Interest label determination method and device

Abstract

The invention provides an interest label determination method and device, and relates to the technical field of artificial intelligence, in particular to the technical fields of large models, automatic driving, intelligent traffic and the like. According to the specific implementation scheme, an interaction behavior sequence of a target object in a map application is determined; determining candidate interest tags associated with the interactive behaviors in the interactive behavior sequence; and determining a target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with the interaction behaviors in the interaction behavior sequence. According to the mode, the accuracy of the generated target interest tag is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, specifically to technical fields such as large models, autonomous driving, and intelligent transportation, and more particularly to a method and device for determining interest tags. Background Art

[0002] How to model user interests to support the commercial search and push of map applications is the key to whether map applications can achieve profit and loss balance and strategic goals. In other words, the effectiveness of user interest modeling in map scenarios directly affects key indicators such as user stickiness, daily active users, and revenue of map applications. Summary of the Invention

[0003] Embodiments of the present disclosure provide a method, apparatus, device, and storage medium for determining interest tags.

[0004] In a first aspect, an embodiment of the present disclosure provides a method for determining an interest tag, the method comprising: determining an interaction behavior sequence of a target object within a map application; determining candidate interest tags associated with each interaction behavior in the interaction behavior sequence; and determining a target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence.

[0005] In the second aspect, an embodiment of the present disclosure provides an interest tag determination device, which includes: a first determination module, a second determination module and a third determination module, wherein the first determination module is configured to determine the interaction behavior sequence of the target object within the map application; the second determination module is configured to determine the candidate interest tags associated with each interaction behavior in the interaction behavior sequence; the third determination module is configured to determine the target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence.

[0006] In a third aspect, an embodiment of the present disclosure provides an electronic device comprising one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by the one or more processors, the one or more processors implement a method for determining an interest tag as in any embodiment of the first aspect.

[0007] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the interest tag determination method of any embodiment of the first aspect.

[0008] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program, which, when executed by a processor, implements the interest tag determination method of any embodiment of the first aspect.

[0009] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 is an exemplary system architecture diagram in which the present disclosure may be applied; Figure 2 is a flowchart of an embodiment of a method for determining interest tags according to the present disclosure; Figure 3 is a flowchart of another embodiment of the method for determining interest tags according to the present disclosure; Figure 4 is a flowchart of another embodiment of the method for determining interest tags according to the present disclosure; Figure 5 is a schematic diagram of an application scenario of the interest tag determination method according to the present disclosure; Figure 6 is a schematic diagram of an embodiment of an apparatus for determining interest tags according to the present disclosure; Figure 7 It is a structural diagram of a computer system suitable for implementing the electronic device of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0011] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0012] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0013] Figure 1 An exemplary system architecture 100 is shown to which an embodiment of the interest tag determination method of the present disclosure can be applied.

[0014] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. Network 104 is a medium for providing communication links between terminal devices 101, 102, 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0015] Users can use terminal devices 101 , 102 , 103 to interact with server 105 via network 104 to receive or send messages, etc.

[0016] Terminal devices 101, 102, and 103 may be hardware or software. When terminal devices 101, 102, and 103 are software, they may be installed in the electronic devices listed above. They may be implemented as multiple software programs or software modules, or as a single software program or software module. This is not specifically limited here.

[0017] Server 105 can be a server that provides various services, for example, determining the interaction behavior sequence of the target object in the map application; determining the candidate interest tags associated with each interaction behavior in the interaction behavior sequence; and determining the target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence.

[0018] It should be noted that server 105 can be either hardware or software. When server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When the server is software, it can be implemented as multiple software programs or software modules (for example, to provide interest tag determination services), or as a single software program or software module. No specific limitations are imposed here.

[0019] It should be noted that the interest tag determination method provided in the embodiments of the present disclosure can be executed by the server 105, or by the terminal devices 101, 102, 103, or by the server 105 and the terminal devices 101, 102, 103 in cooperation with each other. Accordingly, the various parts (such as various units, sub-units, modules, and sub-modules) included in the interest tag determination apparatus can be all set in the server 105, or all set in the terminal devices 101, 102, 103, or can be set in the server 105 and the terminal devices 101, 102, 103 respectively.

[0020] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0021] Figure 2 The process 200 of an embodiment of a method for determining an interest tag is shown. The method for determining an interest tag may specifically include the following steps: Step 201: Determine the interaction behavior sequence of the target object in the map application.

[0022] In this embodiment, the execution subject (for example, Figure 1The server 105 or the terminal devices 101, 102, 103) can collect the interactive behaviors of the target object in the map application in real time or periodically, and generate an interactive behavior sequence based on the collection time of the interactive behaviors.

[0023] Interactions can be any interactions between a target object and a map application, such as check-in, stay, search, navigation, route calculation, hotel reservation, reading reviews, location sharing, etc. Interactions can be based on points of interest or not, and this application does not limit this. It should be noted that the collection of interaction behaviors is performed with user authorization.

[0024] Here, the number of the interaction behavior sequences may be one or more. If the number of the interaction behavior sequences is multiple, the collection time periods corresponding to the interaction behavior sequences are different.

