Point of interest determination method and apparatus, electronic device, and storage medium

Abstract

The present disclosure provides a point of interest determination method and device, electronic equipment and storage medium, relates to the field of artificial intelligence, in particular to the field of deep learning and intelligent transportation. The specific implementation scheme is that the electronic equipment obtains target demand information of a user for querying scene information. The electronic equipment can determine associated demand information matched with the target demand information according to a mapping relationship including a plurality of first demand information and a plurality of associated demand information, and determine a target POI matched with the associated demand information according to a second mapping relationship including a second demand information and an associated relationship after a plurality of POI information. Finally, the electronic equipment inputs one or more POI corresponding to the target POI information.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and more particularly to the fields of deep learning and intelligent transportation technology, specifically to a method, apparatus, electronic device, and storage medium for determining points of interest. Background Technology

[0002] When users search using electronic maps or search engines, they enter their needs to retrieve points of interest (POIs) that match those needs.

[0003] Typically, electronic devices that host electronic maps or search engines can use tag-based recall to find Points of Interest (POIs) that meet a user's needs. For example, multiple tags can be manually assigned to each POI, and then the user's needs can be matched against these tags. If a user's needs match a particular tag, then the POI corresponding to that tag is the one that meets the user's needs.

[0004] However, due to the limited number of POI tags, tag-based retrieval methods cannot accurately match user needs in some scenarios. For example, if a user inputs "boiled fish," they might want to know where restaurants serve this dish. But with tag-based retrieval, the electronic device might retrieve "how to make boiled fish," rather than restaurants that offer it, thus failing to meet the user's needs. Summary of the Invention

[0005] This disclosure provides a method, apparatus, electronic device, and storage medium for determining points of interest.

[0006] According to a first aspect of this disclosure, a method for determining points of interest is provided, comprising:

[0007] The electronic device acquires the user's target demand information for querying location information. Then, based on a first mapping relationship that includes associations between multiple first demand information items and multiple related demand information items, the electronic device determines the related demand information that matches the target demand information. The multiple first demand information items include the target demand information. Furthermore, based on a second mapping relationship that includes associations between multiple second demand information items and multiple POI information items, the electronic device determines the target POI information that matches the associations, where the multiple second demand information items include the related demand information. Thus, after determining the target POI information that matches the user's demand information, the electronic device can output one or more POIs corresponding to that target POI information.

[0008] According to a second aspect of this disclosure, an apparatus for determining points of interest (POIs) is provided, comprising: an acquisition unit for acquiring user demand information, the demand information being used to query location information; a determination unit for determining, based on a first mapping relationship including associations between a plurality of first demand information and a plurality of associated demand information, associated demand information, the plurality of first demand information including the target demand information; and a determination unit for determining target POI information matching the associations based on a second mapping relationship including associations between a plurality of second demand information and a plurality of POI information, the plurality of second demand information including the associated demand information; and an output unit for outputting one or more POIs corresponding to the target POI information.

[0009] According to a third aspect of this disclosure, an electronic device is provided, comprising:

[0010] At least one processor; and

[0011] A memory that is communicatively connected to at least one processor; wherein,

[0012] The memory stores instructions that can be executed by at least one processor, such that the at least one processor is able to perform any of the methods in the first aspect.

[0013] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, comprising:

[0014] Computer instructions are used to cause the computer to perform any of the methods in the first aspect.

[0015] According to a fifth aspect of this disclosure, a computer program product is provided, comprising:

[0016] A computer program, any one of the methods of a computer program in the first aspect of being executed by a processor.

[0017] The technical solution disclosed herein solves the problem that the POIs retrieved do not meet the user's needs, thereby improving the user experience.

[0018] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0019] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0020] Figure 1 This is a schematic diagram of a method for constructing a semantic mapping relationship according to an embodiment of this disclosure;

[0021] Figure 2 This is a flowchart illustrating a method for determining points of interest according to an embodiment of this disclosure;

[0022] Figure 3 This is a schematic diagram of a search interface of an electronic device according to an embodiment of the present disclosure;

[0023] Figure 4 This is a flowchart illustrating a method for determining points of interest according to an embodiment of this disclosure;

[0024] Figure 5 This is a schematic diagram of another method for determining points of interest according to an embodiment of this disclosure;

[0025] Figure 6 This is an application scenario of a semantic mapping relationship in an embodiment of this disclosure;

[0026] Figure 7 This is a schematic diagram of the structure of a point of interest determination device disclosed herein;

[0027] Figure 8 This is a block diagram of an electronic device for a method of determining points of interest provided in an embodiment of this disclosure. Detailed Implementation

[0028] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0029] Before providing a detailed description of the method for determining points of interest in the embodiments of this disclosure, the application scenarios of the embodiments of this disclosure will be explained first.

[0030] When users use electronic maps or search engines, the information they input can include precise needs (such as searching for a specific location), general needs (such as restaurants, hotels, universities), and chain needs (such as convenience store A). However, in some scenarios, users will perform semantic searches. For example, when a user inputs "boiled fish," their need is to find restaurants that serve this dish. Similarly, when a user inputs "product A," their need is to find stores or convenience stores that sell product A.

[0031] To meet users' information needs, electronic devices containing electronic maps or search engines can perform semantic searches based on users' needs to determine the Points of Interest (POIs) corresponding to those needs.

[0032] In one example, such as Figure 1 As shown, when a user enters their needs through the search box, the electronic device can send that information to the server. Upon receiving the needs information, the server can perform a semantic search to determine the Point of Interest (POI) corresponding to the needs. After determining the POI, the server can send the confirmed POI to the electronic device. The electronic device, upon receiving the POI from the server, can then input the POI.

[0033] In some embodiments, in order to determine the POI corresponding to the demand information, the server can query the POI corresponding to the demand information through methods such as inverted index recall, tag-based recall, and channel page classification.

[0034] Inverted indexing can refer to searching for text within a user's needs and identifying Points of Interest (POIs) that include that text. For example, if the user's needs are "Where to go this weekend?", inverted indexing might retrieve POIs containing "play," such as "xx toy store," which are not relevant to the user's actual needs and are therefore not precise enough.

[0035] Tag-based recall involves matching demand information with tags from multiple Points of Interest (POIs) and displaying the POIs whose tags match the demand information to the user. However, due to the limited number of tags for a POI, this method cannot identify POIs accurately, and the query scope is relatively small.

