CN117252206A

CN117252206A - Intention recognition method, apparatus, device, medium, and program product

Info

Publication number: CN117252206A
Application number: CN202311100078.6A
Authority: CN
Inventors: 解鹏; 谢端阳; 卫若灿; 刘一岑; 谢世明
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2023-08-29
Filing date: 2023-08-29
Publication date: 2023-12-19

Abstract

The embodiment of the specification discloses an intention recognition method, device, equipment, medium and product. The method is applied to the client, and comprises the following steps: acquiring current end behavior information of a target user; determining a current intention set corresponding to the target user based on the current end behavior information; determining a target current intention corresponding to the target user based on the current intention set, the target end behavior characteristic corresponding to the target user and the target user information corresponding to the target user; matching the recommendation intention corresponding to the server recommendation result with the target current intention to obtain a target matching result; and determining target to-be-displayed supply information corresponding to the target user based on the target matching result.

Description

Intention recognition method, apparatus, device, medium, and program product

Technical Field

The present disclosure relates to the field of computer technology, and in particular, to a method, apparatus, device, medium, and program product for intent recognition.

Background

In recent years, a large internet platform gradually develops into a super traffic portal, and marketing recommendation based on goods/rights and the like becomes an important means for realizing traffic rendering and profit in the internet industry. The multi-dimensional information of the user is collected through a big data technology, the specific marketing recommendation scene information is combined, meanwhile, the platform goods supply is associated, and then the relation among the three is built through machine learning model learning, so that the most suitable goods or rights are recommended to the user at proper time and scene, and further the service conversion of the platform GMV and the like is brought.

The traditional marketing recommendation decision link is mainly completed on a cloud server. The user accesses the client, triggers the client to initiate a request to the server, and after the server recalls the supplied content, the goods which are ordered in front are returned to the client for final display through the marketing recommendation strategy of the server.

Disclosure of Invention

The embodiment of the specification provides an intention recognition method, device, equipment, medium and program product, which can consume and utilize abundant user behavior characteristics on a client under the premise of controllable network resource consumption and protection of user privacy, timely capture real-time intention of the client user, enable a client side to evolve from single rendering display of a server recommendation result to the client to have thinking decision-making capability, improve accuracy of intention recognition when the user interacts with the client, optimize target to be displayed supply information to be displayed to the user, and promote user experience and supply conversion effect of the client. The technical scheme is as follows:

in a first aspect, embodiments of the present disclosure provide an intent recognition method, where the method is applied to a client; the method comprises the following steps:

Acquiring current end behavior information of a target user;

determining a current intention set corresponding to the target user based on the current end behavior information;

determining a target current intention corresponding to the target user based on the current intention set, the target end behavior characteristic corresponding to the target user and the target user information corresponding to the target user;

matching the recommendation intention corresponding to the server recommendation result with the target current intention to obtain a target matching result;

and determining target to-be-displayed supply information corresponding to the target user based on the target matching result.

In a second aspect, embodiments of the present disclosure provide an intent recognition device, where the device is applied to a client; the intention recognition device includes:

the acquisition module is used for acquiring the current end behavior information of the target user;

the first determining module is used for determining a current intention set corresponding to the target user based on the current end behavior information;

the second determining module is used for determining the target current intention corresponding to the target user based on the current intention set, the target end behavior characteristic corresponding to the target user and the target user information corresponding to the target user;

The intention matching module is used for matching the recommendation intention corresponding to the server side recommendation result with the target current intention to obtain a target matching result;

and the third determining module is used for determining target to-be-displayed supply information corresponding to the target user based on the target matching result.

In a third aspect, embodiments of the present disclosure provide an electronic device, including: a processor and a memory;

the processor is connected with the memory;

the memory is used for storing executable program codes;

the processor executes a program corresponding to the executable program code stored in the memory by reading the executable program code for performing the method provided by the first aspect of the embodiments of the present specification or any one of the possible implementations of the first aspect.

In a fourth aspect, embodiments of the present specification provide a computer storage medium having stored thereon a plurality of instructions adapted to be loaded by a processor and to carry out the method provided by the first aspect of embodiments of the present specification or any one of the possible implementations of the first aspect.

In a fifth aspect, embodiments of the present description provide a computer program product comprising instructions which, when run on a computer or a processor, cause the computer or the processor to perform the method provided by the first aspect of embodiments of the present description or any one of the possible implementations of the first aspect.

In the embodiment of the specification, on one hand, by determining the current intention set corresponding to the target user based on the current end behavior information of the target user and determining the corresponding target current intention on the client side based on the current intention set, the target end behavior characteristics corresponding to the target user and the target user information, the rich target end behavior characteristics on the client side can be consumed and utilized on the premise that the network resource consumption is controllable (i.e. the request is not required to be sent to the server side in real time according to the current end behavior information) and the privacy of the user is protected (i.e. the intention recognition on the client side can be realized without uploading the user information to the server side), the real-time behaviors of the client user and the real-time calculation cognitive user are better perceived, and the real-time intention (target current intention) of the client user is captured in time; on the other hand, by matching the recommendation intention corresponding to the server recommendation result with the target current intention identified by the client, and determining the corresponding target to-be-displayed supply information according to the matching result, the client side is enabled to evolve from rendering display of a single server recommendation result to have thinking decision-making capability on the client, so that when a user interacts with the client, the supply information to be displayed on the client is decided in real time according to the target current intention identified on the client, intervention is completed on the server recommendation result cached by the client, accuracy of intention identification is improved, the supply information to be displayed to the user is optimized, and user experience and supply conversion effect of the client are improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present description, the drawings that are required in the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present description, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of an intent recognition system according to an exemplary embodiment of the present disclosure;

FIG. 2 is a flow chart of an intent recognition method according to an exemplary embodiment of the present disclosure;

FIG. 3 is a schematic diagram of the effect of intent recognition triggering provided in an exemplary embodiment of the present disclosure;

FIG. 4 is a schematic diagram of an implementation flow for determining a current intent set according to an exemplary embodiment of the present disclosure;

FIG. 5 is a schematic diagram of an implementation process for determining a current intent set according to an exemplary embodiment of the present disclosure;

FIG. 6 is a schematic diagram of another implementation of the current intent set provided in an exemplary embodiment of the present disclosure;

FIG. 7 is a schematic diagram of an implementation flow for determining a current intention of a target according to an exemplary embodiment of the present disclosure;

FIG. 8 is a schematic diagram of an implementation process for determining target offer information to be presented according to an exemplary embodiment of the present disclosure;

FIG. 9 is a schematic diagram of another implementation process for determining target offer information to be presented according to an exemplary embodiment of the present disclosure;

FIG. 10 is a schematic diagram of another implementation process for determining target offer information to be presented according to an exemplary embodiment of the present disclosure;

FIG. 11 is a flowchart of another method for identifying intent provided in an exemplary embodiment of the present disclosure;

FIG. 12 is a schematic diagram of an implementation process of an intent recognition method according to an exemplary embodiment of the present disclosure;

fig. 13 is a schematic structural view of an intention recognition device according to an exemplary embodiment of the present disclosure;

fig. 14 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present disclosure.

Detailed Description

The technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification.