[0025] For example, the number of interactive behavior sequences can be four, namely, long-term behavior sequence (collection time period t1-t10), medium-long-term behavior sequence (collection time period t3-t10), medium-short-term behavior sequence (collection time period t5-t10) and short-term behavior sequence (collection time period t7-t10); among which, the length of the collection time periods corresponding to the long-term behavior sequence, medium-long-term behavior sequence, medium-short-term behavior sequence and short-term behavior sequence are shortened successively.

[0026] Step 202: Determine candidate interest tags associated with each interactive behavior in the interactive behavior sequence.

[0027] In this embodiment, after determining the sequence of interactive behaviors, the executing entity may first determine the associated object of each interactive behavior based on the interactive behavior, and then determine the candidate interest tags associated with the interactive behavior and the weights of the candidate interest tags based on the interest tags of the associated object of the interactive behavior and the weights of the interest tags.

[0028] The number of interest tags of the associated object can be one or more, which is not limited in this application. Interest tags are keywords used to describe the user's hobbies, areas of interest or preferences, such as listening to concerts, watching operas, traveling, reading, etc.

[0029] Here, the executing entity may directly determine the interest tags and interest tag weights of the associated objects of the interactive behavior as the candidate interest tags and the weights of the candidate interest tags associated with the interactive behavior; or may determine the interest tags and interest tag weights with the highest weights among the interest tags of the associated objects of the interactive behavior as the candidate interest tags and the weights of the candidate interest tags associated with the interactive behavior. This application does not limit this.

[0030] Among them, the associated objects of interactive behaviors may include multiple types, such as points of interest, user-generated content (such as notes, reviews, etc.), coordinates / locations (non-points of interest), etc.

[0031] Specifically, the associated objects of interactive behaviors such as search and route calculation are points of interest; the associated objects of interactive behaviors such as reading notes and reading reviews are user-generated content; and the associated objects of interactive behaviors such as location navigation and location sharing are coordinates / locations.

[0032] The interest tags of the associated objects and the weights of the interest tags may be determined based on the basic attribute information and associated information of the associated objects.

[0033] Specifically, if the associated object is a point of interest, the interest tags and weights of the associated object can be determined based on the basic attribute information of the point of interest, such as place type, location, name, introduction, picture, etc., and associated attribute information, such as user comments, ratings, check-in data, consumption data, etc.; if the associated object is user-generated content, the interest tags and weights of the associated object can be determined based on the basic attribute information of the user-generated content, such as title, text, keyword frequency, category, etc., and associated information, such as keywords mentioned in user comments / forwards, number of comments / forwards, etc.; if the associated object is a coordinate / location, the interest tags and weights of the associated object can be determined based on the basic attribute information of the coordinate / location, such as name, longitude and latitude, etc., and associated attribute information, such as search records, click records, distribution and type of surrounding points of interest, street view pictures / satellite pictures, etc.

[0034] Step 203 : determining a target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence.

[0035] In this embodiment, the executing entity may directly determine the target interest tag of the target object based on the ranking information of the interactive behavior in the interactive behavior sequence and the candidate interest tags associated with the interactive behavior; or it may determine the target interest tag of the target object based on the ranking information of the interactive behavior in the interactive behavior sequence, the candidate interest tags associated with the interactive behavior, and the weight of the candidate interest tags. This application does not limit this.

[0036] Here, the execution subject may determine the target interest tag of the target object based on the sorting information of the interactive behavior in the interactive behavior sequence and the candidate interest tags associated with the interactive behavior in a variety of ways, for example, directly inputting the sorting information of the interactive behavior and the candidate interest tags associated with the interactive behavior into a preset generation model to generate the target interest tag of the target object; determining the target interest tag of the target object based on the sorting information of the interactive behavior, the candidate interest tags associated with the interactive behavior, and a preset mapping relationship table between the sorting information of the interactive behavior, the candidate interest tags associated with the interactive behavior and the interest tags.

[0037] Among them, the preset generation model can be any model for generating interest tags, such as a generative adversarial network, a large language model, etc.

[0038] It should be pointed out that, here, the execution subject may also determine the weight of the target interest tag based on the ranking information of the interactive behavior in the interactive behavior sequence and the candidate interest tags associated with the interactive behavior.

[0039] The interest tag determination method provided by the embodiments of the present disclosure determines the interaction behavior sequence of the target object in the map application; determines the candidate interest tags associated with each interaction behavior in the interaction behavior sequence; determines the target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence, takes into account the influence of the timing information of the interaction behavior and the interest tags of the objects associated with the interaction behavior on the generated target interest tags, thereby improving the accuracy of the generated target interest tags.

[0040] In some optional methods, based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence, the target interest tag of the target object is determined, including: based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence, the current interest tag of the target object is determined from the candidate interest tags; based on the current interest tag, the target interest tag of the target object is generated.

[0041] In this implementation, the execution subject may first determine the current interest tag of the target object from the candidate interest tags based on the sorting information of the interactive behavior in the interactive behavior sequence and the candidate interest tags associated with the interactive behavior.

[0042] The current interest tag of the target object may be one or more of the candidate interest tags.