[0036] The channel page categorization method involves displaying multiple categories of Points of Interest (POIs) to users through a webpage. Examples include categories like restaurants, hotels, and parks. Each category can include multiple POIs. However, this method is cumbersome, requiring users to spend time searching and cannot be linked to user input.

[0037] Therefore, embodiments of this application provide a method for determining points of interest (POIs), used to accurately identify POIs that match a user's needs. After receiving user-inputted needs information, the POI determining device can query one or more POIs that match the user's needs information through multiple mapping relationships. For example, the one or more POIs may have a synonymous relationship, a contextual relationship, or a sequential relationship with the user's needs information. In this way, the electronic device can accurately match the user's needs to the POIs through the association between the needs information and the POI information.

[0038] For example, if a user inputs "boiled fish," the semantic mapping can determine which restaurants offer this dish. Furthermore, the hierarchical relationships within the semantic mapping can be used to identify Points of Interest (POIs) that offer boiled fish. These POIs are more closely matched to the user's needs.

[0039] It should be noted that the electronic devices described in this disclosure are not limited to any particular type. The electronic devices described in this disclosure may be tablet computers, mobile phones, desktop computers, laptops, handheld computers, notebook computers, ultra-mobile personal computers (UMPCs), netbooks, cellular phones, personal digital assistants (PDAs), augmented reality (AR) / virtual reality (VR) devices, in-vehicle devices, etc. This disclosure does not impose any special limitations on the specific form of the electronic device.

[0040] The execution entity of the method for determining points of interest provided in this disclosure can be a determining device, which can be an electronic device or a server. Furthermore, the executing device can also be the central processing unit (CPU) of the electronic device or server, or a control module within the electronic device or server for determining points of interest.

[0041] The method for determining points of interest provided in the embodiments of this disclosure will be described in detail below with reference to the accompanying drawings.

[0042] The method for determining points of interest provided in this disclosure may include: a process of constructing semantic mapping relationships through semantic mapping mining (hereinafter referred to as the "semantic mapping relationship construction process") and a process of determining POIs that match the user's needs information based on the semantic mapping relationships (hereinafter referred to as the "point of interest determination process").

[0043] The following section describes the "semantic mapping relationship construction process".

[0044] like Figure 1 As shown, the "semantic mapping relationship construction process" includes: S101 to S112.

[0045] S101. Obtain multiple sample data.

[0046] The sample data may include demand information where page views (PV) are greater than a first threshold but click counts are less than a second threshold, and demand information where the conversion rate is lower than a preset value. The first threshold is greater than the second threshold, and both thresholds can be set as needed; for example, the first threshold can be 50 and the second threshold can be 40, without restriction.

[0047] PV can refer to the number of times the original demand information was entered, deleted, and replaced with new demand information. Click count can refer to the number of times the POI corresponding to the original demand information was clicked. Conversion rate can be the ratio of the number of clicks on the POI corresponding to the new demand information to the number of clicks on the POI corresponding to the original demand information. The new demand information is related to the original demand information, or the original demand information is a generalized demand, and the new demand information is a specific demand information. For example, the original demand information is "Where to go this weekend?", and the new demand information is "Have a picnic this weekend." Another example is that the original demand information is "What to eat for lunch?", and the new demand information is "Boiled fish."

[0048] In one example, a certain requirement information A is entered 100 times within a preset time period. Of these 100 entries, requirement information A is deleted and requirement information B is entered 90 times (i.e., the page view (PV) of requirement information A is 90). Requirement information A and requirement information B are related. If the POI corresponding to requirement information A is clicked 30 times and the POI corresponding to requirement information B is clicked 60 times, then the click count for requirement information A is 30, resulting in a conversion rate of 60 / 30 = 2. This indicates that users are more interested in the POI corresponding to requirement information B, or in other words, the POI corresponding to requirement information A does not meet the user's actual needs. Based on this, the server can use requirement information A as sample data. Subsequently, based on the technical solution of this embodiment, if a user enters requirement information A again, the server can output the POI corresponding to requirement information B, thus satisfying the user's needs. The specific implementation method of this technical solution can be referred to in the following description, and will not be repeated here.

[0049] In one possible implementation, the server can obtain sample data from one or more data sources. Sample sources may include session logs, search logs, etc. Specifically, session logs and search logs can refer to user query and search behavior on the search platform.

[0050] S102. Obtain multiple sample POIs and determine the POI information corresponding to each sample POI.

[0051] These multiple sample POIs can include POIs related to the user's needs. POI information can include one or more POIs belonging to the same type (or industry). For example, POI types can include zoos, forest parks, farmhouses, fruit picking gardens, etc. Zoo-related POI information can include Zoo A, Zoo B, Zoo C, etc.

[0052] In one possible implementation, the server can obtain multiple sample POIs from data sources such as POIs in the user's query information or user comments. For example, user comments may include POI-related fields or words. After obtaining multiple sample POIs, the server can respond to an annotation operation to determine the POI information corresponding to each sample POI. The annotation operation can refer to labeling each sample POI with its corresponding POI information or the POI's industry.

[0053] It should be noted that in this embodiment, the query can be a series of queries entered by the user. For example, if a user enters a first query but does not click on the POI determined by the first query, and then enters a second query and clicks on the POI determined by the second query, then the first query is the sample data, and the second query is the query data. The POI corresponding to the second query is the sample POI.

[0054] For example, if a user first enters "Where to go this weekend", the returned POIs are: XX Toy Store, etc.; then the user deletes "Where to go this weekend" and re-enters "Outdoor Picnic", the returned POIs are: Park A, Park B, Farmhouse C, etc., and the user clicks on "Park A" and "Farmhouse C" in the POIs, then "Where to go this weekend" can be considered as sample data, "Outdoor Picnic" is a different query, and the sample POIs are "Park A" and "Farmhouse C".

[0055] For example, referring to the example in S101 above, the POI corresponding to demand information B can be the sample POI.

[0056] Furthermore, after obtaining multiple sample POIs, to avoid interference from abnormal data in subsequent model training, these sample POIs can be filtered. For example, POIs with a PV lower than a preset value can be deleted. The preset value can be set as needed, for example, it can be 50. Alternatively, the multiple POIs can be sorted from largest to smallest PV, and a preset number of POIs at the bottom of the sort can be deleted, such as deleting the bottom 50% of the POIs. In this way, higher quality sample POIs can be obtained.