The terms first, second, third and the like in the description and in the claims and in the above drawings, are used for distinguishing between different objects and not necessarily for describing a particular sequential or chronological order. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.

It should be noted that, information (including but not limited to user equipment information, user personal information, etc.), data (including but not limited to data for analysis, stored data, presented data, etc.), and signals according to the embodiments of the present disclosure are all authorized by the user or are fully authorized by the parties, and the collection, use, and processing of relevant data is required to comply with relevant laws and regulations and standards of relevant countries and regions. For example, the present-end behavior information, the target user information, and the like referred to in this specification are all acquired with sufficient authorization.

Reference is next made to fig. 1, which is a schematic diagram illustrating an architecture of an intent recognition system according to an exemplary embodiment of the present disclosure. As shown in fig. 1, the intention recognition system includes: a client 110 and a server 120.

Wherein:

client 110 may include one or more clients. User version software may be installed in the client 110 to implement functions of recommending information to the user after being triggered by the user for the user to browse information desired by the user or purchasing products in the information recommended by the client. When the target user inputs the target operations such as calling, clicking, screen cutting, staying, sliding and the like on the client 110 in the current or current preset time period, the client 110 can acquire the current end behavior information of the target user. The current end behavior information includes real-time interaction behavior (current end behavior) of the target user on the client 110 within the current or current preset time period, such as, but not limited to, end-call, click, screen-cutting, stay, slide, etc. behavior occurring within the current or current preset time period, and current offer information related to the real-time interaction behavior (current end behavior), such as, but not limited to, information including products, services, rights, content, etc. related to the click behavior. After the client 110 obtains the current end behavior information of the target user, it determines a current intention set corresponding to the target user based on the current end behavior information, determines a target current intention corresponding to the target user based on the current intention set, the target end behavior feature corresponding to the target user and the target user information, and finally matches the recommendation intention corresponding to the server recommendation result with the target current intention, and determines the target to-be-displayed supply information corresponding to the client 110 based on the target matching result.

Optionally, the client 110 may further display the determined target to-be-displayed offer information on a to-be-displayed page corresponding to the obtained current end behavior information of the target user.

Further, the client 110 may establish a data relationship with the network, and establish a data connection relationship with the server 120 through the network, for example, send a to-be-recommended supply request, a current intention of a to-be-recommended target, etc., receive and cache a server recommendation result corresponding to the to-be-recommended supply request, receive second supply information corresponding to the current intention of the to-be-recommended target issued by the server, etc. The above-mentioned to-be-recommended offer request may be, but not limited to, a request triggered to be sent to the server 120 when the client 110 acquires the historical terminal behavior information that is the same as or similar to the current terminal behavior information before acquiring the current terminal behavior information of the target user, or a request triggered to be sent to the server 120 when the client 110 acquires the current terminal behavior information of the target user, or the like. The target current intention to be recommended is used for representing the target current intention which is not matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention corresponding to the target user.

It should be understood that any one of the clients 110 may be, but not limited to, a mobile phone, a tablet computer, a notebook computer, etc. with user software installed.

The server 120 may be a server capable of providing various intention recognition and supply recommendation, and may receive, through a network, data such as a supply request to be recommended, a current intention of a target to be recommended, and the like sent by any one user side in the client 110, recognize the intention of the target user based on historical end behavior information or current end behavior information carried in the supply request to be recommended, target base information of the target user, and the like, and determine, for each display position in at least one display position included in a page to be displayed corresponding to the historical end behavior information or the current end behavior information, first supply information to be recommended, that is, a server recommendation result. In addition, if the server 120 receives the current intention of the target to be recommended sent by the client 110, the server may recall the corresponding second supply information from the cloud supply pool based on the current intention of the target to be recommended, and send the second supply information to the client 110, so that the client 110 can determine the corresponding target to be displayed supply information based on the second supply information. The cloud supply pool includes a plurality of supply information. It is to be appreciated that the server may be, but is not limited to being, a hardware server, a virtual server, a cloud server, or the like.

The network may be a medium that provides a communication link between any one of the data sources in the data source cluster and the intent recognition distributed service cluster or between the user side and the intent recognition distributed service cluster, or may be the internet including network devices and transmission media, but is not limited thereto. The transmission medium may be a wired link, such as, but not limited to, coaxial cable, fiber optic and digital subscriber lines (digital subscriber line, DSL), etc., or a wireless link, such as, but not limited to, wireless internet (wireless fidelity, WIFI), bluetooth, a mobile device network, etc.

It will be appreciated that the number of clients 110 and servers 120 in the intent recognition system shown in fig. 1 is by way of example only, and that any number of clients 110 and servers 120 may be included in the intent recognition system in a particular implementation, as embodiments of the present description are not specifically limited in this regard. For example, and without limitation, client 110 may be a client cluster of multiple clients and server 120 may be a server cluster of multiple servers.

Next, taking the client execution intention recognition as an example, the intention recognition method provided in the embodiment of the present specification will be described with reference to fig. 1. Referring specifically to fig. 2, a flowchart of an intent recognition method according to an exemplary embodiment of the present disclosure is shown. As shown in fig. 2, the intention recognition method includes the steps of:

S202, obtaining current end behavior information of a target user.

Specifically, after a decision point in user version software installed on a client is triggered by a target user, the client receives current end behavior information corresponding to the target user. The decision point is used for representing a behavior capable of triggering the client to display the supply information, and can include, but is not limited to, calling end, clicking, screen cutting, staying, sliding and the like corresponding to the user software. Triggering the decision point by the target user corresponds to the behavior of the target user, which the decision point refers to, occurring on the client. The calling terminal refers to opening user edition software through a certain means under certain scenes and loading out content which is expected to be displayed to a target user. The above-mentioned current end behavior information may include, but is not limited to, real-time interactive behavior (current end behavior) generated on the client 110 within a current or current preset period of time of the target user, current offer information related to the real-time interactive behavior (current end behavior), and the like. The real-time interactive behavior (current end behavior) includes, for example and without limitation, behaviors such as calling end, clicking, screen cutting, staying, sliding, etc., and the current supply information includes, for example and without limitation, at least one of information such as products, services, rights, contents, etc. related to the real-time interactive behavior (current end behavior). The current preset time period may be a time period corresponding to a first time to a second time, where the second time is a time corresponding to when the decision point is triggered by the target user, and the first time may be, but is not limited to, a time corresponding to 1 minute, 5 minutes, 1 hour, or the like before the second time.

It can be understood that the intention recognition method in the embodiment of the present disclosure may not only perform intention recognition for the search type of interaction behavior, but also perform intention recognition for various interaction behaviors such as clicking, staying time, searching, sliding out and sliding in, that is, the application interaction scenario of the intention recognition method provided in the embodiment of the present disclosure is richer and has stronger applicability.

As can be appreciated, in order to more accurately identify the terminal intention of the target user with respect to the current terminal behavior of the target user on the client, the current terminal behavior information may only include the interaction behavior of the target user currently generated on the client 110, the current supply information related to the interaction behavior, and so on; for example, the current intention of the target user is considered more fully, that is, the current end intention of the target user on the client is recognized more fully, and the current end intention information may include the interactive actions generated by the target user on the client in the current preset time period, and the current supply information related to the interactive actions (that is, the supply information related to all the interactive actions on the client in the current preset time period, etc.).