[0043] Furthermore, the execution entity may generate a target interest tag of the target object based on the current interest tag of the target object.

[0044] Here, the target interest tag is different from the candidate interest tags.

[0045] Specifically, the execution subject can directly input the current interest tag of the target object into the preset generation model to generate the target interest tag of the target object.

[0046] This implementation method determines the current interest tag of the target object from the candidate interest tags based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence; based on the current interest tag, the target interest tag of the target object is generated, thereby realizing the prediction of the future interest tag of the target object.

[0047] In some optional methods, determining candidate interest tags associated with each interactive behavior in an interactive behavior sequence includes: determining an initial interest tag associated with each interactive behavior in the interactive behavior sequence; in response to determining that the initial interest tag is an open tag, mapping the initial interest tag to a closed interest tag to obtain candidate interest tags associated with each interactive behavior.

[0048] In this implementation, for each interactive behavior in the interactive behavior sequence, the execution subject may first determine the initial interest tag associated with the interactive behavior based on the interest tag of the object associated with the interactive behavior.

[0049] Furthermore, for the initial interest tag, the executing entity can determine whether the initial interest tag is an open interest tag. If so, the embedded vector associated with the interest tag and the similarity of the vector can be used to map the initial interest tag to a closed interest tag to obtain the candidate interest tags associated with each interactive behavior.

[0050] Among them, open tags refer to tags that do not have a fixed range or options pre-set, and users can freely add or modify content according to their needs; closed tags refer to a fixed set of tags that are pre-set, and users can only choose from these options and cannot add new tags on their own.

[0051] This implementation method determines the initial interest tag associated with each interactive behavior in the interactive behavior sequence; in response to determining that the initial interest tag is an open tag, the initial interest tag is mapped to a closed interest tag to obtain candidate interest tags associated with each interactive behavior. It utilizes the characteristics of closed tags that are highly structured and easy to manage data, while reducing noise, effectively improving the efficiency and accuracy of generating target tags based on candidate interest tags.

[0052] Further references Figure 3 , which shows Figure 2 The process 300 of another embodiment of the method for determining interest tags is shown. In this embodiment, the process 300 of the method for determining interest tags may include the following steps: Step 301: Determine the interaction behavior sequence of the target object in the map application.

[0053] In this embodiment, the implementation details and technical effects of step 301 can be found in the description of step 201 and will not be repeated here.

[0054] Step 302: Determine candidate interest tags associated with each interactive behavior in the interactive behavior sequence.

[0055] In this embodiment, the implementation details and technical effects of step 302 can be found in the description of step 202 and will not be repeated here.

[0056] Step 303: Based on the ranking information and categories of the interactive behaviors in the interactive behavior sequence, the initial weights of the candidate interest tags associated with the interactive behaviors are adjusted to obtain target weights of the candidate interest tags.

[0057] In this embodiment, for each interactive behavior in the interactive behavior sequence, the executing entity can determine the candidate interest tags associated with the interactive behavior based on the interest tags of the objects associated with the interactive behavior; and determine the initial weight of the candidate interest tags associated with the interactive behavior based on the weight of the interest tags of the objects associated with the interactive behavior.

[0058] Furthermore, the execution entity may adjust the initial weight of the candidate interest tag according to the ranking information and category of the interaction behavior corresponding to the candidate interest tag in the interaction behavior sequence to obtain the target weight of the candidate interest tag.

[0059] Here, different ranking information adjusts the initial weights by different amounts. Generally, the later the ranking information, the larger the corresponding adjustment. That is, the closer the interaction behavior is collected to the current moment, the greater the impact on the weight of the candidate interest tag. For example, in the interaction behavior sequence, the ranking information corresponding to each interaction is first, second, third, and fourth, and the corresponding adjustment amounts for each ranking information are +0.1, +0.2, +0.3, and +0.4, respectively.

[0060] Different categories have different adjustment ranges for the initial weights. For example, the adjustment range for the initial weights of point of interest interaction behaviors is greater than the adjustment range for the initial weights of content consumption interaction behaviors.

[0061] Specifically, the adjustment range of the initial weight for the point of interest interaction behavior is +0.3, and the adjustment range of the initial weight for the content consumption interaction behavior is +0.1.

[0062] Here, it should be pointed out that the category of the interactive behavior can be determined based on a preset classification method of the interactive behavior category.

[0063] There are multiple ways to classify the preset interactive behaviors, for example, classification according to associated objects of the interactive behaviors, classification according to the interactive modes corresponding to the interactive behaviors, and the like.

[0064] Specifically, if the preset interactive behavior classification method is to classify according to the associated objects of the interactive behavior, the interactive behavior categories may include point of interest interaction category, content consumption category, etc., among which the point of interest interaction category may include point of interest inquiry category and financial consumption category; if the preset interactive behavior classification method is to classify according to the interactive method corresponding to the interactive behavior, the interactive behavior categories may include residence category, search category, check-in category, etc.

[0065] Step 304 : Based on the target weights of the candidate interest tags, determine the current interest tag of the target object and the weight of the current interest tag from among the candidate interest tags.