[0057] S103. In response to the association operation, determine the semantic mapping relationship between multiple training sample data and multiple POI information.

[0058] The association operation refers to the process by which developers associate a portion of the sample data (i.e., the training sample data) with multiple POI information. Semantic mapping relationships can include the association between multiple sample data and multiple POI information, the association between multiple sample data, and the association between multiple POI information.

[0059] For example, developers can first select training sample data from multiple sample data sets. For this training sample data set, they can manually label the relationships between the multiple training sample data sets and POI information, the relationships between the multiple training sample data sets, and the relationships between the multiple POI information sets.

[0060] The relationship between training sample data and multiple POI information can include a successor relationship. A successor relationship can refer to the ability of POI information to provide sample data. For example, if the sample data is about an outdoor picnic, then POI information that can provide information about outdoor picnics can include parks and farmhouses. As another example, if the sample data is about feeding alpacas, then POI information that can provide information about feeding alpacas can include zoos.

[0061] The relationships between training sample data can include similarity relationships. A similarity relationship can refer to the similarity between multiple training sample data. For example, if the sample data is about picking strawberries, then sample data similar to picking strawberries could include strawberry picking.

[0062] The relationships between multiple Points of Interest (POIs) can include contextual relationships. A contextual relationship can refer to multiple POIs including parent and child POIs. For example, if sample data A is a park, then the child POIs of sample data A could include forest parks, scenic parks, theme parks, etc., while the parent POIs of sample data A could include scenic areas, leisure areas, etc. As another example, if sample data B is a strawberry picking garden, then the parent POIs of sample data B could include picking gardens, fruit orchards, etc.

[0063] S104. Based on the preset training model, train multiple training sample data and multiple POI information to obtain the initial semantic matching model.

[0064] The preset training model can be the Ernir-GeoL model, or other learning models, such as Natural Language Processing (NLP) algorithms.

[0065] In one example, the server can input multiple training sample data and multiple POI information into a pre-set training model to obtain an initial semantic matching model. The input to the initial semantic matching model can be demand information, and the output can be one or more POI information corresponding to that demand information. In this way, the server can mine demand information with high page views but low conversion rates, and train a semantic matching model based on the demand information and associated POI information to meet the user's demand information.

[0066] Furthermore, to increase the accuracy and applicability of the semantic matching model, the server can process the initial semantic matching model. For example, it can perform sample optimization and model distillation on the initial semantic matching model. The sample optimization process can include:

[0067] S105. Input the test sample data from multiple sample data into the initial semantic matching model to obtain the semantic mapping relationship.

[0068] The semantic mapping relationship can include the associated sample data corresponding to the test sample data, the first sample POI information corresponding to the test sample data, and the second sample POI information corresponding to the first sample POI information.

[0069] The test sample data refers to the sample data other than the training sample data from a pool of sample data. That is, the test sample data is sample data that has not been manually labeled or sample data that does not have associated POI information.

[0070] S106. If the semantic mapping relationship is abnormal, the server responds to the modification operation, modifies the semantic mapping relationship, and obtains the modified semantic mapping relationship.

[0071] Among them, semantic mapping relationship anomalies may include one or more of the following: the second sample data and the associated sample data have no similarity or the similarity is lower than a preset value; the second sample data and the first sample POI information have no successor relationship; and the first sample POI information and the second sample POI information have no hierarchical relationship.

[0072] In one example, if the second sample data is not associated with the first POI information, then the second POI information corresponding to the second sample data is determined.

[0073] In this context, the lack of correlation between the second sample data and the first POI information can also be described as the second sample data being unrelated to the first POI information. For example, if the second sample data is "what to eat for lunch," and inputting "what to eat for lunch" into the initial semantic model yields POI information such as "park" or "zoo," then this indicates that the second sample data and the first POI information are not related.

[0074] In one example, after receiving the POI information output by the initial semantic model, the server can output the second sample data and the first POI information in a correspondence format. For example, the second sample data and the first POI information can be output in the form of a table or array. For instance, the server could output: {Sample Data 1: POI Information 1}, {Sample Data 2: POI Information 2}, {Sample Data 3: POI Information 3}, ..., {Sample Data n: POI Information n}, where n is a positive integer. In this way, developers can determine whether there is a correlation between the sample data and the POI information.

[0075] In one possible implementation, in response to a modification operation, the server can determine the POI information corresponding to the second sample data. The modification operation can refer to modifying the POI information corresponding to the second sample data, and there is a correlation between the second sample data and the modified POI information. For example, the second sample data might be "What to eat for lunch," and the modified POI might be "restaurants," "dining," or similar POI information. In this way, the server can accurately determine the POI information that is related to the second sample data.

[0076] S107. Iteratively train the initial semantic matching model based on the modified semantic mapping relationship to obtain the semantic matching model.

[0077] Iterative training refers to using the modified semantic mapping relationship, including sample data and sample POIs, as training data to continue training the initial semantic matching model until the POI information output by the semantic matching model has a sequential relationship with the sample data, a similarity relationship between the sample data, a hierarchical relationship between the POI information, or the accuracy of the output results reaches a preset threshold. The preset threshold can be set as needed and is not limited. In this way, by actively learning and continuously optimizing the training samples, the accuracy of the model is improved.

[0078] Furthermore, to improve the applicability of the model, model distillation can be performed on the semantic matching model to obtain a processed semantic matching model. Model distillation is used to compress the model size without changing the model's effectiveness and performance. Specific methods for model distillation can be found in existing techniques and will not be elaborated upon here.

[0079] Furthermore, after obtaining the semantic matching model, semantic mapping relationships can be constructed based on this model in practical applications. These semantic mapping relationships can be either knowledge graphs or semantic vectors. Knowledge graphs can be constructed using method one, while semantic vectors can be established using method two.

[0080] Method 1: Construct a knowledge graph based on a semantic matching model; Method 2: Determine semantic vectors based on a semantic matching model. The following explains Method 1 and Method 2.

[0081] Method 1: Construct a knowledge graph based on a semantic matching model.

[0082] In one example, such as Figure 1 As shown, a knowledge graph can be constructed using S108 and S109.

[0083] S108. Based on the semantic model, determine the POI information corresponding to each requirement information among multiple requirement information.