Illustratively, as shown in fig. 3 (left diagram), if the target user performs an end-call behavior (current end-action) on the target application 310 on the client, for example, but not limited to, clicking on a presentation area corresponding to the target application 310, and the like. After acquiring the current end behavior information corresponding to the target user, the client performs corresponding intention recognition, and determines that the current end behavior information corresponds to the supply information to be displayed, i.e. the target supply information to be displayed, of each display position 321 in the page to be displayed (e.g. right diagram of fig. 3).

S204, determining a current intention set corresponding to the target user based on the current end behavior information.

Specifically, as shown in fig. 4, the implementation procedure of S204 may include the following steps:

s402, extracting corresponding current supply information from the current end behavior information.

Specifically, the current offer information is used to characterize recommendation information associated with the current end behavior of the target user in the current end behavior information, for example, but not limited to, at least one of products, services, rights, content, and the like related to the current end behavior.

For example, if the current end behavior of the target user is to click on the display position corresponding to the recommended product a in the user version software, the current supply information may be product information corresponding to the product a. The display may be, but is not limited to, a display area or display control, etc.

S404, determining a current intention set corresponding to the target user based on the current supply information.

Specifically, the current intention set includes at least one current intention, and each current intention in the at least one current intention has a corresponding relationship with information in the current supply information. The current intention is the intention corresponding to the current end behavior information of the target user, namely the purpose which the target user hopes to achieve when performing the current end behavior.

According to the method and the device for determining the intention, the intention set which is possibly present at present of the target user can be directly determined on the client according to the current supply information corresponding to the current end behavior information of the target user without returning data or requests to the server, the server does not need to conduct intention recognition according to the behaviors of the client in real time, the pressure of the server is reduced, the information of the target user is not required to be uploaded to the server, the privacy of the target user is protected, the effectiveness and instantaneity of the intention recognition are guaranteed, and the accuracy of the intention recognition is improved.

Optionally, the current supply information is correspondingly marked with one or more attributes, and the attributes may characterize the product, service, rights, content and other information included in the current supply information. As shown in fig. 5, the implementation process of determining the current intention set corresponding to the target user based on the current supply information in S404 may include: extracting at least one corresponding intention label from the attribute of the current supply information, for example, but not limited to extracting all attribute labels or attribute labels marked as intention in advance in the attribute of the current supply information to obtain at least one corresponding intention label; the current intent set corresponding to the target user is then determined based on the at least one intent tag, for example and without limitation, at least one intent corresponding to the at least one intent tag may be determined directly as the current intent, resulting in a current intent set including the at least one current intent.

For example, if the supply information includes information of the product a, and the product a corresponds to the 3 attribute tags of the source a, the seller b, and the product type c, the 3 attribute tags may be directly extracted to be directly used as the intention tags, and it is determined that the current intention set corresponding to the target user includes 3 current intentions of the source a, the seller b, and the product type c, which respectively indicate that the target user currently wants to view or purchase the product corresponding to the source a, the seller b, or the product type c. If the content included in the current supply information is the ancient poem B, the attribute corresponding to the ancient poem B has the author d and the theme "homeland", 2 intention labels of the author d and the theme "homeland" can be directly extracted from the attribute of the ancient poem B, and it is determined that the current intention set corresponding to the target user includes 2 current intentions of the author d and the theme "homeland", which means that the target user currently wants to refer to the poem set of the author d or the poem set of the theme "homeland".

According to the embodiment of the specification, the corresponding at least one intention label is directly extracted from the attribute of the current supply information, so that the current intention set corresponding to the target user is determined, and the determination of the current intention set of the target user can be simply and quickly realized according to the current supply information without introducing additional information, so that the applicability is strong, the realization cost is low, and the efficiency of client intention recognition can be improved to a certain extent.

Alternatively, the client may be preset with a correspondence relationship between the supply information and the intention, that is, a target mapping relationship, for example, but not limited to, a target mapping table or a target mapping function between the supply information and the intention, and the like. As shown in fig. 5, the implementation process of determining the current intent set corresponding to the target user based on the current supply information in S404 may also include: the current supply information is mapped according to the target mapping relation between the supply information and the intention to obtain a current intention set corresponding to the target user, so that the determination of the current intention set of the target user is realized according to the current supply information by presetting the target mapping relation between the supply information and the intention, the effectiveness of each current intention in the identified current intention set is ensured, and the accuracy of intention identification is improved.

It is to be understood that the correspondence between the supply information and the intention may be a one-to-one relationship or a one-to-many relationship, that is, one supply information may have one or more intentions corresponding to the supply information, which is not limited in the embodiment of the present specification.

Optionally, the client may have a current cloud intention set corresponding to the target user cached in advance, where the current cloud intention set includes at least one current cloud intention. The current cloud intention set is determined by the server based on the historical end behavior information of the client in a historical time period before the current time corresponding to the current end behavior, and the historical time period can be, but is not limited to, a week or a month before the current time. As shown in fig. 6, the implementation process of determining the current intent set corresponding to the target user based on the current supply information in S404 may also include: firstly, determining a current end intention set corresponding to a target user based on current supply information, wherein the implementation process of determining the current end intention set corresponding to the target user is, for example, but not limited to, any one of two implementation processes similar to that shown in fig. 5, and is not repeated here; and then, carrying out fusion recall on the current end intention set and the current cloud intention set corresponding to the target user cached in advance by the client, namely, aggregating each current end intention in the current end intention set and each current cloud intention in the current cloud intention set to obtain the current intention set corresponding to the target user. The current intention set may include, but is not limited to, current intention corresponding to various types of current behaviors including calling, clicking, screen cutting, staying, sliding, and the like, where the current intention set is a union of a current intention set and a current cloud intention set.

In the embodiment of the specification, the current intention set of the target user is determined by combining the current cloud intention set corresponding to the historical end behavior information in the historical time period cached in advance on the client and the current end intention set of the target user identified at the current moment of the client, so that the comprehensiveness of the current intention set is determined, and the accuracy of intention identification is improved.

Referring next to fig. 2, as shown in fig. 2, after determining the current intent set corresponding to the target user based on the current end behavior information in S204, the intent recognition method further includes:

s206, determining the target current intention corresponding to the target user based on the current intention set, the target end behavior characteristic corresponding to the target user and the target user information corresponding to the target user.

Specifically, the target behavior feature may be, but is not limited to, a feature corresponding to the current behavior information or a feature corresponding to the historical terminal behavior information in a historical time period before the current time corresponding to the current behavior. The target user information includes target basic information and target sensitive information of a target user on a client, the target basic information may include, but is not limited to, age, male and female, occupation, native place and the like, the target sensitive information may include, but is not limited to, sensitive information related to a target offer corresponding to a target end behavior feature, for example, when the target offer is a financial product, the target sensitive information may include sensitive information such as balance, total amount of financial, and details of financial holding.