[0066] In this embodiment, after determining the target weights of the candidate interest tags associated with each interactive behavior in the interactive behavior sequence, the executing entity may perform a weighted sum of the target weights of the same candidate interest tags, that is, merge the same candidate interest tags into one candidate interest tag, and determine the weighted sum of the weights of the same candidate interest tags as the new weight of the merged candidate interest tag, sort the merged candidate interest tags in descending order according to the new weights, and determine a preset number (for example, 3, 2, etc.) of the merged candidate interest tags with a higher ranking as the current interest tag, and determine the weights of the preset number of merged candidate interest tags with a higher ranking as the weight of the current interest tag.

[0067] Step 305: Generate a target interest tag for the target object based on the current interest tag and the weight of the current interest tag.

[0068] In this embodiment, the execution entity may input the current interest tag and the weight of the current interest tag into a preset generation model to generate a target interest tag for the target object.

[0069] Here, the preset generation model can be any model for generating interest tags, such as a generative adversarial network, a large language model, etc.

[0070] The embodiments of the present disclosure adjust the initial weights of candidate interest tags associated with interactive behaviors based on the ranking information and categories of interactive behaviors in the interactive behavior sequence to obtain target weights of candidate interest tags; based on the target weights of candidate interest tags, the current interest tags of the target object and the weights of the current interest tags are determined in the candidate interest tags; based on the current interest tags and the weights of the current interest tags, the target interest tags of the target object are generated, which fully considers the influence of the categories of interactive behaviors in the interactive behavior sequence on the weights of interest tags and improves the accuracy of the determined target tags.

[0071] In some optional manners, the categories of interactive behaviors may include: point of interest inquiry, financial consumption, and content consumption.

[0072] In this implementation, the categories of interactive behaviors may include multiple categories, such as interest point inquiry, financial consumption, and content consumption.

[0073] Among them, the point of interest inquiry interaction behavior is used to indicate the behavior of querying information about specific places or services through map applications, such as searching for restaurants, navigating to hotels, etc.

[0074] Financial consumption interaction behaviors are used to indicate the completion of consumption-related operations through map applications, such as booking a hotel or taking a taxi.

[0075] Content consumption interaction behaviors are used to indicate behaviors of obtaining content information related to a geographic location through a map application, such as reading reviews and reading notes.

[0076] This implementation method improves the richness of interactive behaviors and achieves effective division of interactive behaviors by setting the categories of interactive behaviors to include: point of interest inquiry category, financial consumption category, and content consumption category.

[0077] In some optional methods, the initial weights of the candidate interest tags associated with the interactive behaviors are adjusted based on the sorting information and categories of the interactive behaviors in the interactive behavior sequence to obtain the target weights of the candidate interest tags, including: adjusting the initial weights of the candidate interest tags associated with the interactive behaviors based on the sorting information of the interactive behaviors in the interactive behavior sequence to obtain the intermediate weights; in response to determining that there is a specified behavior in the interactive behavior sequence, adjusting the intermediate weights of the candidate interest tags associated with the specified behavior according to the time point corresponding to the specified behavior and the interactive behavior before the specified behavior in the interactive behavior sequence that is the same as the candidate interest tag associated with the specified behavior to obtain the target weight.

[0078] In this implementation, for each interactive behavior in the interactive behavior sequence, the execution entity may first adjust the initial weight of the candidate interest tag associated with the interactive behavior according to the sorting information of the interactive behavior in the interactive behavior sequence to obtain the intermediate weight of the candidate interest tag.

[0079] Furthermore, the executing entity can determine whether there is a specified behavior in the interactive behavior sequence. If so, the intermediate weight of the candidate interest tag associated with the specified behavior can be adjusted based on the time point corresponding to the specified behavior and the interactive behavior before the specified behavior in the interactive behavior sequence that is the same as the candidate interest tag associated with the specified behavior to obtain the target weight of the candidate interest tag.

[0080] Here, the designated behavior is a financial consumption interaction behavior, such as booking a hotel, taking a taxi, paying for shopping, buying tickets, paying for parking, etc.

[0081] Among them, the time point corresponding to the specified behavior and the interactive behavior before the specified behavior with the same candidate interest tag associated with the specified behavior are used to determine the difficulty parameter for the specified behavior to achieve the execution goal.

[0082] For example, if the designated behavior is taking a taxi, the difficulty of achieving the execution goal during peak hours in the morning and evening is higher than that during non-peak hours; if the designated behavior is shopping payment, and the candidate interest tag of the designated behavior is "buying clothes", the more interactive behaviors (such as searching, reading notes, taking a taxi, etc.) with the associated candidate interest tag "buying clothes" before the designated behavior, the greater the difficulty parameter of achieving the execution goal of the designated behavior.

[0083] Here, the difficulty parameter for a given behavior to achieve the execution goal is usually positively correlated with the adjustment amplitude of the intermediate weight, that is, the larger the difficulty parameter, the larger the adjustment amplitude. For example, when the difficulty parameters are 2 and 5, the adjustment amplitudes are +0.2 and +0.5 respectively.

