Information recommendation method and device based on large model, intelligent agent and electronic equipment

Abstract

The invention provides an information recommendation method and device based on a large model, an intelligent agent and electronic equipment, and relates to the technical field of artificial intelligence, in particular to the technical field of large models, AI assistants and the like. The specific implementation scheme of the information recommendation method comprises the steps of receiving historical experience information of a second object input by a first object and position information of the second object; analyzing the historical experience information by using a large model to generate capability feature information of the second object; and determining a first target mechanism for recommending to the first object from the plurality of candidate mechanisms by analyzing a plurality of paths of the second object to the plurality of candidate mechanisms based on the position information and the position information of the plurality of candidate mechanisms by using the large model.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to the fields of large models, AI assistants, etc., and specifically relates to an information recommendation method, device, intelligent agent, and electronic device based on a large model. Background Art

[0002] With the increasing abundance of information resources, a variety of resources provide users with multiple choices. In order to improve the user experience, there is an urgent need to provide an information recommendation method that can meet the diverse needs of users. Summary of the Invention

[0003] The present disclosure provides an information recommendation method, device, intelligent agent, and electronic device based on a large model.

[0004] According to one aspect of the present disclosure, there is provided an information recommendation method based on a large model, including: receiving historical experience information of a second object and location information of the second object input by a first object; analyzing the historical experience information by using a large model to generate ability characteristic information of the second object; and using the large model, based on the location information and location information of multiple candidate institutions, by analyzing multiple paths for the second object to go to the multiple candidate institutions, determining a first target institution for recommending to the first object from the multiple candidate institutions.

[0005] According to another aspect of the present disclosure, there is provided an information recommendation device based on a large model, including a receiving module, an analyzing module, and a determining module.

[0006] The receiving module is configured to receive historical experience information of a second object and location information of the second object input by a first object.

[0007] The analyzing module is configured to analyze the historical experience information by using a large model to generate ability characteristic information of the second object.

[0008] The determining module is configured to use the large model, based on the location information and location information of multiple candidate institutions, by analyzing multiple paths for the second object to go to the multiple candidate institutions, determining a first target institution for recommending to the first object from the multiple candidate institutions.

[0009] According to another aspect of the present disclosure, there is provided an intelligent agent for information recommendation, including: an input module, a processing module, and an output module.

[0010] The input module is configured to receive historical experience information of a second object and location information of the second object input by a first object.

[0011] A processing module, configured to determine a target task based on historical experience information and location information received by an input module, determine a target large model based on the target task, and execute the method described above by invoking the target large model to obtain information of a first target institution.

[0012] An output module, configured to output the information of the first target institution obtained by the processing module.

[0013] According to another aspect of the present disclosure, there is provided an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described above.

[0014] According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the method described above.

[0015] According to another aspect of the present disclosure, there is provided a computer program product, including a computer program, where the computer program, when executed by a processor, implements the method described above.

[0016] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. Description of the Drawings

[0017] The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. Among them:

[0018] Figure 1 Schematically shows an exemplary system architecture to which the information recommendation method and device based on a large model according to an embodiment of the present disclosure can be applied;

[0019] Figure 2 Schematically shows a flowchart of the information recommendation method based on a large model according to an embodiment of the present disclosure;

[0020] Figure 3A Schematically shows a schematic diagram of determining candidate institutions based on historical experience information using a large model according to an embodiment of the present disclosure;

[0021] Figure 3B Schematically shows a schematic diagram of determining candidate institutions based on historical experience information using a large model according to another embodiment of the present disclosure;

[0022] Figure 4Schematically shows a schematic diagram of using a large model for path generation and analysis according to an embodiment of the present disclosure;

[0023] Figure 5 Schematically shows a schematic diagram of generating a path in combination with travel preferences according to an embodiment of the present disclosure;

[0024] Figure 6 Schematically shows a schematic diagram of using a large model for path safety analysis according to an embodiment of the present disclosure;

[0025] Figure 7 Schematically shows a schematic diagram of using a large model for credibility analysis according to an embodiment of the present disclosure;

[0026] Figure 8 Schematically shows a block diagram of an information recommendation device based on a large model according to an embodiment of the present disclosure;

[0027] Figure 9 Schematically shows a block diagram of an agent for information recommendation according to an embodiment of the present disclosure; and

[0028] Figure 10 Schematically shows a block diagram of an electronic device suitable for implementing an information recommendation method based on a large model according to an embodiment of the present disclosure. Detailed implementation manners

[0029] The following describes exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted below.

[0030] In related examples, when making institutional recommendations, it is usually necessary to obtain information about various aspects of the institution, such as promotional information and evaluation information, etc. Then, based on these fragmented information, recommendations are made based on keywords or historical preferences input by the user. Due to the lack of in-depth understanding of the user's needs or intentions, the matching degree between the recommended results and the user's intentions is low, reducing the user experience.