It can be understood that in the intention recognition method provided by the embodiment of the specification, the target sensitive information of the target user is not sent to the server, and only the intention recognition of the client side is participated in on the used client, so that the problem that the target sensitive information is easy to leak in the network transmission process is avoided, the effects of more strongly protecting the user privacy and realizing more effective and accurate intention recognition by richer target user information are achieved.

Specifically, as shown in fig. 7, the implementation process of determining the target current intent corresponding to the target user based on the current intent set, the target end behavior feature corresponding to the target user, and the target user information corresponding to the target user in S206 may include:

s702, inputting the current intention set, the target end behavior characteristics corresponding to the target user and the target user information corresponding to the target user into an intention recognition model, and outputting an intention recognition result corresponding to the target user.

Specifically, the intention recognition model is obtained by training based on terminal behavior information, terminal behavior logs and user information corresponding to each of the plurality of users. Before the intention recognition, the server may be used to train based on the corresponding terminal behavior information, terminal behavior log and user information of the multiple users, and the intention recognition model may be obtained by using the relation between the user information, the behaviors in the terminal behavior information, the supplies related to the behaviors and the corresponding intentions of the supplies as the target of the behaviors performed by the users recorded in the terminal behavior log. And then compressing and converting the trained intention recognition model, and transmitting the intention recognition model to a client so that the intention recognition model can be deployed on the client. The terminal behavior information may include, but is not limited to, behaviors such as calling, clicking, screen cutting, staying, sliding, etc. of the user for a period of time, and supply information corresponding to each behavior, and intentions corresponding to each supply information. In the above-mentioned end behavior log, in a period of time corresponding to the end behavior information, behaviors generated by the user on each piece of supply information corresponding to the end behavior information are recorded, for example, but not limited to, whether the user clicks a piece of supply information, if so, the corresponding behavior is recorded as 1 in the end behavior log, and if not, the corresponding behavior is recorded as 0 in the end behavior log.

It can be appreciated that, in order to ensure accuracy and effectiveness of intent recognition, the target end behavior feature should be an end behavior feature associated with an intent recognition model, such as, but not limited to, a feature corresponding to a click behavior applied by the target end behavior feature when the intent recognition model is an intent recognition model for a click behavior; when the intention recognition model is an intention recognition model for click behavior and stay behavior, the target end behavior feature applies a feature corresponding to the click behavior or stay behavior, and the like.

S704, determining the target current intention corresponding to the target user based on the intention recognition result.

Specifically, the target current intention may be, but is not limited to, a current intention of which the bias degree of the target user in the current intention set is highest or greater than a preset threshold.

Alternatively, the intent recognition result may include a target current intent that the intent recognition model recognizes from the current intent set based on target-end behavioral characteristics corresponding to the target user and target user information corresponding to the target user. In S704, the implementation process of determining the target current intention corresponding to the target user based on the intention recognition result may be to directly use the target current intention in the intention recognition result as the target current intention corresponding to the target user.

Optionally, the intention recognition result may include a target score corresponding to each current intention in the current intention set, where the target score is used to characterize a bias degree of the target user to the current intention. The implementation process of determining the target current intention corresponding to the target user based on the intention recognition result in S704 may include: determining a target current intention corresponding to a target user based on the current intention set and target scores corresponding to all the current intentions in the current intention set, wherein the target current intention comprises N current intentions with highest corresponding target scores in the current intention set or current intentions with corresponding target scores larger than preset scores in the current intention set. The above N is a positive integer, for example, but not limited to, 1, 2, 3, etc. The value of N should be less than or equal to the total number of current intents included in the current intent set.

It can be understood that, in order to achieve more accurate intention recognition, and further determine the most desired target to be presented supply information of the target user according to the current end behavior of the target user, the N may be set to 1, that is, only the current intention with the highest corresponding target score in the current intention set is determined as the target current intention, and only the supply information to be presented to the target user is truly and finally determined according to the target current intention with the highest corresponding target score in the current intention set; in order to achieve accurate and comprehensive intention recognition, and further achieve determination of supply information corresponding to multiple target current intentions for a target user to select according to the current end behaviors of the target user, the N can be set to be 2, 3 and the like which are larger than 1, and the total value of the current intentions included in the current intention set is smaller, namely, the 2 or 3 current intentions with the highest corresponding target scores in the current intention set are determined to be target current intentions, further achieve the purpose of displaying more comprehensive supply information for the target user, avoid the situation that supply information corresponding to a single target current intention possibly cannot accept the current requirements of the target user, and improve the conversion rate of supply information display after intention recognition.

Next, please refer to fig. 2, as shown in fig. 2, after determining the target current intention corresponding to the target user based on the current intention set, the target end behavior feature corresponding to the target user, and the target user information corresponding to the target user in S206, the intention recognition method further includes:

and S208, matching the recommendation intention corresponding to the server recommendation result with the target current intention to obtain a target matching result.

Specifically, the server side recommendation result may include a plurality of pieces of offer information that are determined to be recommended to the client by the server side based on historical end behavior information and/or current end behavior information carried in an offer request to be recommended of the client and target base information of the target user. The historical terminal behavior information is terminal behavior information which is the same as or similar to the current terminal behavior information and occurs before the current terminal behavior information of the target user. The recommendation intent corresponding to the server-side recommendation result is determined in a similar manner in S404, and will not be described here again.

Further, the page to be displayed corresponding to the current end behavior information comprises at least one display position, and the page to be displayed comprises a page to be displayed by the client after the current end behavior information occurs. The server side recommending result comprises first supply information which is issued by the server side and is recommended corresponding to each display position in at least one display position. The first supply information is used for representing information recommended by the service end based on target basic information and historical end behavior information corresponding to the target user, and the target end behavior information comprises historical end behavior information and/or current end behavior information of the target user. The step S208 is to match the recommendation intention corresponding to the server recommendation result with the target current intention, and before obtaining the target matching result, the client may determine the recommendation intention corresponding to the server recommendation result based on the first supply information corresponding to each display position in the at least one display in a similar manner in step S404, but not limited to.

It will be appreciated that, in order to effectively control network bandwidth resource consumption, the server recommendation result may be determined based on historical terminal behavior information that is the same as or similar to the current terminal behavior information and that occurs before the current terminal behavior information, and cached in advance in the client. At this time, the client side does not send any request to the server side immediately after the current side behavior information is generated, and whether the last server side recommendation result of the server side can accept the target current intention identified by the client side or not can be judged, namely, the recommendation intention corresponding to the server side recommendation result is matched with the target current intention, so that the pressure of the server side is reduced, and the compliance privacy problem is met. In order to achieve more accurate real-time intention recognition, the server side recommendation result can also be determined directly based on the current side behavior information, namely, after the current side behavior information is generated on the client side, a corresponding supply recommendation request is triggered to be sent to the server side, so that the server side recommendation result issued by the server side based on the current side behavior information carried by the supply recommendation request and target basic information of the target user is obtained. Because there is a certain difference between the target basic information of the target user stored in the server and the target user information of the target user stored on the client, for example, but not limited to, the target sensitive information of the target user is not involved in the target basic information, and therefore, there may be a mismatch between the recommendation intention corresponding to the server recommendation result issued by the server and the target current intention identified by the client, so that it is required to determine whether the server recommendation result can accept the target current intention identified by the client to further ensure accuracy of intention identification and effectiveness and conversion rate of the target to be presented with the supply information finally to be presented to the target user.