[0084] This implementation method adjusts the initial weights of candidate interest tags associated with interactive behaviors based on the sorting information of interactive behaviors in the interactive behavior sequence to obtain intermediate weights; in response to determining the presence of a specified behavior in the interactive behavior sequence, the intermediate weights of candidate interest tags associated with the specified behavior are adjusted according to the time point corresponding to the specified behavior and the interactive behavior that precedes the specified behavior in the interactive behavior sequence and is the same as the candidate interest tag associated with the specified behavior to obtain the target weight. This method takes into account the impact of the financial consumption category on the weights of interest tags, achieves further adjustment of the weights of candidate interest tags, and improves the accuracy of the determined target weights of candidate interest tags.

[0085] In some optional methods, a target interest tag of the target object is generated based on the current interest tag and the weight of the current interest tag, including: generating a target interest tag of the target object based on the current interest tag, the weight of the current interest tag and the portrait information of the target object.

[0086] In this implementation, the execution entity may input the current interest tag, the weight of the current interest tag, and the portrait information of the target object into a preset generation model to generate a target interest tag for the target object.

[0087] Here, the portrait information of the target object may include multiple items, such as demographic information (such as gender, age, etc.), user search or query records (such as locations, routes, points of interest, etc.), search frequency and time period (such as querying commuting routes during peak hours in the morning and evening, searching for leisure places on weekends, etc.), commonly used transportation methods (such as driving, bus, subway, etc.), travel time patterns (such as weekday commuting time, weekend travel time), social relationships (such as location sharing records with friends, interest tags synchronized on social platforms for common travel routes, etc.), related consumption data (such as interaction with e-commerce platforms, interaction with life service accounts, etc.), life service preferences (consumption habits, dining preferences, travel scenarios, etc.), interactive behaviors (such as place collection, comments or ratings on points of interest, sharing operations, etc.), etc.

[0088] This implementation method generates the target interest tag of the target object based on the current interest tag, the weight of the current interest tag and the portrait information of the target object, fully considering the impact of the portrait information on the interest tags that the target object may be interested in in the future, and improving the accuracy of the generated target interest tag.

[0089] In some optional methods, a target interest tag for a target object is generated based on the current interest tag, the weight of the current interest tag, and the portrait information of the target object, including: inputting the current interest tag, the weight of the current interest tag, and the portrait information of the target object into a large language model to generate a target interest tag for the target object.

[0090] In this implementation, the execution entity may input the current interest tag, the weight of the current interest tag, and the portrait information of the target object into the large language model to generate the target interest tag of the target object.

[0091] Large Language Model (LLM) refers to a type of natural language processing model trained based on deep learning technology. It has the characteristics of wide knowledge breadth and strong language comprehension and generation capabilities. There are many types of large language models, such as the GPT (Generative Pre-trained Transformer) series and the BERT (Bidirectional Encoder Representations from Transformers) series.

[0092] This implementation method generates the target interest tag of the target object by inputting the current interest tag, the weight of the current interest tag, and the portrait information of the target object into the large language model, making full use of the large model's strong language understanding and generation capabilities, and improving the accuracy of the generated target interest tag.

[0093] Further references Figure 4 , which shows Figure 2 The process 400 of another embodiment of the method for determining interest tags is shown. In this embodiment, the process 400 of the method for determining interest tags may include the following steps: Step 401: Determine the interaction behavior sequence of the target object in the map application.

[0094] In this embodiment, the implementation details and technical effects of step 401 can be found in the description of step 301 and will not be repeated here.

[0095] Step 402: Determine candidate interest tags associated with each interactive behavior in the interactive behavior sequence.

[0096] In this embodiment, the implementation details and technical effects of step 402 can be found in the description of step 302 and will not be repeated here.

[0097] Step 403 : Based on the ranking information and categories of the interactive behaviors in the interactive behavior sequence, the initial weights of the candidate interest tags associated with the interactive behaviors are adjusted to obtain target weights of the candidate interest tags.

[0098] In this embodiment, the implementation details and technical effects of step 403 can be found in the description of step 303 and will not be repeated here.

[0099] Step 404 : Based on the target weights of the candidate interest tags, determine the long-term interest tags of the target object and the first weights of the long-term interest tags from among the candidate interest tags.

[0100] In this embodiment, there are two interactive behavior sequences, namely a long-term behavior sequence and a short-term behavior sequence.

[0101] Among them, the length of the first time period corresponding to the long-term behavior sequence is longer than the length of the second time period corresponding to the short-term behavior sequence. Usually, the second time period is a sub-time period of the first time period, and the starting time of the second time period is closer to the current time than the starting time of the first time period.

[0102] The execution entity can determine the long-term interest tag of the target object and the first weight of the long-term interest tag from the candidate interest tags based on the ranking information, category, etc. of each interactive behavior in the long-term behavior sequence and the candidate interest tags associated with each interactive behavior in the long-term behavior sequence.

[0103] Step 405 : Based on the target weights of the candidate interest tags, determine the short-term interest tags of the target object and the second weights of the short-term interest tags from among the candidate interest tags.

[0104] In this embodiment, the execution entity can determine the short-term interest tag of the target object and the second weight of the short-term interest tag from the candidate interest tags based on the sorting information, category, etc. of each interactive behavior in the short-term behavior sequence and the candidate interest tags associated with each interactive behavior in the short-term behavior sequence.