[0084] In one example, the server can input multiple demand information into the preset semantic model to obtain POI information with correlation for each demand information.

[0085] S109. Construct a semantic mapping relationship based on multiple requirement information and the POI information corresponding to each requirement information.

[0086] The semantic mapping relationship can include a first mapping relationship between multiple demand information, a second mapping relationship between multiple demand information and multiple POI information, and a third mapping relationship between multiple POI information.

[0087] The first mapping relationship can include the association between multiple pieces of demand information, which can include synonyms or appositives. For example, picking strawberries and strawberry harvesting are two pieces of demand information that are synonymous.

[0088] The second mapping relationship can include the association between multiple demand information and multiple POI information, and this association can include a successor relationship. For example, POI information that has a successor relationship with outdoor picnics can include parks, farmhouses, etc.

[0089] The third mapping relationship can include the association between multiple POI information, which can include contextual relationships or hierarchical relationships. For example, the subordinate POI information of a park can include forest parks, scenic parks, nature parks, etc.

[0090] In one example, the server can first merge the same POI information among the POI information corresponding to multiple demand information, so that the demand information can be connected through the same POI information.

[0091] For example, demand information 1 is associated with POI information 1 and POI information 2; demand information 2 is associated with POI information 1 and POI information 3; and demand information 3 is associated with POI information 2 and POI information 4. The server can connect demand information 1 and demand information 2 through POI information 2, and connect demand information 2 and demand information 3 through POI information 4.

[0092] Furthermore, the server can determine and annotate the hierarchical relationships between multiple POI information. For example, the server can determine the hierarchical relationships between multiple POI information in response to an annotation operation.

[0093] In this way, the server can quickly and accurately construct semantic mapping relationships based on the relationships between multiple demand information, multiple POI information, and multiple demand information and multiple POI information.

[0094] Method 2: Determine the semantic vector based on the semantic matching model.

[0095] In one example, such as Figure 1 As shown, the semantic vector can be determined through S110 to S111.

[0096] S110. Based on the preset semantic model, determine the POI information corresponding to each requirement information among multiple requirement information.

[0097] S110 can be referred to as S108 above, and will not be repeated here.

[0098] S111. Determine multiple semantic vectors based on the POI information corresponding to each of the multiple requirement information.

[0099] A semantic vector includes one or more of the following: a demand information and its corresponding POI information, a demand information and its corresponding associated demand information, and a POI information and its corresponding associated POI information.

[0100] Furthermore, after obtaining multiple semantic vectors, the method may also include S112 to facilitate application to products.

[0101] S112. Construct a recall model based on multiple semantic vectors.

[0102] This recall model can be used to query POI information. For example, it can be a semantic recall model based on artificial neural networks (ANN).

[0103] In one example, the server can use these multiple semantic vectors as parameters of the recall model, thereby enabling the POI query function of the recall model.

[0104] In another example, the server can use the AN algorithm to train the multiple semantic vectors to obtain the AN semantic recall model.

[0105] In one scenario, after the AN semantic recall model is built, it can be applied to multiple online products (such as electronic maps, search engines, etc.), thereby increasing the applicability of POI queries.

[0106] Thus, because the semantic matching model can identify POI information that is related to the demand information, multiple demand information items and their corresponding POI information can be obtained based on this model. Therefore, based on the relationships between these multiple demand information items and POI information, a semantic mapping relationship that accurately represents the relationship between the multiple demand information items and POI information can be constructed.

[0107] Based on the above technical solution, by mining user demand information, we can obtain demand information with high query frequency but low conversion rate, as well as the POI information associated with this demand information. Thus, the semantic matching model trained based on this demand information and POI information can accurately determine the POI information corresponding to these demand information. In subsequent applications of this semantic matching model, it can avoid situations where users need to modify their demand information to find matching POIs, thereby improving the user experience.

[0108] The process of "determining points of interest" will be described next.

[0109] After constructing the semantic mapping relationship using the above method, the POI information matching the user's needs can be determined based on this semantic mapping relationship. For example... Figure 2 As shown, it includes S201 to S203.

[0110] S201. Obtain the user's target needs information.

[0111] Among them, the user's target demand information can be used to query location information.

[0112] In one possible implementation, the electronic device can respond to the user's input and obtain the user's target demand information.

[0113] In one example, an electronic device can be equipped with a search box for users to input search information. This search information may include desired information. For example, such as... Figure 3As shown, the search information could refer to "places suitable for taking children to play on weekends" in search box 310.

[0114] The search information can be text information entered by the user via keyboard or touchscreen, or semantic information of the user recognized by the electronic device through voice.

[0115] It's important to note that user needs information can include both scenario-based needs information and concrete needs information. Scenario-based needs information refers to needs that characterize the service or purpose a user requires. For example, scenario-based needs information may include fields indicating a service or query. For instance, a user's need might be: "A place suitable for taking the kids out on the weekend." In this need information, "taking the kids out" refers to a service, and "a place to go" refers to a query. That is, this scenario-based needs information does not include specific location information or fields. Concrete needs information refers to needs that clearly define the user's purpose. For instance, concrete needs information may include fields indicating a specific purpose. For example, a user's needs might be: "Picking strawberries, feeding alpacas, outdoor picnics," etc. These needs contain fields indicating the purpose / service.

[0116] In one example, to determine whether a user's requirement information is scenario-based or specific, the electronic device, after acquiring the user's requirement information, can perform semantic recognition to determine whether the requirement includes fields that clearly define the user's purpose / service. For example, the electronic device can be pre-configured with multiple fields representing purposes / services. The electronic device can then match the user's requirement information with these multiple fields. If the requirement information matches a field, it is determined to be a specific requirement; if it does not match a field, it is determined to be a scenario-based requirement.

[0117] When an electronic device determines that a user's requirement information is concrete, the electronic device can use that statement information as the user's requirement information.

[0118] When an electronic device determines that a user's needs are scenario-based, it can transform those statements into concrete requirements.

[0119] In one example, the electronic device can extract scene elements from the requirement information to obtain the scene elements in the statement information.

[0120] For example, when an electronic device is equipped with a scene element extraction model for extracting scene elements, the electronic device can input the statement information into the preset scene element extraction model to obtain the scene elements for the required information.