It may be understood that the number of recommendation intents corresponding to the server side recommendation result may include one or more, which is not limited in this embodiment of the present disclosure. The matching of the recommendation intention corresponding to the server side recommendation result with the target current intention is to match all recommendation intents corresponding to the server side recommendation result with all target current intents corresponding to the target user one by one, so as to determine whether the server side recommendation result can accept all target current intents of the main target user.

Next, please refer to fig. 2, as shown in fig. 2, in S208, the recommendation intention corresponding to the server recommendation result is matched with the target current intention, and after the target matching result is obtained, the intention recognition method further includes:

s210, determining target to-be-displayed supply information corresponding to the target user based on the target matching result.

Specifically, the target to-be-displayed supply information includes supply information to be displayed in a page to be displayed corresponding to the client after the current end behavior information occurs.

Optionally, the number of the target current intents is at least one. The target matching result includes a target current intention to be recommended, where the target current intention to be recommended is used to characterize a target current intention of at least one target current intention, where the target current intention is not matched with a recommendation intention corresponding to the server side recommendation result, and the mismatch may include, but is not limited to, that the recommendation intention is inconsistent with the target current intention or a similarity between the recommendation intention and the target current intention is less than a preset similarity, that is, the target current intention to be recommended may be understood as a target current intention that the server side recommendation result cannot accept at all. As shown in fig. 8, the implementation process of determining the target to-be-presented supply information corresponding to the target user based on the target matching result in S210 may include: when all of at least one target current intention identified by the client is not matched with each recommendation intention in at least one recommendation intention corresponding to the server recommendation result, the server recommendation result can be considered to completely accept the target current intention of the target user, the target current intention to be recommended can be uploaded to the server, so that the server recalls corresponding second supply information from the cloud supply pool based on the target current intention to be recommended, receives the second supply information issued by the server, and determines target to-be-displayed supply information corresponding to the target user based on the second supply information, for example, but not limited to, directly determining the second supply information as the target to-be-displayed supply information corresponding to the target user. The cloud supply pool includes a plurality of supply information.

Optionally, the number of the target current intents is at least one. The target matching result may further include a recommended target current intention, in addition to the target current intention to be recommended, where the recommended target current intention is used to characterize a target current intention that matches a recommended intention corresponding to the server side recommendation result in the at least one target current intention, that is, the recommended target current intention may be understood as a target current intention that the server side recommendation result has accepted. As shown in fig. 9, in S210, the implementation process of determining the target to-be-presented supply information corresponding to the target user based on the target matching result may also include: when at least one target current intention identified by the client side is partially matched with a certain recommendation intention corresponding to at least one recommendation intention corresponding to the server side recommendation result respectively, and all recommendation intentions in at least one recommendation intention corresponding to the server side recommendation result are not matched with the partial target current intention, the condition that the target current intention of the target user cannot be accepted by the part of the server side recommendation result can be considered, the target current intention to be recommended can be uploaded to the server side, so that the server side recalls corresponding second supply information from the cloud supply pool based on the target current intention to be recommended, receives the second supply information issued by the server side, and determines target to-be-displayed supply information corresponding to the target user based on the first supply information corresponding to the target recommendation intention in the server side recommendation result and the second supply information, for example, but not limited to aggregate the first supply information corresponding to the target recommendation intention in the server side recommendation result and the second supply information, so as to obtain corresponding target to-be-displayed supply information. The target recommendation intention is used for representing the recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server side recommendation result.

Optionally, the number of the target current intents is at least one. The target matching result comprises a recommended target current intention, and the recommended target current intention is used for representing a target current intention matched with a recommendation intention corresponding to the server side recommendation result in at least one target current intention. As shown in fig. 10, the implementation process of determining the target to-be-presented supply information corresponding to the target user based on the target matching result in S210 may also include: when the at least one target current intention identified by the client can find the matched recommendation intention in the at least one recommendation intention corresponding to the server recommendation result, the server recommendation result can be considered to be completely capable of accepting the target current intention of the target user, and then the target to-be-displayed supply information corresponding to the target user can be determined directly based on the first supply information corresponding to the target recommendation intention in the server recommendation result, for example, but not limited to, the first supply information corresponding to the target recommendation intention in the server recommendation result is determined directly as the target to-be-displayed supply information. The target recommendation intention is used for representing the recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server side recommendation result.

Reference is next made to fig. 11, which is a schematic flow chart of another method for identifying intent according to an exemplary embodiment of the present disclosure. As shown in fig. 11, the intention recognition method includes the steps of:

s1102, obtaining the current end behavior information of the target user.

Specifically, S1102 is identical to S202, and will not be described here again.

S1104, determining a current intention set corresponding to the target user based on the current end behavior information.

Specifically, S1104 corresponds to S204, and will not be described here.

S1106, determining the target current intention corresponding to the target user based on the current intention set, the target end behavior characteristic corresponding to the target user and the target user information corresponding to the target user.

Specifically, S1106 corresponds to S206, and will not be described here.

S1108, matching the recommendation intention corresponding to the server recommendation result with the target current intention to obtain a target matching result.

Specifically, S1108 corresponds to S208, and will not be described herein.

S1110, determining target to-be-displayed supply information corresponding to the target user based on the target matching result.

Specifically, S1110 corresponds to S210, and will not be described herein.

S1112, rearranging the plurality of target supply information according to a preset rule to obtain a target rearranging result corresponding to the plurality of target supply information.

Specifically, the target supply information to be displayed includes a plurality of target supply information to be displayed. After determining the target to-be-displayed supply information corresponding to the target terminal behavior information of the target user, the client may rearrange the display sequence corresponding to each of the plurality of target supply information according to a preset rule, where the rearranged preset rule may, but does not include: topping, weighting, inserting, model rearrangement and the like, for example, but not limited to, ranking according to the recommendation score corresponding to each target supply information in the server recommendation result from large to small, or directly topping the target supply information corresponding to the current intention of each target supply information in the server recommendation result, or ranking the recommendation score obtained by weighting and summing the recommendation score corresponding to each target supply information in the server recommendation result and the target score output by the target current intention corresponding to the intention recognition model corresponding to each target supply information, or directly inserting the target supply information corresponding to the current intention of each target with the highest target score under the supply information currently browsed by the target user, or directly inputting the current end behavior information and a plurality of target supply information into the trained end intelligent model to obtain a corresponding rearrangement result and the like. The intelligent terminal model is obtained by training based on a plurality of supply information sets of known rearrangement results and corresponding terminal behavior information. The supply information set includes a plurality of supply information.

Optionally, in order to reduce the pressure of the server and effectively control the network bandwidth resource consumption, a cache rearrangement interval may be preset on the client, for example, but not limited to, a cache rearrangement interval duration is 1 minute, 5 minutes, etc., so that the client, within the cache rearrangement interval duration after obtaining the server recommendation result issued by the server, even if the target user performs an operation on the client, will not trigger the client to send a new recommendation request to the server, and after obtaining the server recommendation result issued by the server and passing the cache rearrangement interval duration, will automatically trigger the server to issue a new server recommendation result, and then trigger a new rearrangement operation after determining that the target corresponding to the target user is to display the supply information.