[0105] Step 406: Generate a target interest tag for the target object based on the long-term interest tag, the first weight, the short-term interest tag, and the second weight.

[0106] In this embodiment, the execution entity can first determine whether the long-term interest tag and the short-term interest tag meet the preset association conditions. If the long-term interest tag and the short-term interest tag meet the preset association conditions, the target interest tag of the target object can be generated based on the long-term interest tag, the first weight of the long-term interest tag, the short-term interest tag and the second weight of the short-term interest tag.

[0107] The preset association conditions may include one of the following: the short-term interest tag and the long-term interest tag are the same, the short-term interest tag and the long-term interest tag have an intersection, and the short-term interest tag is an extended tag of the long-term interest tag.

[0108] Specifically, if the long-term interest tag is "music" or "concert", and the short-term interest tag is "concert", and the long-term interest tag and the short-term interest tag meet the preset association conditions, then the target interest tag of the target object can be generated based on the long-term interest tag, the first weight of the long-term interest tag, the short-term interest tag, and the second weight of the short-term interest tag.

[0109] The above-mentioned embodiments of the present disclosure determine the long-term interest tags of the target object and the first weight of the long-term interest tags among the candidate interest tags based on the target weights of the candidate interest tags; determine the short-term interest tags of the target object and the second weight of the short-term interest tags among the candidate interest tags based on the target weights of the candidate interest tags; generate the target interest tags of the target object based on the long-term interest tags, the first weights, the short-term interest tags, and the second weights, taking into account the influence of the weights of the interest tags of long-term and short-term behavior sequences on the generated target interest tags, thereby further improving the accuracy of the generated target interest tags.

[0110] In some optional methods, a target interest tag for a target object is generated based on the long-term interest tag, the first weight, the short-term interest tag and the second weight, including: in response to determining that the long-term interest tag and the short-term interest tag do not meet the preset association conditions and the second weight is greater than the first weight, a target interest tag for the target object is generated based on the short-term interest tag and the second weight.

[0111] In this implementation, the execution entity may first determine whether the long-term interest tag and the short-term interest tag meet the preset association conditions.

[0112] If not, it can be further determined whether the second weight is greater than the first weight. If it is, that is, the target object has undergone interest migration, the target interest tag of the target object can be generated based on the short-term interest tag and the second weight of the short-term interest tag.

[0113] If it is less than, that is, the target object has occasional interest in the short-term interest tag and no interest migration occurs, then the target interest tag of the target object can be generated according to the long-term interest tag and the first weight of the long-term interest tag.

[0114] Specifically, if the long-term interest tags are "music" and "concert", and the short-term interest tag is "painting", the long-term interest tags and the short-term interest tags do not meet the preset association conditions, and the second weight of the short-term interest tag is greater than the first weight, that is, the target object has undergone interest migration, then the target interest tag of the target object can be generated based on the short-term interest tag and the second weight of the short-term interest tag, for example, "painting skills".

[0115] This implementation method generates a target interest tag for the target object based on the short-term interest tag and the second weight in response to determining that the long-term interest tag and the short-term interest tag do not meet the preset association conditions and the second weight is greater than the first weight, thereby realizing the prediction of the interest tag under the condition that the target object has undergone interest migration and improving the accuracy of the generated target interest tag.

[0116] Continue to see Figure 5 , Figure 53 is a schematic diagram of an application scenario of the method for determining interest tags according to this embodiment.

[0117] The execution subject 501 can determine the interactive behavior sequence of the target object in the map application, for example, searching for place A 502 (time t1, the first interactive behavior) - staying in place B 503 (time t2, the second interactive behavior) - reading notes K 504 (time t3, the third interactive behavior) - consuming in place C 505 (time t4, the fourth interactive behavior), where the interactive behavior at time t1, i.e. searching for place A, is associated with place A, and the interest tag of place A is listening to a concert, with a weight of 0.8; the interactive behavior at time t2, i.e. staying in place B, is associated with place B, and place B The interest tags of are listening to concerts, with a weight of 0.5, and watching stage plays, with a weight of 0.3; the interactive behavior at time t3 is reading note K, and the associated object of is note K, and the interest tag of note K is watching stage plays, with a weight of 0.2; the interactive behavior at time t4 is consumption at place C, and the associated object of is place C, and the interest tags of place A are listening to concerts, with a weight of 0.1, and watching operas, with a weight of 0.8; for each interactive behavior, the interest tags and weights of the associated objects of the interactive behavior can be directly determined as the candidate interest tags associated with the interactive behavior and the initial weights of the candidate interest tags.

[0118] Furthermore, the initial weights of the candidate interest tags may be adjusted according to the categories of the interactive behaviors and the ranking information of the interactive behaviors in the interactive behavior sequence to obtain target weights of the candidate interest tags.

[0119] For example, after adjustment, the weight of the candidate interest tag "listening to a concert" associated with the first interactive behavior is 0.83; the weight of the candidate interest tag "listening to a concert" associated with the second interactive behavior is 0.53, and the weight of "watching a stage play" is 0.33; the weight of the candidate interest tag "watching stage props" associated with the third interactive behavior is 0.21; the weight of the candidate interest tag "listening to a concert" associated with the fourth interactive behavior is 0.15, and the weight of "watching an opera" is 0.85.