[0031] In view of this, the embodiments of the present disclosure utilize a large model to mine the potential capabilities of users based on their historical experience information, and deeply understand user intentions from the perspective of stimulating the potential capabilities of the training objects. Then, considering the convenience and safety of going to the institution, the large model is used to analyze the paths of users going to each candidate institution to determine the first target institution for stimulating the potential capabilities of users, further strengthening the understanding of user intentions from the perspective of user travel, thereby improving the matching degree between the recommendation result and user intentions and enhancing the user experience.

[0032] Figure 1 FIG. schematically shows an exemplary system architecture to which the information recommendation method and apparatus based on a large model according to the embodiments of the present disclosure can be applied.

[0033] It should be noted that Figure 1 The illustration is only an example of the system architecture to which the embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios. For example, in another embodiment, the exemplary system architecture to which the information recommendation method and apparatus based on a large model can be applied may include a terminal device, but the terminal device can implement the information recommendation method and apparatus based on a large model provided by the embodiments of the present disclosure without interacting with the server.

[0034] As Figure 1 shown, the exemplary architecture 100 may include a terminal device 101, an agent 102, and a server 103.

[0035] Various communication client applications may be installed on the terminal device 101, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, AI intelligent assistants, etc. (only as examples).

[0036] The terminal device 101 may be various electronic devices with a display screen and supporting web browsing, including but not limited to smartphones, tablets, laptop portable computers, desktop computers, and the like.

[0037] The agent 102 may be based on a large model, such as a large language model, and output information that meets user needs by identifying user needs.

[0038] The server 103 may be a server providing various services, such as a background management server that supports the content browsed by the user using the terminal device 101 (only as an example). The background management server may analyze and process data such as user requests received, and feedback the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device 101.

[0039] For example, the user can input the historical experience information of the second object and the location information of the second object in the terminal device 101. The terminal device 101 can perform operations such as ability feature mining, path generation, and path analysis by calling the agent 102 multiple times, and finally output the target institution information 110.

[0040] It should be noted that the information recommendation method based on the large model provided by the embodiments of the present disclosure can generally be executed by the terminal device 101. Correspondingly, the information recommendation device based on the large model provided by the embodiments of the present disclosure can also be set in the terminal device 101.

[0041] Alternatively, the information recommendation method based on the large model provided by the embodiments of the present disclosure can generally also be executed by the server 103. Correspondingly, the information recommendation device based on the large model provided by the embodiments of the present disclosure can generally be set in the server 103. The information recommendation method based on the large model provided by the embodiments of the present disclosure can also be executed by a server or a server cluster different from the server 103 and capable of communicating with the terminal device 101 and / or the server 103. Correspondingly, the information recommendation device based on the large model provided by the embodiments of the present disclosure can also be set in a server or a server cluster different from the server 103 and capable of communicating with the terminal device 101 and / or the server 103.

[0042] For example: The terminal device 101 can send the historical experience information of the second object and the location information of the second object input by the user to the server 103. After receiving the historical experience information of the second object and the location information of the second object, the server 103 can generate the target institution information by finally calling the agent 102 by executing the information recommendation method based on the large model of the embodiments of the present disclosure. Finally, the target institution information is fed back to the terminal device 101.

[0043] It should be understood that Figure 1 the numbers of the terminal devices, agents, and servers in

[0044] In the technical solution of the present disclosure, the processing of the collection, storage, use, processing, transmission, provision, disclosure, and application of the user's personal information involved all comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good customs.

[0045] In the technical solution of the present disclosure, before obtaining or collecting the user's personal information, the authorization or consent of the user is obtained.

[0046] Figure 2 Schematically shows a flowchart of the information recommendation method based on the large model according to the embodiments of the present disclosure.

[0047] As Figure 2 shown, the method includes operations S210 to S230.

[0048] In operation S210, historical experience information of a second object and location information of the second object input by a first object are received.

[0049] In operation S220, the large model is used to analyze the historical experience information to generate ability characteristic information of the second object.

[0050] In operation S230, the large model is used to determine, based on the location information and location information of multiple candidate institutions, a first target institution for recommendation to the first object by analyzing multiple paths for the second object to go to the multiple candidate institutions.

[0051] According to an embodiment of the present disclosure, the second object may be a to-be-trained object, and there may be a kinship between the first object and the second object. For example: the first object may be a guardian of the second object, etc. The historical experience information may be unimodal information or multimodal information. For example: at least two of text, image, audio, and video. The text may include text for describing the behavior preferences of the second object. For example: likes singing, dancing, etc. The image may include an image of the award list of various competitions participated by the second object. The audio and video may include audio and video clips for recording the growth experience of the second object.

[0052] In some embodiments, a Prompt (hint) may be constructed based on the historical experience information and input into the large model to output the ability characteristic information. The large model may be a large language model or a multimodal large model.

[0053] According to an embodiment of the present disclosure, the ability characteristic information may include scores of the second object in each predetermined dimension. The predetermined dimension may, for example, include: artistic performance, language expression, social ability, logical reasoning ability, observation ability, etc.