And S1114, rendering and displaying the plurality of target supply information based on the target rearrangement result.

Specifically, after the target rearrangement results of the target supply information are obtained, the target supply information can be rendered according to the display sequence of the target supply information in the target rearrangement results, so that the corresponding target supply information is displayed at the corresponding display position in the page to be displayed corresponding to the current end behavior information, the user experience and the service display effect are improved, and the conversion rate of the target user in the page to be displayed is improved to a certain extent.

Reference is next made to fig. 12, which is a schematic diagram illustrating an implementation process of an intent recognition method according to an exemplary embodiment of the present disclosure. The above intent recognition method is applied to a client, as shown in fig. 12, where the client may include an end intent recall module, an end intent recognition module, an end cache intent match determination module, an end intent recall module, and an end rearrangement module. And the terminal intention recall module is used for carrying out fusion recall on multiple paths of intentions after the terminal decision point is triggered to obtain a current intention set corresponding to the target user. The above-mentioned end intention recall module fuses and recalls the multiple intents, and the detailed description of the implementation process of S202 and S204 is omitted here. The terminal intention recognition module is configured to perform inference recognition on the current intention set based on the trained intention recognition model to obtain the target current intention of the target user, and the specific implementation process of the terminal intention recognition module is referred to S206 and will not be described herein. The terminal cache intention matching judging module is configured to judge whether the current intention of the target is accepted by the server side recommendation result, and finally determine the supply information to be displayed of the target according to the judgment result, and the specific implementation is referred to S208, and details are not repeated here. The terminal intention expanding and recalling module is configured to expand and recall the current intention of the target to be recommended, which is not received by the server recommendation result, from the cloud supply pool to obtain corresponding second supply information, and please refer to the description corresponding to fig. 8, which is not repeated here. The above-mentioned end rearrangement module is configured to rearrange the display order of the plurality of target supply information in the target supply information to be displayed, and specific reference is made to the implementation process of S1112 and S1114, which are not described herein again.

Reference is next made to fig. 13, which is a schematic structural diagram of an intent recognition device according to an exemplary embodiment of the present disclosure. The intention recognition device is applied to the client; as shown in fig. 13, the intention recognition apparatus 1300 includes:

an obtaining module 1310, configured to target current end behavior information of a user;

a first determining module 1320, configured to determine, based on the current end behavior information, a current intent set corresponding to the target user;

a second determining module 1330, configured to determine a target current intent corresponding to the target user based on the current intent set, a target end behavior feature corresponding to the target user, and target user information corresponding to the target user;

the intention matching module 1340 is configured to match a recommendation intention corresponding to a server recommendation result with the target current intention to obtain a target matching result;

and a third determining module 1350, configured to determine, based on the target matching result, target to-be-displayed supply information corresponding to the target user.

In one possible implementation manner, the first determining module 1320 includes:

the extraction unit is used for extracting corresponding current supply information from the current end behavior information; the current supply information is used for representing recommendation information associated with the current end behavior of the target user;

And the first determining unit is used for determining the current intention set corresponding to the target user based on the current supply information.

In one possible implementation manner, the first determining unit includes:

an extracting subunit, configured to extract at least one corresponding intention tag from the attribute of the current supply information;

and the first determining subunit is used for determining the current intention set corresponding to the target user based on the at least one intention label.

In one possible implementation manner, the first determining unit is specifically configured to:

mapping the current supply information according to a target mapping relation to obtain a current intention set corresponding to the target user; the target mapping relationship is used for representing the corresponding relationship between the supply information and the intention.

In one possible implementation manner, the first determining unit includes:

a second determining subunit, configured to determine a current end intention set corresponding to the target user based on the current supply information;

and the fusion recall subunit is used for carrying out fusion recall on the current end intention set and the current cloud intention set corresponding to the target user to obtain the current intention set corresponding to the target user.

In one possible implementation manner, the second determining module 1330 includes:

the intention recognition unit is used for inputting the current intention set, the target end behavior characteristics corresponding to the target user and the target user information corresponding to the target user into an intention recognition model and outputting an intention recognition result corresponding to the target user; the intention recognition model is obtained by training based on terminal behavior information, terminal behavior logs and user information corresponding to each of a plurality of users;

and a second determining unit, configured to determine a target current intention corresponding to the target user based on the intention recognition result.

In one possible implementation manner, the intention recognition result includes a target score corresponding to each current intention in the current intention set; the target score is used for representing the deviation degree of the target user on the current intention; the second determining unit is specifically configured to:

determining a target current intention corresponding to the target user based on the current intention set and target scores corresponding to the current intentions in the current intention set; the target current intention comprises N current intentions with highest corresponding target scores in the current intention set or current intentions with corresponding target scores greater than a preset score in the current intention set; the above-mentioned N is a positive integer.

In one possible implementation manner, the page to be displayed corresponding to the current end behavior information includes at least one display position; the server side recommending result comprises first supply information which is issued by the server side and is recommended corresponding to each display position in the at least one display position; the first supply information is used for representing information recommended by the server based on target basic information and target terminal behavior information corresponding to the target user; the target terminal behavior information comprises historical terminal behavior information of the target user and/or current terminal behavior information;

the intention recognition apparatus 1300 further includes:

and the fourth determining module is used for determining the recommendation intention corresponding to the server side recommendation result based on the first supply information corresponding to the recommendation of each display position in the at least one display.

In one possible implementation, the number of the target current intents is at least one; the target matching result comprises the current intention of the target to be recommended; the target current intention to be recommended is used for representing the target current intention which is not matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention; the third determining module 1350 includes:

The uploading unit is used for uploading the current intention of the target to be recommended to the server, so that the server recalls corresponding second supply information from the cloud supply pool based on the current intention of the target to be recommended; the cloud supply pool includes a plurality of supply information;

the receiving unit is used for receiving the second supply information issued by the server;

and a third determining unit, configured to determine target to-be-displayed supply information corresponding to the target user based on the second supply information.

In one possible implementation, the target matching result further includes a current intention of the recommended target; the recommended target current intention is used for representing the target current intention matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention; the third determining unit is specifically configured to: determining target to-be-displayed supply information corresponding to the target user based on first supply information and the second supply information corresponding to the target recommendation intention in the server recommendation result; the target recommendation intention is used for representing the recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server side recommendation result.

In one possible implementation, the target matching result includes a recommended target current intention; the recommended target current intention is used for representing the target current intention matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention; the third determining unit is specifically configured to:

determining target to-be-displayed supply information corresponding to the target user based on first supply information corresponding to the target recommendation intention in the server recommendation result; the target recommendation intention is used for representing a recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server recommendation result.

In one possible implementation manner, the target to-be-displayed supply information includes a plurality of target supply information to be displayed; the intention recognition apparatus 1300 further includes:

the rearrangement module is used for rearranging the plurality of target supply information according to a preset rule to obtain target rearrangement results corresponding to the plurality of target supply information;

and the display module is used for rendering and displaying the plurality of target supply information based on the target rearrangement result.