[0121] For example, when a server is configured with a scene element model, an electronic device can send the required information to the server. After receiving the required information, the server can input the required information into the scene element pre-model, obtain the scene elements required by the requested information, and then send them to the electronic device. In this way, the electronic device can obtain the scene elements required by the requested information.

[0122] Furthermore, after acquiring the scene elements of the demand information, the electronic device can transform this demand information into concrete demand information through element association. In this way, demand information with a generalized description can be transformed into demand information with a specific requirement. Subsequently, based on this specific requirement, POIs that meet the user's needs can be queried more quickly and accurately.

[0123] In one example, feature association may include one or more methods such as manual generation, POI comment information extraction, and log mining.

[0124] Among them, manual production can instruct the annotation of concrete demand information corresponding to scene elements through manual means. For example, when the scene elements are "weekend" and "playing with children", the concrete demand information to be managed can be: feeding alpacas, picking strawberries, and outdoor picnic.

[0125] POI review information extraction refers to retrieving review information corresponding to a POI to obtain POIs that include scene elements, and then using the services that the POI can provide as concrete demand information. For example, if the review information of a POI includes: "This place is suitable for taking children to play," "Suitable for weekend play," "Nice environment, suitable for family play," etc., and the services that the POI can provide include: feeding alpacas and outdoor picnics, then the concrete demand information corresponding to "Where to go for a weekend play" could be: feeding alpacas and outdoor picnics.

[0126] Log mining refers to traversing a user's search logs, querying the Points of Interest (POIs) corresponding to the user's search information, and using the services that the POI can provide as concrete demand information. For example, if a user's search information is "What are some good restaurants nearby?", and then the user's search information is "XX restaurant", and "XX restaurant" can provide dishes such as boiled fish and Mapo tofu, then the concrete demand information corresponding to "What are some good restaurants nearby?" could be: boiled fish, Mapo tofu, etc.

[0127] In one example, when querying POIs using log mining, a conditional random field (CRF) can be used to traverse the log information, which can reduce query time and save manpower compared to manual methods.

[0128] S202. Based on the first mapping relationship, determine the associated requirement information that matches the target requirement information, and based on the second mapping relationship, determine the target requirement information that matches the associated requirement information.

[0129] The first mapping relationship can be as described above. Figure 1 The semantic mapping relationship shown in the embodiment exhibits a similarity relationship. The second mapping relationship can be as described above. Figure 1 The semantic mapping relationship shown in the embodiment represents a succession relationship. This semantic mapping relationship can refer to the knowledge graph mentioned above, or it can be a semantic vector configured in the recall model mentioned above.

[0130] In one scenario, after acquiring a user's target requirement information, the electronic device can send that information to a server. Upon receiving the target requirement information, the server can determine the associated requirement information that matches the user's requirement information based on a first mapping relationship.

[0131] In another scenario, electronic devices can obtain semantic mapping relationships from the server and determine the associated demand information that matches the user's demand information based on the semantic mapping relationships.

[0132] In another scenario, electronic devices can be configured with a recall model. The electronic device can input the user's demand information into the recall model to determine the associated demand information that matches the user's demand information.

[0133] In one example, when the electronic device / server cannot determine the POI information corresponding to the target demand information based on the second mapping relationship, the electronic device / server can traverse multiple first demand information in the first mapping relationship to determine the associated demand information that is the same as or similar to the target demand information. After determining the associated demand information that is the same as or similar to the target demand information, the electronic device / server can query the POI information based on the association between the associated demand information and the POI information. These multiple first demand information include the target demand information.

[0134] In this context, "requirement information similar to target requirement information" can refer to the target requirement information including scenario elements that are the same as or similar to the target requirement information.

[0135] For example, when a user's target demand information is scenario demand information, such as "a place suitable for taking children to play on the weekend", electronic devices / servers can transform this demand information into concrete demand information such as "time (weekend) and purpose (taking children to play)". Then, target demand information similar to this demand information can be "what to do on the weekend", "where to go on Saturday and Sunday", "where are some fun places", etc.

[0136] For example, if a user's requirement is "feed alpacas", then similar target requirements could be "see alpacas", "where can I see alpacas", "where are there alpacas", etc.

[0137] In one scenario, combined Figure 3 ,like Figure 4 As shown, the user's search information is "places suitable for taking children to play on the weekend" in search box 310. The electronic device sends this search information to the server. Upon receiving the search information, the server can extract the elements of the search information to obtain the scenario elements: time (weekend) and purpose (taking children to play). Based on these new scenario elements, the server performs element association to obtain the user's specific needs information.

[0138] After obtaining the user's specific needs information, the server can determine the POI information that matches the specific needs information based on the second mapping relationship.

[0139] In one example, when a user's requirement information is requirement information A, the server can retrieve the succession relationship in the semantic mapping relationship. If a requirement information similar to or the same as requirement information A is found in the succession relationship, the POI information associated with that requirement information can be used as the POI information that matches requirement information A.

[0140] In one example, referring to the example in S202 above, if no similar or identical requirement information to requirement information A is found, the server can query requirement information B that is synonymous with requirement information A based on the synonym relationships between multiple requirement information pieces, and then query whether the successor relationship includes requirement information B. If requirement information B is still found in the successor relationship, the server continues to query requirement information that is synonymous with requirement information B based on the synonym relationships until requirement information included in the successor relationship is found.

[0141] Then, based on the connection relationship, the server can query the POI information B corresponding to the demand information and use the POI information B as the POI information corresponding to the demand information A.

[0142] Furthermore, after retrieving POI information B corresponding to demand information A, if POI information B has no corresponding POI or includes more than a preset number of POIs, the server can query the POI information associated with POI information B based on the hierarchical relationship in the semantic mapping, and use this POI information as the POI information matching demand information A. Thus, the POI information that satisfies the user's demand information can be accurately determined.

[0143] For example, combining Figure 4The semantic mapping relationship indicates that the user's demand information is for places suitable for taking children to play on weekends. This demand information is scenario-based demand information. The specific users corresponding to this demand information include: feeding alpacas, picking strawberries, and outdoor picnics. The server can traverse the semantic mapping relationship and find feeding alpacas and outdoor picnics, but not picking strawberries.