[0054] For example: the historical experience information of the second object may include "likes singing, dancing, also likes storytelling, has a somewhat introverted personality but is good at observing". By analyzing the historical experience information using the large model, the obtained ability characteristic information may be "[Artistic Performance] 0.85; [Language Expression] 0.80; [Social Ability] 0.45; [Observation Ability] 0.75".

[0055] It can be understood that the behavior preferences of the second object in the above historical experience information are directly described in text form, and such behavior preferences are usually obtained based on the subjective cognition of the first object about the second object.

[0056] In order to further improve the accuracy of mining the ability characteristics of the second object, in the embodiments of the present disclosure, a large model can also be used to analyze multi-modal historical experience information. For example, videos, texts, etc. that record the growth experience of the second object are input into the multi-modal large model to generate ability characteristic information.

[0057] According to the embodiments of the present disclosure, a candidate institution is an institution configured with courses that match the ability characteristic information. For example, if the scores of the art performance dimension and the language expression dimension in the ability characteristic information of the second object are relatively high, it can be determined that the institutions configured with art courses and language expression courses are candidate structures.

[0058] In the embodiments of the present disclosure, the convenience and safety dimensions of the commuting path of the second object to the institution can be analyzed.

[0059] For example, if the path length of the second object to institution In1 is longer than the path length of the second object to institution In2, it can be determined that the convenience degree of the path to institution In1 is lower.

[0060] For example, if the path length of the second object to institution In1 is the same as the path length of the second object to institution In2, and there is a pedestrian overpass configured in the path to institution In1, and there are various roadblocks configured in the path to institution In2 due to road construction, it can be determined that the safety degree of the path to institution In1 is higher.

[0061] In some embodiments, the commuting path can also be analyzed from the dimension of the travel preference of the second object.

[0062] For example, the travel preference of the second object is to take public transportation. There is a public transportation vehicle directly from the object's location to institution In1 configured in the path of the second object to institution In1. In the path of the second object to institution In2, after reaching the public transportation station from the object's location, it is necessary to walk a certain distance to reach institution In2. It can be determined that the convenience degree of the path to institution In1 is higher.

[0063] Therefore, since the traffic condition data is updated dynamically, the large model can be used to dynamically analyze multiple paths of multiple candidate institutions based on the updated traffic condition data, so as to select the institution corresponding to the path that is convenient and safe for travel as the first target institution.

[0064] In the embodiments of the present disclosure, based on the historical experience information of the user, the large model is used to mine the potential capabilities of the user, and the user intention is deeply understood from the perspective of stimulating the potential capabilities of the training object. Then, considering the convenience and safety of going to the institution, the large model analyzes the paths of the user going to each candidate institution to determine the first target institution for stimulating the potential capabilities of the user, further strengthening the understanding of the user intention from the perspective of the user's travel, thereby improving the matching degree between the recommendation result and the user intention and enhancing the user experience.

[0065] The following refers to Figure 3A - Figure 3B and further describes the Figure 2 method shown in conjunction with specific embodiments.

[0066] According to the embodiments of the present disclosure, by using the large model to analyze the historical experience information and generate the ability characteristic information of the second object, the following operations may be included: obtaining the behavior preference information from the historical experience information; and using the large model to analyze the behavior preference information to generate the ability characteristic information.

[0067] Figure 3A Schematically shows a schematic diagram of determining candidate institutions based on historical experience information using a large model according to an embodiment of the present disclosure

[0068] As Figure 3A shown, in this embodiment 300A, first, the behavior preference information 322 "singing, dancing, storytelling, observing" is obtained from the historical experience information 321 "Object A is 5 years old and won the first prize in the children's group dance at the art festival. Participated in a storytelling competition in xx year...".

[0069] Then, the Prompt constructed based on the behavior preference information 322 is input into the large model 320, and the ability characteristic information 323 is output.

[0070] In some embodiments, in order to improve the matching degree between the ability mining direction and the user intention, evaluation dimension information may be added to the Prompt. The evaluation dimension information may be pre-configured or obtained based on the user's selection on the interaction interface.

[0071] Next, the candidate institution 324A may be determined based on the scores of each dimension in the ability characteristic information 323.

[0072] For example: if the scores of artistic performance and language expression in the ability characteristic information 323 are relatively high, an institution configured with at least one of art courses and language courses may be determined as a candidate institution.

[0073] In an embodiment of the present disclosure, by using a large model to extract behavioral preferences based on historical experience information, the accuracy of potential ability mining is further improved, thereby reducing the impact of the subjective cognition of the object on the accuracy of information recommendation.

[0074] To further improve the matching degree between the information recommendation result and the user's intention, it is also possible to receive the intention information selected by the first object; use the large model to analyze the intention information and the ability feature information to generate a course type that matches the ability feature information; and determine multiple candidate institutions based on the course type.

[0075] Figure 3B A schematic diagram showing the determination of candidate institutions using a large model based on historical experience information according to another embodiment of the present disclosure is schematically illustrated.