The division of the modules in the intent recognition device is for illustration only, and in other embodiments, the intent recognition device may be divided into different modules as needed to perform all or part of the functions of the intent recognition device. The implementation of the respective modules in the intention recognition apparatus provided in the embodiments of the present specification may be in the form of a computer program. The computer program may be run on a client. Program modules of the computer program may be stored in the memory of the client. Which when executed by a processor, implements all or part of the steps of the intent recognition method described in the embodiments of the present specification.

Next, please refer to fig. 14, which is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present disclosure. As shown in fig. 14, the electronic device 1400 may include: at least one processor 1410, at least one communication bus 1420, a user interface 1430, at least one network interface 1440, a memory 1450.

Wherein a communication bus 1420 may be used to enable connection communication of the various components described above.

The user interface 1430 may include a Display (Display) and a Camera (Camera), and the optional user interface may also include a standard wired interface, a wireless interface, among others.

The network interface 1440 may optionally include, among other things, a bluetooth module, a near field communication (Near Field Communication, NFC) module, a wireless fidelity (Wireless Fidelity, wi-Fi) module, and the like.

Wherein processor 1410 may include one or more processing cores. The processor 1410 utilizes various interfaces and lines to connect various portions of the overall electronic device 1400, perform various functions of the routing electronic device 1400, and process data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 1450, and invoking data stored in the memory 1450. Alternatively, the processor 1410 may be implemented in hardware in at least one of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 1410 may integrate one or a combination of several of a processor (Central Processing Unit, CPU), an image processor (Graphics Processing Unit, GPU), and a modem, etc. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 1410, but may be implemented by a single chip.

The Memory 1450 may include a random access Memory (Random Access Memory, RAM) or a Read-Only Memory (ROM). Optionally, the memory 1450 includes a non-transitory computer-readable medium. The memory 1450 may be used to store instructions, programs, code, sets of codes, or instruction sets. The memory 1450 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as an acquisition function, an intent matching function, a determination function, etc.), instructions for implementing the various method embodiments described above, and the like; the storage data area may store data or the like referred to in the above respective method embodiments. The memory 1450 may also optionally be at least one storage device located remotely from the aforementioned processor 1410. As shown in fig. 14, an operating system, network communication modules, user interface modules, and program instructions can be included in the memory 1450, which is a type of computer storage medium.

In some possible embodiments, the electronic device 1400 may be the intent recognition apparatus described above, and the processor 1410 may be configured to invoke program instructions stored in the memory 1450 and specifically perform the following operations:

Acquiring current end behavior information of a target user; determining a current intention set corresponding to the target user based on the current end behavior information; determining a target current intention corresponding to the target user based on the current intention set, the target end behavior characteristic corresponding to the target user and the target user information corresponding to the target user; matching the recommendation intention corresponding to the server recommendation result with the target current intention to obtain a target matching result; and determining target to-be-displayed supply information corresponding to the target user based on the target matching result.

In some possible embodiments, when the processor 1410 executes the determining, based on the current end behavior information, a current intent set corresponding to the target user, the method specifically is used to execute: extracting corresponding current supply information from the current end behavior information; the current supply information is used for representing recommendation information associated with the current end behavior of the target user; and determining a current intention set corresponding to the target user based on the current supply information.

In some possible embodiments, when the processor 1410 executes the determining, based on the current supply information, a current intent set corresponding to the target user, the method specifically includes: extracting at least one corresponding intention label from the attribute of the current supply information; and determining a current intention set corresponding to the target user based on the at least one intention label.

In some possible embodiments, when the processor 1410 determines, based on the current supply information, a current intent set corresponding to the target user, the method specifically includes: mapping the current supply information according to a target mapping relation to obtain a current intention set corresponding to the target user; the target mapping relationship is used for representing the corresponding relationship between the supply information and the intention.

In some possible embodiments, when the processor 1410 executes the determining, based on the current supply information, a current intent set corresponding to the target user, the method specifically includes: determining a current end intention set corresponding to the target user based on the current supply information; and carrying out fusion recall on the current end intention set and the current cloud intention set corresponding to the target user to obtain the current intention set corresponding to the target user.

In some possible embodiments, when the processor 1410 determines the target current intent corresponding to the target user based on the current intent set, the target end behavior feature corresponding to the target user, and the target user information corresponding to the target user, the method specifically includes: inputting the current intention set, the target end behavior characteristics corresponding to the target user and the target user information corresponding to the target user into an intention recognition model, and outputting an intention recognition result corresponding to the target user; the intention recognition model is obtained by training based on terminal behavior information, terminal behavior logs and user information corresponding to each of a plurality of users; and determining the target current intention corresponding to the target user based on the intention recognition result.

In some possible embodiments, the intention recognition result includes a target score corresponding to each current intention in the current intention set; the target score is used for representing the deviation degree of the target user on the current intention; the processor 1410 is specifically configured to execute, when executing the determination of the target current intention corresponding to the target user based on the intention recognition result: determining a target current intention corresponding to the target user based on the current intention set and target scores corresponding to the current intentions in the current intention set; the target current intention comprises N current intentions with highest corresponding target scores in the current intention set or current intentions with corresponding target scores greater than a preset score in the current intention set; the above-mentioned N is a positive integer.

In some possible embodiments, the page to be displayed corresponding to the current end behavior information includes at least one display position; the server side recommending result comprises first supply information which is issued by the server side and is recommended corresponding to each display position in the at least one display position; the first supply information is used for representing information recommended by the server based on target basic information and target terminal behavior information corresponding to the target user; the target terminal behavior information comprises historical terminal behavior information of the target user and/or current terminal behavior information;

The processor 1410 performs the matching between the recommendation intention corresponding to the server recommendation result and the target current intention, and before obtaining the target matching result, is further configured to perform: and determining the recommendation intention corresponding to the server side recommendation result based on the first supply information of the recommendation corresponding to each display position in the at least one display.

In some possible embodiments, the number of target current intents is at least one; the target matching result comprises the current intention of the target to be recommended; the target current intention to be recommended is used for representing the target current intention which is not matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention; when the processor 1410 executes the target to-be-presented supply information corresponding to the target user determined based on the target matching result, the method is specifically configured to execute: uploading the current intention of the target to be recommended to the server, so that the server recalls corresponding second supply information from the cloud supply pool based on the current intention of the target to be recommended; the cloud supply pool includes a plurality of supply information; receiving the second supply information issued by the server; and determining target to-be-displayed supply information corresponding to the target user based on the second supply information.

In some possible embodiments, the target matching result further includes a recommended target current intention; the recommended target current intention is used for representing the target current intention matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention; when the processor 1410 executes the determining, based on the second offer information, the offer information to be displayed corresponding to the target user, the method specifically includes: determining target to-be-displayed supply information corresponding to a target user based on first supply information and second supply information corresponding to the target recommendation intention in the server recommendation result; the target recommendation intention is used for representing the recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server side recommendation result.