[0144] For feeding alpacas and outdoor picnics, the server determines the POI information corresponding to feeding alpacas as "zoo" based on this semantic mapping relationship, and the POI information corresponding to outdoor picnics includes parks, zoos, and farmhouses. If the server finds the POI in a pre-configured list of multiple POI information (each POI information includes one or more POIs), it can use the POI corresponding to the POI information as the POI that meets the user's needs and return it to the electronic device. For example, if the pre-configured POI information includes zoos and farmhouses, the server can return the POIs included in zoos (e.g., D Zoo) and the POIs included in farmhouses (e.g., C Amusement Park) to the electronic device.

[0145] If the server does not find a POI among the pre-configured POI information, it can check if the POI has parent or child POI information. If so, the server can use one or more POIs included in the parent and / or child POI information as the POI that meets the user's needs. This allows for more precise identification of the POI that meets the user's requirements.

[0146] For example, such as Figure 4 As shown, the pre-configured POI information does not include parks. The server can query the lower-level POI information of parks based on context. For example, if the server can find that the lower-level POI information of a park includes a forest park, the server can use the POI included in the forest park (B Forest Park) as the POI that meets the user's needs.

[0147] For strawberry picking, the server can use semantic mapping to find synonyms between multiple demand information items and query demand information that is synonymous with strawberry picking. For example, in the context of strawberry harvesting, the server can query whether the demand relationship includes strawberry harvesting. Figure 4 As shown, this relationship includes strawberry picking, and the corresponding POI information for strawberry picking is a strawberry picking garden. Since the strawberry picking garden is not among the pre-configured POI information, the server can query the parent POI information of the strawberry picking garden in the hierarchical relationship, such as the picking garden itself. Because the pre-configured POI information includes picking gardens, the server can determine that the POI information matching strawberry picking is a picking garden. Furthermore, the server can determine that the POI included by the picking garden is Strawberry Picking Garden A.

[0148] In this way, the server can obtain POI information matching the user's specific needs (places suitable for taking children to play on weekends) based on semantic mapping relationships. These POIs include: zoos, forest parks, farmhouses, and fruit picking gardens. For example... Figure 4 As shown, the POIs corresponding to the POI information include A Strawberry Picking Garden, B Forest Park, C Amusement Park, and D Zoo.

[0149] It should be noted that in this embodiment, the POI information can be determined based on the POI's basic information. For example, POI information may include zoos, forest parks, farmhouses, fruit picking gardens, etc. The basic information of a POI may include its name, alias, address, etc. For example, the basic information of Strawberry Picking Garden A may include A (name), alias (Strawberry Picking Garden), and address (latitude and longitude); the basic information of Forest Park B may include B (name), alias (Forest Park), and address (latitude and longitude); the basic information of Amusement Park C may include C (name), alias (Farmhouse), and address (latitude and longitude); and the basic information of Zoo D may include D (name), alias (Zoo), and address (latitude and longitude). Wherein, the POI information of Strawberry Picking Garden A is "Fruit Picking Garden," the POI information of Forest Park B is "Forest Park," the POI information of Amusement Park C is "Farmhouse," and the POI information of Zoo D is "Zoo."

[0150] After receiving the Point of Interest (POI) information corresponding to the specific demand information, the server can return one or more POIs included in the POI information to the electronic device. In this way, the electronic device can determine one or more POIs corresponding to the user's demand information.

[0151] S203. Output one or more POIs corresponding to the target POI information.

[0152] In one example, such as Figure 4 As shown, after receiving one or more POIs corresponding to the target POI information from the server, the electronic device can display the one or more POIs in the search results box 320. Each POI can be linked to its corresponding description information. That is, in response to a click on a POI, the electronic device can obtain the description information corresponding to that POI.

[0153] based on Figure 2The technical solution allows electronic devices to determine Points of Interest (POIs) that meet user needs based on semantic mapping relationships after acquiring the user's demand information. Since semantic mapping relationships include the associations between multiple demand information items and the associations between multiple demand information items and multiple POIs, the retrieval scope of demand information can be expanded based on these associations. This avoids the problem of insufficient associations between demand information items and POIs, resulting in the inability to retrieve corresponding POI information. For example, if the user's demand information is not found in the associations between multiple demand information items and multiple POI information, the electronic device can query the associated demand information and continue to query for matching POI information based on that associated demand information. In other words, by using the associations between multiple demand information items and the associations between multiple demand information items and multiple POI information, POI information matching the demand information can be retrieved. For instance, if the user's demand information is "boiled fish," and this demand information is not found in the associations between multiple demand information items and multiple POI information, the electronic device can query associated demand relationships matching "boiled fish," such as restaurants that serve boiled fish, based on the associations between multiple demand information items. Then, the electronic device can search for restaurants that serve boiled fish and determine the specific restaurant (i.e., POI) that serves boiled fish.

[0154] In some embodiments, such as Figure 5 As shown, embodiments of this application may further include S501 and S502.

[0155] S501, Obtain user profile information.

[0156] The user profile information can include user profiles and / or scenario profiles. User profiles can be used to identify user characteristics. For example, a user profile may include the user's mode of transportation, age, and historical POIs (Points of Interest). Scenario profiles may include location information, time information, etc. For example, location information may refer to whether the user is located locally (or their latitude and longitude), and the time of the incident may refer to whether it occurred on a holiday or a weekday.

[0157] In one possible implementation, electronic devices or servers can determine a user's profile information based on the user's registration information or historical behavior.

[0158] The user's registration information can refer to information entered when using applications with search functions or other applications. Historical behavior can include historical search behavior and driving behavior.

[0159] For example, when a user registers, the electronic device can respond to the user's registration action by displaying a registration page. This registration page can include information fields that the user needs to fill in. These information fields may include information related to the user's profile.

[0160] For example, electronic devices or servers can obtain user profile information from user search logs. This could include the mode of transportation a user chooses when planning a route using an electronic map, or the points of interest (POIs) the user frequently searches for.

[0161] S502. Based on the user's profile information, determine the target POI that matches the user's profile information from one or more POIs corresponding to the target POI information.

[0162] Among them, the target POI can be one or more POIs that are related to the user's profile information.

[0163] For example, when user profile information includes a user profile, the electronic device can filter one or more POIs based on the user profile to determine the target POI. For instance, one or more POIs might include: Park A, Forest Park B, Amusement Park C, and Zoo D. User profile information includes public transportation usage. The electronic device can determine the target POI as the one with the fewest transfers or the most bus routes based on the user's location information (such as the location marked as home on an electronic map) and a preset public transportation route plan.