[0076] As Figure 3B shown, the difference between this embodiment 300B and embodiment 300A is that an interaction interface 301 is added. In the interaction interface 301, the course type can be displayed to the first object, for example: art performance type 3011, logical reasoning type 3012, physical fitness improvement type 3013, etc.

[0077] When the first object selects the logical reasoning type 3012, it means that the first object wants to cultivate the second object's ability in logical reasoning, and the intention information 325 can be determined to stimulate the second object's logical reasoning ability.

[0078] Then, the intention information 325 and the ability feature information 323 can be input into the large model 320. At this time, the large model can generate a course type that matches the ability feature by comprehensively analyzing the user's intention tendency and the ability feature of the second object.

[0079] It can be understood that in embodiment 300B, since the user's intention tendency lies in the logical reasoning type of courses, therefore, even if the score of the "observation" dimension in the ability feature information 323, which is "0.75", is lower than the scores of the "language expression" dimension, which is "0.80", and the "art performance" dimension, which is "0.85", the course type 326 output by the large model mainly focuses on the "observation" dimension, for example: thinking type 3261 and programming type 3262.

[0080] In some embodiments, the intention information selected by the user on the interaction interface may not match the ability feature information. For example: the intention information can be "physical fitness improvement type", however, the score of the [limb coordination] dimension in the ability feature information is lower than the predetermined threshold, indicating that the second object lacks the basic ability to take physical fitness improvement type courses. At this time, when the large model generates the course type, it will generate the course type based on the dimension in the ability feature information whose score is greater than the predetermined threshold.

[0081] Finally, according to the course type 326, the candidate institution 324B is determined. The institutions in the candidate institution 324B are at least configured with one course in the course type 326.

[0082] In the embodiments of the present disclosure, in addition to considering the ability characteristics, intention information is added, further improving the matching degree between the result of information recommendation and the user's intention.

[0083] The following combines Figure 4 - Figure 6 , in combination with specific embodiments, further illustrates Figure 2 the method shown.

[0084] Using a large model, based on the location information and the location information of multiple candidate institutions, by analyzing multiple paths for the second object to go to multiple candidate institutions, to determine the first target institution for recommending to the first object from multiple candidate institutions, may include the following operations: using a large model, based on the location information and the location information of multiple candidate institutions, generating multiple paths; and using a large model, by analyzing the multiple paths, determining the first target institution from multiple candidate institutions.

[0085] Figure 4 Schematically shows a schematic diagram of using a large model for path generation and analysis according to an embodiment of the present disclosure.

[0086] As Figure 4 shown, in operation S230, first, based on the candidate institution 324B and the object location 331, a large model can be called to generate paths 332A between the object location T and each candidate institution.

[0087] In some embodiments, there can be multiple paths from the object location T to each candidate institution. For example: the path from the object location T to the candidate institution In1 can include path Pa and path Pb.

[0088] To reduce the data processing volume when the large model analyzes the paths, one path can also be generated for each candidate institution when generating the paths. For example: in the path from the object location to the candidate institution In1, when the length of path Pa is less than the length of path Pb, only path Pa can be generated.

[0089] Then, call the large model to analyze each path, and the analysis dimensions include but are not limited to convenience and safety. The comprehensive score of each analysis dimension can be used as the comprehensive score of the path, and the institution corresponding to the path with the highest comprehensive score is determined as the first target institution. For example: institution Inj 333.

[0090] In the embodiments of the present disclosure, by using a large model to analyze each path to a candidate institution, the target institution is determined from the comprehensive dimensions of travel convenience and travel safety, further improving the matching degree of the recommendation result with the user's commuting path requirements and further enhancing the user experience.

[0091] In some embodiments, when generating the path of the second object to the candidate institution, the travel preferences of the user can also be considered. For example: some users like to travel by public transportation, then paths with bus stops are preferentially generated. Some users like to travel on foot, then paths with shorter lengths or overpasses are preferentially generated.

[0092] Therefore, by using a large model, based on the location information and the location information of multiple candidate institutions, generating multiple paths may include the following operations: receiving the travel preference information selected by the first object; and using the large model to generate multiple paths based on the travel preference information, location information, and the location information of multiple candidate institutions.

[0093] Figure 5 A schematic diagram showing the generation of paths in combination with travel preferences according to an embodiment of the present disclosure is schematically illustrated.

[0094] As Figure 5 shown, first, the first object can select the travel preference 334 from public transportation 3021, private car 3022,..., walking 302n through the interaction interface 302. For example: it can be the private car 3022.

[0095] Then, the large model can be called to generate multiple paths 332B based on the locations of the candidate institutions in the candidate institution 324B and the object location 331, in combination with the travel preference 334.

[0096] It can be understood that when the large model generates the path of the second object to the candidate institution in combination with the travel preference information, it can generate the path by considering the factors affecting the object's travel based on the travel preference type.

