue information 0 and the first preset prompt information are input into the large model to obtain initial attribute information, and the initial attribute information is determined as the target attribute information;

[0060] After receiving the dialogue information 1, the target attribute information determined according to the dialogue information 0 is used as the current attribute information; the first part is to construct the initial attribute information of the dialogue information, and the dialogue information 1 and the first preset prompt information are input into the large model to obtain the initial attribute information; the second part is to update the current attribute information, and the initial attribute information, the current attribute information and the second preset prompt information are input into the large model to update the current attribute information, and the updated attribute information output by the large model is obtained, and the target attribute information is determined by the updated attribute information output by the large model;

[0061] After receiving dialogue information n, the target attribute information updated according to dialogue information n-1 is used as the current attribute information; dialogue information n and the first preset prompt information are input into the large model to obtain the initial attribute information; the initial attribute information, the current attribute information and the second preset prompt information are input into the large model to update the current attribute information to obtain the updated attribute information output by the large model, and the updated attribute information output by the large model is used to determine the target attribute information. In this way, dynamic update is achieved, and when the next dialogue information is input, the above steps are repeated to update the attribute information of the target object. In the dialogue flow, dialogue information 0, dialogue information 1 and dialogue information n are arranged in the order of time from early to late.

[0062] In the above dynamic update process, in order to make the initial attribute information extracted by the large model more accurate, the large model can be used to evaluate and optimize the first preset prompt information and the generated initial attribute information, and update the first preset prompt information. Specifically, the initial attribute information, the first preset prompt information, and the preset evaluation prompt information are input into the large model for evaluation and optimization processing to obtain the optimization information of the first preset prompt information; the first preset prompt information is optimized according to the optimization information of the first preset prompt information to obtain the optimized first preset prompt information, and the optimized first preset prompt information is used by the large model to extract attribute information for the next target information to be processed. In this way, the first preset prompt information is optimized to improve the accuracy of the target attribute information, thereby improving the accuracy of the recommended information.

[0063] The preset evaluation prompt information is used to instruct the large model to evaluate the initial attribute information and the first preset prompt information and propose corresponding optimization information. Exemplarily, the preset evaluation prompt information may include a task description, an evaluation template, etc. The information recommendation device scores the initial attribute information, including a comprehensive evaluation of the completeness, accuracy, and recall rate of the initial attribute information; determines the score of the first preset prompt information based on the initial attribute information, and evaluates whether the first preset prompt information can enable the large model to output more accurate and comprehensive initial attribute information.

[0064] For details, please refer to Figure 4 , input the initial attribute information, the first preset prompt information and the preset evaluation prompt information into the large model, obtain the score and optimization information output by the large model, optimize the first preset prompt information according to the optimization information, and obtain the first preset prompt information after optimization. The large model uses the first preset prompt information after optimization to extract attribute information from the next target information to be processed.

[0065] In other embodiments, the information recommendation device may also periodically evaluate and optimize the first preset prompt information, and the evaluation and optimization may be performed not only by manual evaluation and optimization, but also by a neural network model. Of course, the above evaluation and optimization process may also evaluate and optimize the second preset prompt information and the preset evaluation prompt information.

[0066] After obtaining the target attribute information of the target object, the target recommendation information can be determined according to the action attribute in the target attribute information to improve the user experience. The step of determining the target recommendation information includes: determining the action attribute of the target object from the target attribute information; determining the target recommendation information according to the intensity of the intention of each action intention in the action attribute of the target object, and sending the target recommendation information to the target object end. Thus, determining the target recommendation information by the action attribute can improve the accuracy of information recommendation, and realize active recommendation according to the intensity of the intention of the action intention.