[0120] Furthermore, the execution entity may merge the same candidate interest tags into one candidate interest tag, determine the sum of the same candidate interest tags as the weight of the merged candidate interest tag, and rank the merged candidate interest tags according to the corresponding weights of the merged interest tags, such as attending a concert (weight of 1.51 = 0.83 + 0.53 + 0.15), watching an opera (weight of 0.85), and watching a stage play (weight of 0.54 = 0.33 + 0.21). The execution entity may determine the candidate interest tag with the highest ranking as the current behavior tag 506, and determine the weight of the interest tag with the highest ranking as the weight of the current interest tag, such as attending a concert, with a weight of 1.51, and watching an opera, with a weight of 0.85.

[0121] Furthermore, the execution entity may determine the target interest tag 507 of the target object according to the current interest tag and the weight of the current interest tag, such as participating in international art festivals and music festivals.

[0122] Further references Figure 6 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of an interest tag determination device. Figure 2 The method embodiment shown corresponds to the embodiment shown.

[0123] like Figure 6 As shown, the interest tag determination device 600 of this embodiment includes: a first determination module 601 , a second determination module 602 and a theme update module 603 .

[0124] The first determining module 601 may be configured to determine an interactive behavior sequence of the target object within the map application.

[0125] The second determining module 602 may be configured to determine a candidate interest tag associated with each interaction behavior in the interaction behavior sequence.

[0126] The third determining module 603 may be configured to determine a target interest tag of a target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence.

[0127] In some optional embodiments, the third determination module includes a determination unit and a generation unit, wherein the determination unit is configured to determine the current interest tag of the target object from the candidate interest tags based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence; and the generation unit is configured to generate the target interest tag of the target object based on the current interest tag.

[0128] In some optional embodiments, the determination unit includes a first determination subunit and a second determination subunit. The first determination subunit is configured to adjust the initial weights of the candidate interest tags associated with the interactive behaviors based on the sorting information and categories of the interactive behaviors in the interactive behavior sequence to obtain the target weights of the candidate interest tags; the second determination subunit is configured to determine the current interest tags of the target object and the weight of the current interest tags among the candidate interest tags based on the target weights of the candidate interest tags; and the generation unit is configured to generate the target interest tags of the target object based on the current interest tags and the weight of the current interest tags.

[0129] In some optional manners, the categories of interactive behaviors include: point of interest inquiry, financial consumption, and content consumption.

[0130] In some optional embodiments, the first determination subunit is further configured to adjust the initial weights of the candidate interest tags associated with the interactive behaviors based on the sorting information of the interactive behaviors in the interactive behavior sequence to obtain an intermediate weight; in response to determining that there is a specified behavior in the interactive behavior sequence, the intermediate weights of the candidate interest tags associated with the specified behavior are adjusted according to the time point corresponding to the specified behavior and the interactive behavior before the specified behavior in the interactive behavior sequence that is the same as the candidate interest tag associated with the specified behavior to obtain a target weight.

[0131] In some optional embodiments, the second determination subunit is further configured to determine the long-term interest tag of the target object and the first weight of the long-term interest tag among the candidate interest tags based on the target weight of the candidate interest tags; determine the short-term interest tag of the target object and the second weight of the short-term interest tag among the candidate interest tags based on the target weight of the candidate interest tags; and the generation unit is further configured to generate the target interest tag of the target object based on the long-term interest tag, the first weight, the short-term interest tag, and the second weight.

[0132] In some optional embodiments, the generation unit is further configured to generate a target interest tag for the target object based on the short-term interest tag and the second weight in response to determining that the long-term interest tag and the short-term interest tag do not meet the preset association conditions and the second weight is greater than the first weight.

[0133] In some optional embodiments, the generating unit is further configured to generate a target interest tag of the target object based on the current interest tag, the weight of the current interest tag, and the portrait information of the target object.

[0134] In some optional embodiments, the generation unit is further configured to input the current interest tag, the weight of the current interest tag, and the portrait information of the target object into a preset large language model to generate a target interest tag for the target object.

[0135] In some optional embodiments, the second determination module is configured to determine the initial interest tag associated with each interactive behavior in the interactive behavior sequence; in response to determining that the initial interest tag is an open tag, the initial interest tag is mapped to a closed interest tag to obtain candidate interest tags associated with each interactive behavior.

[0136] In the technical solutions disclosed herein, the acquisition, storage, and application of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0137] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0138] Figure 7A schematic block diagram of an example electronic device 700 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0139] like Figure 7 As shown, device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. RAM 703 may also store various programs and data required for the operation of device 700. Computing unit 701, ROM 702, and RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to bus 704.

[0140] Various components in device 700 are connected to I / O interface 705, including an input unit 706, such as a keyboard, mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a magnetic disk, optical disk, etc.; and a communication unit 709, such as a network card, modem, wireless communication transceiver, etc. The communication unit 709 allows device 700 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0141] The computing unit 701 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as the method for determining interest tags. For example, in some embodiments, the method for determining interest tags may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed onto the device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the computing unit 701, one or more steps of the method for determining interest tags described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the method for determining interest tags in any other suitable manner (e.g., via firmware).