[0097] For example: when the travel preference type is a private car, the factors affecting the object's travel may include the path length, the traffic congestion status, and the number of traffic lights configured, etc., and a path with fewer traffic congestion times and fewer traffic lights configured is generated.

[0098] In some embodiments, in addition to selecting the travel preference based on the interaction interface, with the user's authorization, the travel information of the user in the historical period can also be obtained. Then, the large model can be used to analyze the travel information in the historical period to generate the travel preference.

[0099] When generating a path, considering the user's travel preferences, more candidate paths that better meet the user's preferences can be generated before path analysis, thereby reducing the amount of data analyzed by the large model in the path analysis process and further improving the efficiency of information recommendation.

[0100] In actual application scenarios, the main factor affecting travel is travel safety. Therefore, the embodiments of the present disclosure analyze the travel safety of each path by combining historical traffic information and traffic setting status information, further improving the matching degree between the recommended result and the user's travel safety requirements.

[0101] According to an embodiment of the present disclosure, using a large model, by analyzing multiple paths, determining a first target institution from multiple candidate institutions may include the following operations: obtaining the historical traffic information and traffic facility status information of each of the multiple paths; using the large model to analyze the historical traffic information and traffic facility status information of each of the multiple paths to generate the travel safety of each of the multiple paths; and based on the travel safety, determining the first target institution from multiple candidate institutions.

[0102] According to an embodiment of the present disclosure, the historical traffic information may include: historical traffic condition information, for example: the number of congestion times, the duration of each congestion, etc. It may also include: the probability information of traffic accidents, the information of traffic restricted vehicle types, etc. The traffic setting status information may include: the number of traffic signal configurations, whether there is a pedestrian overpass or an underground pedestrian passage configured, the type of public transportation stations, etc.

[0103] Figure 6 Schematically shows a schematic diagram of path safety analysis using a large model according to an embodiment of the present disclosure.

[0104] As Figure 6 shown, for each path in the multiple paths 332A for the second object to go to each candidate institution, a large model is called to generate the travel safety 343 of each path based on the historical traffic information 341 and traffic setting status information 342 of each path.

[0105] For example: the travel safety of path Pa for the second object to go to candidate institution In1 is 0.8, and the travel safety of path Pb is 0.85. In some embodiments, the travel safety of the second object to go to candidate institution In1 may be determined based on the average value of the travel safety of multiple paths to the same candidate path, for example: 0.825.

[0106] Since the travel safety of the second object to go to candidate institution In1 is 0.825, which is higher than the travel safety of the second object to go to other candidate institutions, therefore, the first target institution can be determined as institution In1 344.

[0107] In some embodiments, the travel safety level of the second object going to each candidate institution can also be displayed to the first object through an interactive interface. The first object can select preference weights for each path on the interactive interface. The sum of the preference weights of all paths can be 1. Then, based on the preference weights selected by the first object and the travel safety level of each path, the score of each path can be determined. So as to determine the first target institution according to the scores of each path.

[0108] For example: The preference weight selected by the first object for path Pa is 0.2, and the preference weight for path Pb is 0.2. Then the score of candidate institution In1 is 0.2×0.825 = 0.165. The preference weight selected by the second object for path Pc is 0.4. Then the score of candidate institution Inj is 0.4×0.5 = 0.2. The preference weight selected by the second object for path Pd is 0.3, and the score of candidate institution InM is 0.3×0.6 = 0.18. The preference weight selected by the second object for path Pe is 0.1, and the score of candidate institution In2 is 0.1×0.6 = 0.06. Therefore, the first target institution can be determined as the institution Inj with the highest score.

[0109] In the embodiments of the present disclosure, by dynamically analyzing the travel safety level of the path of the second object going to each candidate institution in combination with historical traffic information and traffic setting status information, the matching degree between the recommended result and the travel safety needs of the user is further improved, and the user experience is improved.

[0110] The premise of information recommendation based on the publicity information released by candidate institutions on the Internet and the evaluation information of different users on candidate institutions is that the publicity information and evaluation information are highly credible. However, due to the lack of a systematic evaluation mechanism for the authenticity of the information published by each application platform, the matching degree between the recommended result and the user needs is reduced.

[0111] In view of this, the embodiments of the present disclosure may further include the following operations: extracting information to be verified from the publicity information of multiple candidate institutions; obtaining associated information associated with the information to be verified from a predetermined data source; using a large model to verify the information to be verified based on the associated information to generate the credibility of multiple candidate institutions; and determining a second target institution from multiple candidate institutions based on the credibility.

[0112] In some embodiments, the publicity information of each candidate institution can be constructed as a knowledge graph. For example: The nodes in the knowledge graph can include teacher nodes, teaching experience nodes of teachers, institution environment nodes, institution qualification nodes, etc. The information to be verified can include at least two nodes and edge information for associating the at least two nodes.

[0113] Figure 7 Schematically shows a schematic diagram of credibility analysis using a large model according to an embodiment of the present disclosure.