In some possible embodiments, the target matching result includes a recommended target current intention; the recommended target current intention is used for representing the target current intention matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention; when the processor 1410 executes the target to-be-presented supply information corresponding to the target user determined based on the target matching result, the method is specifically configured to execute: determining target to-be-displayed supply information corresponding to the target user based on first supply information corresponding to the target recommendation intention in the server recommendation result; the target recommendation intention is used for representing a recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server recommendation result.

In some possible embodiments, the target to-be-displayed offer information includes a plurality of target offer information to be displayed; after executing the determining, by the processor 1410, the target to-be-presented supply information corresponding to the target user based on the target matching result, the method is further configured to execute: rearranging the plurality of target supply information according to a preset rule to obtain a target rearranging result corresponding to the plurality of target supply information; rendering and displaying the plurality of target supply information based on the target rearrangement result.

The present description also provides a computer-readable storage medium having instructions stored therein, which when executed on a computer or processor, cause the computer or processor to perform one or more steps of the above embodiments. The respective constituent modules of the above-described intention recognition apparatus may be stored in the above-described computer-readable storage medium if implemented in the form of software functional units and sold or used as independent products.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product described above includes one or more computer instructions. When the computer program instructions described above are loaded and executed on a computer, the processes or functions described in accordance with the embodiments of the present specification are all or partially produced. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted across a computer-readable storage medium. The computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage media may be any available media that can be accessed by a computer or a data storage device such as a server, data center, or the like that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy Disk, a hard Disk, a magnetic tape), an optical medium (e.g., a digital versatile Disk (Digital Versatile Disc, DVD)), or a semiconductor medium (e.g., a Solid State Disk (SSD)), or the like.

Those skilled in the art will appreciate that implementing all or part of the above-described embodiment methods may be accomplished by way of a computer program, which may be stored in a computer-readable storage medium, instructing relevant hardware, and which, when executed, may comprise the embodiment methods as described above. And the aforementioned storage medium includes: various media capable of storing program code, such as ROM, RAM, magnetic or optical disks. The technical features in the present examples and embodiments may be arbitrarily combined without conflict.

The above-described embodiments are merely preferred embodiments of the present disclosure, and do not limit the scope of the disclosure, and various modifications and improvements made by those skilled in the art to the technical solution of the disclosure should fall within the scope of protection defined by the claims.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims and description may be performed in an order different from that in the embodiments recited in the description and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. An intent recognition method, the method being applied to a client, the method comprising:

acquiring current end behavior information of a target user;

2. The method of claim 1, the determining the current set of intents corresponding to the target user based on the current end behavior information, comprising:

extracting corresponding current supply information from the current end behavior information; the current supply information is used for representing recommendation information associated with the current end behavior of the target user;

and determining a current intention set corresponding to the target user based on the current supply information.

3. The method of claim 2, the determining the current set of intents corresponding to the target user based on the current feed information, comprising:

Extracting at least one corresponding intention label from the attribute of the current supply information;

and determining a current intention set corresponding to the target user based on the at least one intention label.

4. The method of claim 2, the determining the current set of intents corresponding to the target user based on the current feed information, comprising:

5. The method of claim 2, the determining the current set of intents corresponding to the target user based on the current feed information, comprising:

determining a current end intention set corresponding to the target user based on the current supply information;

and carrying out fusion recall on the current end intention set and the current cloud intention set corresponding to the target user to obtain the current intention set corresponding to the target user.

6. The method of claim 1, the determining the target current intent corresponding to the target user based on the current intent set, the target-end behavioral characteristics corresponding to the target user, and the target user information corresponding to the target user, comprising:

Inputting the current intention set, the target end behavior characteristics corresponding to the target user and the target user information corresponding to the target user into an intention recognition model, and outputting an intention recognition result corresponding to the target user; the intention recognition model is obtained by training based on terminal behavior information, terminal behavior logs and user information corresponding to each of a plurality of users;

and determining the target current intention corresponding to the target user based on the intention recognition result.

7. The method of claim 6, the intent recognition result comprising a target score for each current intent in the current intent set; the target score is used for representing the deviation degree of the target user on the current intention;

the determining the target current intention corresponding to the target user based on the intention recognition result comprises the following steps:

determining a target current intention corresponding to the target user based on the target score corresponding to each current intention in the current intention set; the target current intention comprises N current intentions with highest corresponding target scores in the current intention set or current intentions with corresponding target scores greater than a preset score in the current intention set; and N is a positive integer.

8. The method of claim 1, wherein the page to be displayed corresponding to the current behavior information includes at least one display; the server side recommending result comprises first supply information which is issued by the server side and is recommended corresponding to each display position in the at least one display position; the first supply information is used for representing information recommended by the server based on target basic information corresponding to the target user and target behavior information; the target terminal behavior information comprises historical terminal behavior information and/or current terminal behavior information of the target user;

the recommendation intention corresponding to the server recommendation result is matched with the target current intention, and before the target matching result is obtained, the method further comprises the following steps:

and determining the recommendation intention corresponding to the server side recommendation result based on the first supply information corresponding to the recommendation of each display position in the at least one display.

9. The method of claim 8, the number of target current intents being at least one; the target matching result comprises the current intention of the target to be recommended; the target current intention to be recommended is used for representing the target current intention which is not matched with the recommendation intention corresponding to the server side recommendation result in at least one target current intention;

The determining the target to-be-displayed supply information corresponding to the target user based on the target matching result comprises the following steps:

uploading the current intention of the target to be recommended to the server, so that the server recalls corresponding second supply information from a cloud supply pool based on the current intention of the target to be recommended; the cloud supply pool includes a plurality of supply information;

receiving the second supply information issued by the server;

and determining target to-be-displayed supply information corresponding to the target user based on the second supply information.

10. The method of claim 9, the target matching result further comprising a recommended target current intent; the recommended target current intention is used for representing the target current intention matched with the recommendation intention corresponding to the server-side recommendation result in at least one target current intention;

the determining, based on the second supply information, target supply information to be displayed corresponding to the target user includes:

determining target to-be-displayed supply information corresponding to the target user based on first supply information and second supply information corresponding to the target recommendation intention in the server recommendation result; the target recommendation intention is used for representing a recommendation intention matched with the current intention of the recommended target in the recommendation intention corresponding to the server recommendation result.

11. The method of claim 8, the target matching result comprising a recommended target current intent; the recommended target current intention is used for representing the target current intention matched with the recommendation intention corresponding to the server-side recommendation result in at least one target current intention;

12. The method of claim 1, the target offer information to be presented comprising a plurality of target offer information to be presented;

after the target to-be-displayed supply information corresponding to the target user is determined based on the target matching result, the method further comprises:

rearranging the plurality of target supply information according to a preset rule to obtain a target rearranging result corresponding to the plurality of target supply information;

Rendering and displaying the plurality of target supply information based on the target rearrangement result.

13. An intent recognition device, the device being applied to a client; the device comprises:

14. An electronic device, comprising: a processor and a memory;

the processor is connected with the memory;

the memory is used for storing executable program codes;

the processor runs a program corresponding to executable program code stored in the memory by reading the executable program code for performing the method according to any one of claims 1-12.

15. A computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method steps of any of claims 1-12.

16. A computer program product comprising instructions which, when run on a computer or processor, cause the computer or processor to perform the method of any of claims 1-12.