[0067] The action attributes include the action intention of the target object and the corresponding intention intensity. Action intention refers to the action that the target object wants to perform in the future. Intention intensity refers to the determination and intensity of the target object to perform the corresponding action. As an example, the intensity of intention is divided into 5 levels with a full score of 10:

[0068] 1-2 points (average): The target information to be processed indicates that the action intention is not obvious or is rarely mentioned;

[0069] 3-4 points (normal): The intention of the action is mentioned, but there is no strong motivation or need;

[0070] 5-6 points (positive): The target information to be processed shows a certain degree of enthusiasm and willingness;

[0071] 7-8 points (strong): The action intention is clear, and there is a clear motivation or need in the dialogue;

[0072] 9-10 points (very strong): The action intention is passionate and shows very strong intention and determination.

[0073] The information recommendation device determines whether to make an active recommendation according to the intensity of the action intention, and judges whether the intensity of the action intention in the action attribute of the target object is greater than a preset intensity threshold; if so, the target recommendation information is determined according to the corresponding action intention, and the target recommendation information is sent to the target object end. The preset intensity threshold can be preset manually, and can be set to 6 points, 7 points or 8 points, etc.

[0074] In other embodiments, the information recommendation device may further extract action attributes from the target attribute information and determine the target recommendation information respectively according to each action intention in the action attributes.

[0075] As an example, if the target information to be processed is a dialogue message, the dialogue message is as follows:

[0076] p1: Do you have any plans for tomorrow? If you have nothing to do, come and go hiking with us. The weather will be very good tomorrow.

[0077] p2: No, I am going to play basketball with my friends in Binjiang tomorrow. I haven't been there for a long time. I will watch a documentary with my children in the evening. The teacher has assigned homework.

[0078] The action attributes of the target object p2 are extracted from the target information to be processed:

[0079] Action intention: social activity (preparing to play basketball with friends); Intention intensity: 8 points (clear arrangements and clear expectations);

[0080] Action intention: family activity (planning to watch a documentary with the children tomorrow night); Intention intensity: 7 points (clear family activity plan).

[0081] The target recommendation information can be personalized recommendation information provided to the target object based on its action attributes. Exemplarily, the action attributes of the target object can be combined, and the large model can be applied to concretize the search question according to the "Reified search prompt" to obtain accurate recommendation information. For example: "Please analyze the action intention of the target object in combination with the target attribute information, sort the action intention according to the intensity of the intention, and generate the corresponding specific search question, such as ["Recommend a women's shoulder bag with a price of about 20,000 and a moderate capacity"]; and make an accurate and concise recommendation for the target object based on the specific search question."

[0082] For the above example, the active recommendation results for the target object p2 are as follows:

[0083] 1. Reservation of basketball courts near City A: It is recommended to use the "a" application software to reserve "City A Sports Park Basketball Court", or visit the official website of "City A Sports Center" to check the available time slots;

[0084] 2. Recommended list of popular science documentaries suitable for children:

[0085] - "A" (exploring the mysteries of the universe);

[0086] - "B" (physical geography knowledge);

[0087] - “C” (Human culture and living environment around the world).

[0088] In addition, in order to cope with different needs, the target information to be processed can be information of different modes. Specifically, the information recommendation device obtains initial information to be processed, which includes information to be processed of the target object and information to be processed of other objects; in response to receiving the first mode instruction, the initial information to be processed is determined as the target information to be processed, and the initial attribute information associated with the target object is extracted from the target information to be processed; in response to receiving the second mode instruction, the information to be processed of the target object is extracted from the initial information to be processed, and the information to be processed of the target object is determined as the target information to be processed, and the initial attribute information associated with the target object is extracted from the target information to be processed. In this way, the information to be processed is determined by mode selection to adapt to different scenarios and needs.

[0089] The initial information to be processed may be unprocessed information, or it may be information that has been grammatically checked, such as wrong spelling detection, multi-word detection, and word order detection. The initial information to be processed may include the information to be processed of the target object and the information to be processed of other objects. As an example, when the information to be processed is dialogue information, the initial information to be processed is the interactive dialogue information between the target object and other objects. For example:

[0090] p1: What are we going to eat tonight?

[0091] p2: Let's go eat something light. I'm feeling a little irritated.

[0092] p1: Let’s eat Cantonese food, but it’s raining outside now.