[0114] As Figure 7 shown, for the candidate institution In M 351, the information 352 to be verified extracted from the candidate institution In M 351 can be "Teacher A won the Excellent Teacher Award in City B xx years ago" 352.

[0115] Then, the associated information 353 associated with the information 352 to be verified obtained from the predetermined data source 302 can be "List of Excellent Teacher Award Winners in City B in xx years". The predetermined data source can be a database storing information with relatively high credibility.

[0116] Next, use the large model 320 to verify the information to be verified based on the associated information. For example: check whether "Teacher A" is in the "List of Excellent Teacher Award Winners in City B in xx years". When it is determined that "Teacher A" is in the "List of Excellent Teacher Award Winners in City B in xx years", the generated verification result can be 1. When it is determined that "Teacher A" is not in the "List of Excellent Teacher Award Winners in City B in xx years", the generated verification result can be 0.

[0117] In the embodiment of the present disclosure, there can be multiple pieces of information to be verified, and the credibility 354 can be the average value of the verification results of multiple pieces of information to be verified.

[0118] Finally, based on the credibility 354, an institution with a credibility greater than the predetermined credibility threshold can be determined as the second target institution 355. It is also possible to display, through an interaction interface, institutions with a credibility greater than the predetermined credibility threshold to the first object, so that the first object can select an institution that meets their intended needs as the second target institution 355.

[0119] In the embodiment of the present disclosure, the second target institution 355 can be used as the candidate institution for path analysis in the method described above, thereby reducing the amount of redundant information processed by the large model and improving the information recommendation efficiency. The second target institution 355 can also be, based on the first target institution determined in the method described above, through credibility verification, finally determine whether to recommend the first target institution to the first object. Thereby further improving the credibility of the recommended information.

[0120] In the embodiment of the present disclosure, by performing credibility verification on the information of the candidate institution, the credibility of the recommendation result is further improved, and the user experience is improved.

[0121] Figure 8 Schematically shows a block diagram of an information recommendation device based on a large model according to an embodiment of the present disclosure.

[0122] As Figure 8 shown, the device 800 may include: a receiving module 810, an analysis module 820, and a determination module 830.

[0123] A receiving module 810, configured to receive the historical experience information of a second object and the location information of the second object input by a first object.

[0124] An analysis module 820, configured to analyze the historical experience information by using a large model to generate the ability characteristic information of the second object.

[0125] A determination module 830, configured to use a large model to determine, based on the location information and the location information of multiple candidate institutions, a first target institution for recommending to the first object by analyzing multiple paths for the second object to go to the multiple candidate institutions.

[0126] According to an embodiment of the present disclosure, the analysis module 820 includes a first generation sub-module and a first determination sub-module.

[0127] The first generation sub-module is configured to generate multiple paths by using a large model based on the location information and the location information of multiple candidate institutions.

[0128] The first determination sub-module is configured to determine the first target institution from multiple candidate institutions by analyzing the multiple paths by using a large model.

[0129] According to an embodiment of the present disclosure, the first determination sub-module includes: an acquisition unit, an analysis unit, and a determination unit.

[0130] The acquisition unit is configured to acquire the historical traffic information and the traffic facility status information of each of the multiple paths.

[0131] The analysis unit is configured to analyze the historical traffic information and the traffic facility status information of each of the multiple paths by using a large model to generate the travel safety degree of each of the multiple paths.

[0132] The determination unit is configured to determine the first target institution from multiple candidate institutions based on the travel safety degree.

[0133] According to an embodiment of the present disclosure, the first generation sub-module includes: a reception unit and a generation unit.

[0134] The reception unit is configured to receive the travel preference information selected by the first object.

[0135] The generation unit is configured to generate multiple paths by using a large model based on the travel preference information, the location information, and the location information of multiple candidate institutions.

[0136] According to an embodiment of the present disclosure, the analysis module 820 includes: an acquisition sub-module and a second generation sub-module.

[0137] The acquisition sub-module is configured to acquire the behavior preference information from the historical experience information.

[0138] A second generation sub-module, configured to analyze the behavioral preference information using a large model to generate ability characteristic information.

[0139] According to an embodiment of the present disclosure, the above device further includes: an intention receiving module, a generation module, and a candidate institution determination module.

[0140] The intention receiving module is configured to receive the intention information selected by the first object.

[0141] The generation module is configured to analyze the intention information and the ability characteristic information using a large model to generate a course type that matches the ability characteristic information.

[0142] The candidate institution determination module is configured to determine a plurality of candidate institutions based on the course type.

[0143] According to an embodiment of the present disclosure, the above device further includes: an extraction module, an acquisition module, a verification module, and a target determination module.

[0144] The extraction module is configured to extract information to be verified from the promotional information of a plurality of candidate institutions.

[0145] The acquisition module is configured to obtain associated information associated with the information to be verified from a predetermined data source.

[0146] The verification module is configured to use a large model to verify the information to be verified based on the associated information and generate the credibility of a plurality of candidate institutions.

[0147] The target determination module is configured to determine a second target institution from a plurality of candidate institutions based on the credibility.