[0164] For example, when a user's profile includes a scene profile, the electronic device can filter one or more POIs based on the scene profile to determine the target POI. For instance, one or more POIs might include: Park A, Forest Park B, Amusement Park C, and Zoo D. The user's scene profile might include having a car and it being Sunday. The electronic device can then determine the target POI based on the road information and pedestrian traffic information of these one or more POIs.

[0165] Specifically, POI road information can refer to historical road information between the user's location and the POI. Historical road information can refer to the traffic congestion between the user's location and the POI on multiple weekends within a historical time period. Pedestrian traffic information can refer to the average number of people per POI during multiple weekends within a historical time period. The target POI can be the POI with the least traffic congestion and / or the least pedestrian traffic among one or more POIs on weekends.

[0166] Furthermore, the aforementioned S203 may specifically include S503.

[0167] S503, Output target POI.

[0168] In one example, after identifying a target POI, the electronic device can mark the target POI (e.g., add a marker to the location of the POI on an electronic map) or adjust the order of the target POI to the top of the list of one or more POIs. This can increase user attention and thus improve the user experience.

[0169] Based on this embodiment, electronic devices / servers can further filter one or more Points of Interest (POIs) that meet the user's needs based on the user's profile information, selecting POIs that match the user's profile information. In this way, by combining the user's needs information and user profile information, POIs that match the user's needs can be retrieved more accurately, improving the user experience.

[0170] In some embodiments, such as Figure 6 The image shows a specific application of a method for determining points of interest provided in this application.

[0171] Figure 6 In this context, the server can construct semantic mapping relationships based on user profiles, scenario profiles, user demand information, and POIs. Based on the storage and retrieval capabilities of these semantic mapping relationships, it can determine a list of POIs that can meet the user's demand information, a recommendation query, and filter tags.

[0172] It should be noted that, Figure 6 In this context, extracting information and semantically mapping demand information can yield tags for that demand information (such as industry tags and attribute tags). Industry tags for demand information can refer to the type of scenario element to which the demand information belongs (e.g., play, eat, learn, etc.), while attribute tags can refer to the purpose of the demand information (e.g., take children out to play, go out to eat, etc.). Similarly, extracting information and classifying POI basic information or POI review information by industry can yield POI tags (such as industry tags and attribute tags). A POI's industry tag can refer to the industry it belongs to (e.g., park, zoo, restaurant, etc.), while its attribute tags can refer to its capabilities (e.g., the types of services it can provide (entertainment, dining, etc.)).

[0173] The POI list can be used for semantic recall on electronic maps, displaying POIs that meet users' needs. Recommendation queries can be used to recommend items to users, reducing their input. Filter tags can be applied to filters.

[0174] Specifically, the products to which the above-mentioned POI list, recommendation query, and filter tags can be applied include list page POI results, list page filters, list page recommendation queries, sug page POI results, and sug page hot words.

[0175] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0176] The foregoing primarily describes the solutions provided by the embodiments of this disclosure from the perspective of computer devices. It is understood that, in order to achieve the above functions, the computer device includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, based on the form recognition method steps described in conjunction with the embodiments disclosed herein, this disclosure can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0177] This disclosure embodiment can divide the method for determining points of interest into functional modules or functional units based on the above method examples. For example, each function can be divided into its own functional modules or functional units, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module or functional unit. The module or unit division in this disclosure embodiment is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used.

[0178] like Figure 7 The diagram shown is a structural schematic of a point-of-interest (POI) determination device provided in an embodiment of this disclosure. The POI determination device may include: an acquisition unit 701, a determination unit 702, and an output unit 703.

[0179] The acquisition unit 701 is used to acquire the user's target demand information, which is used to query location information.

[0180] The determining unit 702 is used to determine the associated demand information that matches the target demand information according to the first mapping relationship. The first mapping relationship includes the association relationship between multiple first demand information and multiple associated demand information, and the multiple first demand information includes the target demand information.

[0181] The determining unit 702 is further configured to determine the target POI information that matches the associated demand information according to the second mapping relationship. The second mapping relationship includes the association relationship between multiple second demand information and multiple POI information. The multiple second demand information includes the associated demand information.

[0182] Output unit 703 is used to output one or more POIs corresponding to the target POI information.

[0183] Optionally, the determining unit 702 is specifically used to: determine the first POI information that matches the associated demand information according to the second mapping relationship; if the first POI information does not have a corresponding POI, determine the target POI information associated with the first POI information according to the third mapping relationship, wherein the third mapping relationship includes the association relationship between multiple POI information, and the multiple POI information includes the first POI information and the target POI information.

[0184] Optionally, the acquisition unit 701 is specifically used for: responding to the user's input operation, acquiring the search information input by the user; when the search information is scene requirement information, inputting the search information into a preset scene element extraction model to obtain the scene elements corresponding to the scene requirement information, and determining the target requirement information based on the scene elements; when the search information is concrete requirement information, using the search information as the target requirement information.

[0185] Optionally, the determining unit 702 is further configured to determine a semantic mapping relationship based on a semantic matching model, the semantic mapping relationship including a first mapping relationship, a second mapping relationship and a third mapping relationship.

[0186] Optionally, the device further includes a processing unit 704 and an acquisition unit 701, which are also used to acquire multiple sample data and multiple sample POI information. The sample data includes multiple demand information where the PV is greater than a first threshold but the number of clicks is less than a second threshold, and the second threshold is less than the first threshold.

[0187] The acquisition unit 701 is also used to acquire the semantic mapping relationship between multiple sample data and multiple sample POI information.

[0188] The processing unit 704 is used to train a semantic matching model based on multiple sample data and multiple sample POI information according to a preset training model.

[0189] Optionally, the semantic matching model is an optimized model, which includes sample optimization and / or model distillation.

[0190] Optionally, the acquisition unit 701 is also used to acquire user profile information, which includes user profile and / or scene profile. The user profile is used to characterize the user's features, and the scene profile includes location information and time information.

[0191] The determining unit 702 is further configured to determine, based on the user's profile information, a target POI that matches the user's profile information from one or more POIs corresponding to the target POI information.

[0192] Output unit 703 is specifically used to output the target POI.