[0093] p2: We can order takeout.

[0094] If the first mode instruction is received, the initial information to be processed is determined as the target information to be processed. The first mode instruction can also be called non-specified object input, and the target information to be processed can include not only the information to be processed of the target object, but also the information to be processed of other objects other than the target object. The target information to be processed thus determined has rich dialogue logic, from which the initial attribute information of the specified target object is extracted, and is not affected by the number of objects.

[0095] If the second mode instruction is received, the target object's to-be-processed information is extracted from the initial to-be-processed information, that is, information irrelevant to the target object is excluded. The second mode can also be called a specified object input, and the target to-be-processed information only includes the target object's to-be-processed information. As an example, when the second mode instruction is received, the above initial to-be-processed information is processed to obtain the target to-be-processed information as follows:

[0096] p1: What shall we eat tonight? Let's have Cantonese food, but it's raining outside now.

[0097] p2: Let's have Cantonese food, but it's raining outside now. We can order takeout.

[0098] Therefore, by integrating the information to be processed of the target object, the confusion of feature information between different objects can be avoided in complex scenarios.

[0099] In order to elaborate on the information recommendation method of the present application, Figure 5 The framework diagram shown further illustrates it, as follows:

[0100] First, input the initial information to be processed;

[0101] The first part is a mode selection module, which includes a first mode and a second mode. If the first mode instruction is received, the initial information to be processed is determined as the target information to be processed; if the second mode instruction is received, the information to be processed of the target object is extracted from the initial information to be processed to obtain the target information to be processed;

[0102] The second part is an attribute information extraction and update module, which inputs the target information to be processed and the first preset prompt information into the large model to obtain the initial attribute information output by the large model; the initial attribute information, the current attribute information of the target object and the second preset prompt information are input into the large model for combined update processing to obtain the target attribute information of the target object;

[0103] The third part is the active recommendation module, which extracts the action attributes of the target object from the target attribute information, where the action attributes include action intention and intention intensity; determines whether the intention intensity of each action intention is greater than the preset intensity threshold. If so, it triggers active recommendation, combines the corresponding action intention, applies the big model to determine the target recommendation information according to the specific search prompts, and sends the target recommendation information to the target object end to realize active recommendation of information.

[0104] See also Figure 6 , Figure 6 6 is a schematic diagram of an exemplary embodiment of an information recommendation device shown in the present application. The information recommendation device 600 includes an extraction module 610, an update module 620 and a recommendation module 630. The extraction module 610 is used to extract the initial attribute information associated with the target object from the acquired target information to be processed; the update module 620 is used to update the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; the recommendation module 630 is used to determine the target recommendation information according to the action attribute in the target attribute information, and send the target recommendation information to the target object end.

[0105] In the above scheme, the information recommendation device extracts the initial attribute information associated with the target object from the acquired target information to be processed; updates the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; determines the target recommendation information according to the action attribute in the target attribute information, and sends the target recommendation information to the target object end. In the above scheme, the current attribute information of the target object is first updated according to the initial attribute information, which can improve the timeliness of the target object information update and ensure the accuracy of the target attribute information; then, the action attribute in the updated target attribute information is used to recommend information, so that the recommended information is more in line with the intention of the target object, and the accuracy of the information recommendation is improved.

[0106] The functions of each module can be found in the information recommendation method embodiment, which will not be described in detail here.

[0107] In order to implement the information recommendation method of the above embodiment, the present application proposes another electronic device, for details, please refer to Figure 7 , Figure 7 It is a structural schematic diagram of an embodiment of an electronic device provided by the present application.

[0108] The electronic device 700 includes a memory 710 and a processor 720 , wherein the memory 710 and the processor 720 are coupled.

[0109] The memory 710 is used to store program data, and the processor 720 is used to execute the program data to implement the information recommendation method of the above embodiment.

[0110] In this embodiment, the processor 720 may also be referred to as a CPU (Central Processing Unit). The processor 720 may be an integrated circuit chip having signal processing capabilities. The processor 720 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. A general-purpose processor may be a microprocessor or the processor 720 may also be any conventional processor, etc.