[0148] According to an embodiment of the present disclosure, the present disclosure also provides an agent for information recommendation, an electronic device, a readable storage medium, and a computer program product.

[0149] According to an embodiment of the present disclosure, an agent for information recommendation includes: an input module, a processing module, and an output module.

[0150] The input module is configured to receive the historical experience information of the second object and the location information of the second object input by the first object.

[0151] The processing module is configured to determine a target task based on the historical experience information and the location information received by the input module, determine a target large model based on the target task, and execute the method described above by calling the target large model to obtain information about the first target institution.

[0152] The output module is configured to output the information about the first target institution obtained by the processing module.

[0153] According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method as described above.

[0154] According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method as described above.

[0155] According to an embodiment of the present disclosure, a computer program product includes a computer program, and the computer program implements the method as described above when executed by a processor.

[0156] Figure 9 A block diagram of an agent for information recommendation according to an embodiment of the present disclosure is schematically shown.

[0157] As Figure 9 shown, in an embodiment of the present disclosure, inspired by the von Neumann architecture in modern computer theory, as Figure 9 shown, the AI agent 900 may include three core modules: an input module 910, an output module 920, and a processing module 930. The processing module 930 may include a control unit 931, a storage unit 932, and an arithmetic unit 933.

[0158] The input module 910 is responsible for receiving or perceiving information such as queries, requests, instructions, signals, or data from the outside world (such as a user or an external environment), and converting it into a format that the AI agent 900 can understand and process. The input module 910 is the primary link for the AI agent 900 to interact with the outside world, enabling the AI agent 900 to efficiently and accurately obtain necessary "sensory" information from the outside world and respond to this information.

[0159] In an example, the input information received by using the input module 910 may be the historical experience information of the second object and the location information of the second object described above.

[0160] In an example, the processing module 930 is the core support for the AI agent 900 to handle complex tasks. The processing module 930 may determine a target task based on the input information received by the input module 910, determine a large model based on the target task, and output target institution information by invoking the large model to execute the information recommendation method based on the large model described above.

[0161] In the example, during operation, the control unit 931 in the processing module 930 will continuously interact with the storage unit 932, the arithmetic unit 933, and / or the output module 920. However, it should be noted that in the embodiments of the present disclosure, the control unit 931 acts as a single initiator to initiate communication with the storage unit 932, the arithmetic unit 933, and / or the output module 920, and there is no communication coupling between the storage unit 932, the arithmetic unit 933, and the output module 920.

[0162] In the example, the performance of the control unit 931 can be closely related to the large model on which the AI agent 900 is based. To fully utilize the capabilities of the large language model, the internal structure of the control unit 1031 can be designed to be highly configurable and extensible to handle various different types of tasks and requirements in real-world scenarios.

[0163] The storage unit 932 can be responsible for memorizing information such as historical conversations and event streams. The configuration information, target text, and data resources generated in each round as described above can be included in the storage unit 932.

[0164] In the example, after the AI agent 900 obtains a configuration generation request, the AI agent 1000 can use an intent recognition model to determine the configuration intent from the initial text. The configuration intent can be stored in the storage unit 932. The AI agent 900 can retrieve relevant data resources from the storage unit 1032 and feedback them to the control unit 931. Then, the control unit 931 can use the feedback data resources to obtain configuration data corresponding to the initial text. Relevant text data can also be retrieved from the storage unit 932 and feedback to the control unit 931. Then, the control unit 931 can use the returned text data to obtain the target text. And the target text and configuration data are passed to the output module 920.

[0165] The arithmetic unit 1033 can be regarded as a predefined tool library. Renderers, display controls, etc. as described above can be included in the arithmetic unit 933.

[0166] In the example, when the AI agent 1000 needs to render multiple output data, relevant renderers and display controls can be called from the operation unit 933 and fed back to the control unit 932. Then, the control unit 932 can use the fed-back renderers and display controls to render the first search result and pass the first search result to the output module 920. It can be understood that although large language models have excellent language understanding and generation capabilities, like humans, the tasks they can solve without any tools are very limited. When the AI agent 900 is given the ability to call tools, it can perform tasks such as completing mathematical operations with the help of a calculator, performing data analysis with the help of Python, and completing prediction tasks with the help of a search engine.

[0167] In the example, the output module 920 can output the target institution information described above.

[0168] The AI agent 900 according to the embodiments of the present disclosure can simply and effectively improve the degree of intelligence and enhance flexibility and versatility.