[0142] Various embodiments of the systems and techniques described above 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), system-on-chip systems (SOCs), 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 are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0143] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0144] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, 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, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0145] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0146] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0147] A computer system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a host product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosts and virtual private server (VPS) services. Servers can also be classified as distributed system servers or servers integrated with blockchain.

[0148] According to the technical solution of the embodiment of the present disclosure, the accuracy of the generated target interest tags is improved.

[0149] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions provided by this disclosure can be achieved. This is not limited herein.

[0150] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A method for determining an interest tag, comprising: Determine the target object's interactive behavior sequence within the map application; Determining candidate interest tags associated with each interactive behavior in the interactive behavior sequence; Based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence, a target interest tag of the target object is determined.

2. The method according to claim 1, wherein The determining of the target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence includes: Based on the interaction behavior sequence and candidate interest tags associated with each interaction behavior in the interaction behavior sequence, determining a current interest tag of the target object from the candidate interest tags; Based on the current interest tag, a target interest tag of the target object is generated.

3. The method according to claim 2, wherein: The determining, based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence, the current interest tag of the target object from the candidate interest tags includes: Based on the ranking information and categories of the interactive behaviors in the interactive behavior sequence, adjusting the initial weights of the candidate interest tags associated with the interactive behaviors to obtain target weights of the candidate interest tags; Based on the target weights of the candidate interest tags, determining the current interest tag of the target object and the weight of the current interest tag from among the candidate interest tags; and The generating a target interest tag of the target object based on the current interest tag includes: A target interest tag of the target object is generated based on the current interest tag and the weight of the current interest tag.

4. The method according to claim 3, wherein: The categories of the interactive behaviors include: point of interest inquiry, financial consumption, and content consumption.

5. The method according to claim 4, wherein The adjusting the initial weights of the candidate interest tags associated with the interactive behaviors based on the ranking information and categories of the interactive behaviors in the interactive behavior sequence to obtain target weights of the candidate interest tags includes: Based on the ranking information of the interactive behaviors in the interactive behavior sequence, adjusting the initial weights of the candidate interest tags associated with the interactive behaviors to obtain intermediate weights; In response to determining that a specified behavior exists in the interactive behavior sequence, the intermediate weights of the candidate interest tags associated with the specified behavior are adjusted according to the time point corresponding to the specified behavior and the interactive behavior with the same candidate interest tags associated with the specified behavior before the specified behavior in the interactive behavior sequence to obtain the target weight, wherein the specified behavior is a financial consumption interactive behavior.

6. The method according to claim 3, wherein: The number of the interactive behavior sequences is two, namely a long-term behavior sequence and a short-term behavior sequence. The target weight based on the candidate interest tags determines the current interest tag of the target object and the weight of the current interest tag from among the candidate interest tags, including: Based on the target weights of the candidate interest tags, determining the long-term interest tags of the target object and the first weights of the long-term interest tags from the candidate interest tags; Based on the target weights of the candidate interest tags, determining the short-term interest tags of the target object and the second weights of the short-term interest tags from the candidate interest tags; Furthermore, generating a target interest tag of the target object based on the current interest tag and the weight of the current interest tag includes: A target interest tag for the target object is generated based on the long-term interest tag, the first weight, the short-term interest tag, and the second weight.

7. The method according to claim 6, wherein: The generating a target interest tag of the target object based on the long-term interest tag, the first weight, the short-term interest tag, and the second weight includes: In response to determining that the long-term interest tag and the short-term interest tag do not meet the preset association condition and the second weight is greater than the first weight, a target interest tag of the target object is generated according to the short-term interest tag and the second weight.

8. The method according to claim 3, wherein: The generating a target interest tag of the target object based on the current interest tag and the weight of the current interest tag includes: A target interest tag for the target object is generated based on the current interest tag, the weight of the current interest tag, and the portrait information of the target object.

9. The method according to claim 8, wherein The generating a target interest tag of the target object based on the current interest tag, the weight of the current interest tag, and the portrait information of the target object includes: The current interest tag, the weight of the current interest tag, and the portrait information of the target object are input into a preset large language model to generate a target interest tag for the target object.

10. The method according to any one of claims 1 to 9, wherein: Determine candidate interest tags associated with each interaction behavior in the interaction behavior sequence, including: Determining an initial interest tag associated with each interaction behavior in the interaction behavior sequence; In response to determining that the initial interest tag is an open tag, the initial interest tag is mapped to a closed interest tag to obtain candidate interest tags associated with each interactive behavior.

11. An interest tag determination device, comprising: A first determining module is configured to determine an interactive behavior sequence of a target object in a map application; A second determining module is configured to determine a candidate interest tag associated with each interactive behavior in the interactive behavior sequence; The third determining module is configured to determine the target interest tag of the target object based on the interaction behavior sequence and the candidate interest tags associated with each interaction behavior in the interaction behavior sequence.

12. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.

13. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 10.

14. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 10.

Images