[0111] The present application also provides a computer-readable storage medium, such as Figure 8 As shown, the computer-readable storage medium 800 is used to store program data 810. When the program data 810 is executed by the processor, it is used to implement the information recommendation method in the method embodiment of the present application.

[0112] The method involved in the information recommendation method embodiment of the present application, when implemented in the form of a software functional unit and sold or used as an independent product, can be stored in a device, such as a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or part of the contribution to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0113] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An information recommendation method, characterized in that: The information recommendation method comprises: Extracting initial attribute information associated with the target object from the acquired target information to be processed; updating the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object; Target recommendation information is determined according to the action attribute in the target attribute information, and the target recommendation information is sent to the target object end.

2. The information recommendation method according to claim 1, characterized in that: The step of extracting initial attribute information associated with the target object from the acquired target information to be processed includes: Inputting the target information to be processed and the first preset prompt information into the large model to obtain the attribute information associated with the target object output by the large model; The attribute information associated with the target object output by the large model is determined as initial attribute information.

3. The information recommendation method according to claim 1, characterized in that: The step of updating the current attribute information of the target object according to the initial attribute information to obtain the target attribute information of the target object includes: Inputting the initial attribute information, the current attribute information and the second preset prompt information into the large model to obtain updated attribute information output by the large model; The updated attribute information output by the large model is determined as the target attribute information of the target object.

4. The information recommendation method according to claim 2, characterized in that: The method further comprises: Inputting the initial attribute information, the first preset prompt information and the preset evaluation prompt information into the large model for evaluation optimization processing to obtain optimization information of the first preset prompt information; The first preset prompt information is optimized according to the optimization information of the first preset prompt information to obtain the optimized first preset prompt information, and the optimized first preset prompt information is used by the large model to extract attribute information of the next target information to be processed.

5. The information recommendation method according to claim 1, characterized in that: The action attribute includes at least one action intention and a corresponding intention intensity, and the step of determining target recommendation information according to the action attribute in the target attribute information and sending the target recommendation information to the target object end includes: Determining an action attribute of the target object from the target attribute information; The target recommendation information is determined according to the intention intensity of each action intention in the action attribute of the target object, and the target recommendation information is sent to the target object end.

6. The information recommendation method according to claim 5, characterized in that: The step of determining the target recommendation information according to the intention intensity of each action intention in the action attribute of the target object, and sending the target recommendation information to the target object end includes: Determine whether the intensity of each action intention in the action attribute of the target object is greater than a preset intensity threshold; If so, the target recommendation information is determined according to the corresponding action intention, and the target recommendation information is sent to the target object end.

7. The information recommendation method according to claim 1, characterized in that: The step of extracting initial attribute information associated with the target object from the acquired target information to be processed includes: Acquire initial information to be processed, wherein the initial information to be processed includes information to be processed of the target object and information to be processed of other objects; In response to receiving the first mode instruction, determining the initial information to be processed as the target information to be processed, and extracting initial attribute information associated with the target object from the target information to be processed; In response to receiving the second mode instruction, the to-be-processed information of the target object is extracted from the initial to-be-processed information, and the to-be-processed information of the target object is determined as the target to-be-processed information, and the initial attribute information associated with the target object is extracted from the target to-be-processed information.

8. An information recommendation device, characterized in that: The device comprises: An extraction module, used to extract initial attribute information associated with the target object from the acquired target information to be processed; An updating module, configured to update the current attribute information of the target object according to the initial attribute information to obtain target attribute information of the target object; The recommendation module is used to determine target recommendation information according to the action attribute in the target attribute information, and send the target recommendation information to the target object end.

9. An electronic device, characterized in that: include: A memory and a processor, wherein the memory stores program instructions, and the processor retrieves the program instructions from the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: include: Program data is stored, and when the program data is executed by a processor, it is used to implement the method according to any one of claims 1 to 7.

Images