[0169] Figure 10 FIG. shows a schematic block diagram of an example electronic device 1000 that can be used to implement the embodiments of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0170] As Figure 8 shown, the device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0171] Multiple components in device 1000 are connected to I / O interface 1005, including: input unit 1006, such as a keyboard, mouse, etc.; output unit 1007, such as various types of displays, speakers, etc.; storage unit 1008, such as a disk, optical disc, etc.; and communication unit 1009, such as a network card, modem, wireless communication transceiver, etc. Communication unit 1009 allows device 1000 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0172] Computing unit 1001 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Computing unit 1001 executes the various methods and processes described above, such as the information recommendation method based on a large model. For example, in some embodiments, the information recommendation method based on a large model can be implemented as a computer software program that is tangibly contained in a machine-readable medium, such as storage unit 1008. In some embodiments, part or all of the computer program can be loaded and / or installed onto device 1000 via ROM 1002 and / or communication unit 1009. When the computer program is loaded into RAM 1003 and executed by computing unit 1001, one or more steps of the information recommendation method based on a large model described above can be executed. Alternatively, in other embodiments, computing unit 1001 can be configured to execute the information recommendation method based on a large model in any other suitable manner (e.g., by means of firmware).

[0173] Various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor, and can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0174] The program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing device, such that the program codes, when executed by the processor or controller, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine, or entirely on the remote machine or server.

[0175] In the context of the present disclosure, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0176] In order to provide interaction with a user, the systems and techniques described herein may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0177] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), and the Internet.

[0178] A computer system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. 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, a server of a distributed system, or a server incorporating a blockchain.

[0179] It should be understood that various forms of the processes shown above can be used, steps can be reordered, added, or deleted. For example, the steps recited in this disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and no limitations are imposed herein.

[0180] The above specific embodiments do not constitute a limitation on the protection scope 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 principle of this disclosure shall be included within the protection scope of this disclosure.

Claims

1. An information recommendation method based on a large model, comprising: Receiving historical experience information of a second object and location information of the second object input by a first object; Analyzing the historical experience information by using the large model to generate ability characteristic information of the second object; And Using the large model, based on the location information and location information of multiple candidate institutions, by analyzing multiple paths for the second object to go to the multiple candidate institutions, determining a first target institution for recommending to the first object from the multiple candidate institutions.

2. The method according to claim 1, wherein, The step of using the large model, based on the location information and location information of multiple candidate institutions, by analyzing multiple paths for the second object to go to the multiple candidate institutions, determining a first target institution for recommending to the first object from the multiple candidate institutions includes: Using the large model, based on the location information and location information of multiple candidate institutions, to generate the multiple paths; and Using the large model, by analyzing the multiple paths, determining the first target institution from the multiple candidate institutions.

3. The method according to claim 2, wherein The step of using the large model, by analyzing the multiple paths, determining the first target institution from the multiple candidate institutions includes: Obtaining historical traffic information and traffic facility status information of each of the multiple paths; Analyzing the historical traffic information and traffic facility status information of each of the multiple paths by using the large model to generate travel safety levels of each of the multiple paths; and Based on the travel safety levels, determining the first target institution from the multiple candidate institutions.

4. The method according to claim 2 or 3, wherein The step of using the large model, based on the location information and location information of multiple candidate institutions, to generate the multiple paths includes: Receiving travel preference information selected by the first object; and Using the large model, based on the travel preference information, the location information and location information of multiple candidate institutions, to generate the multiple paths.

5. The method according to claim 1, wherein The step of analyzing the historical experience information by using the large model to generate ability characteristic information of the second object includes: Obtaining behavior preference information from the historical experience information; and Analyzing the behavior preference information by using the large model to generate the ability characteristic information.

6. The method according to any one of claims 1-5, wherein The method further includes: Receiving intention information selected by the first object; Analyzing the intention information and the ability characteristic information by using the large model to generate a course type matching the ability characteristic information; and Based on the course type, determining the multiple candidate institutions.

7. The method according to any one of claims 1-6, wherein, The method further includes: Extracting information to be verified from promotional information of the multiple candidate institutions; Obtaining associated information associated with the information to be verified from a predetermined data source; Using the large model, based on the associated information, to verify the information to be verified and generate credibility levels of the multiple candidate institutions; and Based on the credibility levels, determining a second target institution from the multiple candidate institutions.

8. An information recommendation device based on a large model, comprising: A receiving module, configured to receive the historical experience information of a second object and the location information of the second object input by a first object; An analysis module, configured to analyze the historical experience information by using a large model to generate the capability characteristic information of the second object; A determination module, configured to use the large model, based on the location information and the location information of multiple candidate institutions, by analyzing multiple paths for the second object to go to the multiple candidate institutions, determine a first target institution for recommendation to the first object from the multiple candidate institutions.

9. An intelligent agent for information recommendation, comprising: An input module, configured to receive the historical experience information of a second object and the location information of the second object input by a first object; A processing module, configured to determine a target task based on the historical experience information and the location information received by the input module, determine a target large model based on the target task, and execute the method according to any one of claims 1-7 by calling the target large model to obtain the information of the first target institution; And An output module, configured to output the information of the first target institution obtained by the processing module.

10. An electronic device, comprising: At least one processor; And A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method according to any one of claims 1-7.

11. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the method according to any one of claims 1-7.

12. A computer program product, comprising a computer program, where the computer program, when executed by a processor, implements the method according to any one of claims 1-7.

Images