[0193] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0194] Figure 8 A schematic block diagram of an example electronic device 800 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 may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0195] like Figure 8 As shown, device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 802 or a computer program loaded from storage unit 808 into random access memory (RAM) 803. RAM 803 may also store various programs and data required for the operation of device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via bus 804. Input / output (I / O) interface 805 is also connected to bus 804.

[0196] Multiple components in device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of monitors, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0197] The computing unit 801 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as method XXX. For example, in some embodiments, method XXX may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of method XXX described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform method XXX by any other suitable means (e.g., by means of firmware).

[0198] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0199] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0200] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0201] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0202] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0203] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0204] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0205] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for determining points of interest, comprising: Obtain search information entered by the user; The search information includes demand information; If the demand information can be matched with multiple fields representing the purpose or service, then the search information is concrete demand information; if the demand information cannot be matched with any fields, then the search information is scenario demand information. When the search information is the scene requirement information, the search information is input into a preset scene element extraction model to obtain the scene elements corresponding to the scene requirement information, and the target requirement information is determined based on the scene elements. The target demand information is used to query location information; When the search information is the concrete demand information, the search information is taken as the target demand information; Based on the first mapping relationship, associated demand information matching the target demand information is determined. The first mapping relationship includes the association between multiple first demand information and multiple associated demand information, and the multiple first demand information includes the target demand information. Based on the second mapping relationship, target point of interest (POI) information that matches the associated demand information is determined. The second mapping relationship includes the association between multiple second demand information and multiple POI information. The multiple second demand information includes the associated demand information. If the POI information corresponding to the target demand information cannot be determined according to the second mapping relationship, the multiple first demand information in the first mapping relationship are traversed to determine the associated demand information that is the same as or similar to the target demand information, and the target POI information is determined based on the association relationship between the associated demand information that is the same as or similar to the target demand information and the POI information. Output one or more POIs corresponding to the target POI information.

2. The method according to claim 1, wherein, The step of determining the target POI information matching the associated demand information based on the second mapping relationship includes: Based on the second mapping relationship, determine the first POI information that matches the associated demand information; If the first POI information does not have a corresponding POI, the target POI information associated with the first POI information is determined according to the third mapping relationship. The third mapping relationship includes the association relationship between multiple POI information, including the first POI information and the target POI.

3. The method according to claim 2, wherein, The method further includes: Based on the semantic matching model, a semantic mapping relationship is determined, which includes the first mapping relationship, the second mapping relationship, and the third mapping relationship.

4. The method according to claim 3, wherein, The method further includes: Acquire multiple sample data and multiple sample POI information. The sample data includes multiple demand information where the page views (PV) are greater than a first threshold but the number of clicks is less than a second threshold, and the second threshold is less than the first threshold. Obtain the semantic mapping relationship between the multiple sample data and the multiple sample POI information; The semantic matching model is obtained by training the multiple sample data and the multiple sample POI information according to the preset training model.

5. The method according to claim 3 or 4, wherein, The semantic matching model is an optimized model, and the optimization process includes sample optimization and / or model distillation.

6. The method according to any one of claims 1-5, wherein, The method further includes: Obtain the user's profile information, which includes a user profile and / or a scene profile. The user profile is used to characterize the user's features, and the scene profile includes location information and time information. Based on the user's profile information, determine the target POI that matches the profile information from one or more POIs corresponding to the target POI information; The step of outputting one or more POIs corresponding to the target POI information includes: Output the target POI.

7. A device for confirming points of interest, comprising: The acquisition unit is used to acquire search information input by the user; The search information includes demand information; If the demand information can be matched with multiple fields representing the purpose or service, then the search information is concrete demand information; if the demand information cannot be matched with any fields, then the search information is scenario demand information. The determining unit is used to input the search information into a preset scene element extraction model when the search information is the scene requirement information, obtain the scene elements corresponding to the scene requirement information, and determine the target requirement information based on the scene elements. The target demand information is used to query location information; when the search information is the concrete demand information, the search information is used as the target demand information. The determining unit is further configured to determine, according to the first mapping relationship, related requirement information that matches the target requirement information, wherein the first mapping relationship includes the association relationship between multiple first requirement information and multiple related requirement information, and the multiple first requirement information includes the target requirement information; The determining unit is further configured to determine, according to the second mapping relationship, target point of interest (POI) information that matches the associated demand information, wherein the second mapping relationship includes the association relationship between multiple second demand information and multiple POI information, and the multiple second demand information includes the associated demand information; The determining unit is further configured to, when it is impossible to determine the POI information corresponding to the target demand information according to the second mapping relationship, traverse multiple first demand information in the first mapping relationship, determine the associated demand information that is the same as or similar to the target demand information, and determine the target POI information based on the association relationship between the associated demand information that is the same as or similar to the target demand information and the POI information; The output unit is used to output one or more POIs corresponding to the target POI information.

8. The apparatus according to claim 7, wherein, The determining unit is specifically used for: Based on the second mapping relationship, determine the first POI information that matches the associated demand information; Based on the third mapping relationship, the target POI information associated with the first POI information is determined. The third mapping relationship includes the association relationship between multiple POI information, including the first POI information and the target POI.

9. The apparatus according to claim 8, wherein, The determining unit is further configured to: Based on the semantic matching model, a semantic mapping relationship is determined, which includes the first mapping relationship, the second mapping relationship, and the third mapping relationship.

10. The apparatus according to claim 9, wherein, The device also includes a processing unit. The acquisition unit is also used to acquire multiple sample data and multiple sample POI information. The sample data includes multiple demand information where the PV is greater than a first threshold but the number of clicks is less than a second threshold, and the second threshold is less than the first threshold. The determining unit is further configured to obtain the semantic mapping relationship between the plurality of sample data and the plurality of sample POI information; The processing unit is used to train the semantic matching model on the plurality of sample data and the plurality of sample POI information according to a preset training model.

11. The apparatus according to claim 9 or 10, wherein, The semantic matching model is an optimized model, and the optimization process includes sample optimization and / or model distillation.

12. The apparatus according to any one of claims 7-11, wherein, The acquisition unit is further configured to acquire the user's profile information, the profile information including a user profile and / or a scene profile, the user profile being used to characterize the user's features, and the scene profile including location information and time information; The determining unit is further configured to determine, based on the portrait information, a target POI that matches the portrait information from one or more POIs corresponding to the target POI information; The output unit is specifically used to output the target POI.

13. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

15. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-6.

Images