CN113377976B

CN113377976B - Resource searching method and device, computer equipment and storage medium

Info

Publication number: CN113377976B
Application number: CN202110936767.5A
Authority: CN
Inventors: 曹效伦
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2021-08-16
Filing date: 2021-08-16
Publication date: 2022-09-09
Anticipated expiration: 2041-08-16
Also published as: CN113377976A

Abstract

The disclosure relates to a resource searching method, a resource searching device, computer equipment and a storage medium, and belongs to the technical field of multimedia. The method comprises the following steps: responding to a search request, and determining a search word and a first multimedia resource corresponding to the search request; acquiring search word characteristics of the search word and first resource characteristics of the first multimedia resource; determining at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of the plurality of candidate multimedia resources; the at least one second multimedia resource is transmitted. The method and the device avoid the multimedia resources which are related to the search word but unrelated to the first multimedia resource from appearing in the search result, so that more accurate search result is obtained, and the search accuracy of the search behavior is improved.

Description

Resource searching method, device, computer equipment and storage medium

Technical Field

The present disclosure relates to the field of multimedia technologies, and in particular, to a resource search method and apparatus, a computer device, and a storage medium.

Background

With the development of multimedia technology and the diversification of search functions, a user can browse multimedia resources anytime and anywhere through an application program on a terminal, and generally, the application program also provides a search function for the multimedia resources, that is, the user inputs an interested search word in the application program so as to conveniently view each multimedia resource related to the search word.

In the above search process, for the same search term, the multimedia resources related to the search term may be various, for example, for the search term "laoshan", the related multimedia resources include: mountain views of Laoshan, self-portrait photos of the Laoshan, Laoshan beer, Qingdao seascape and the like, and how to obtain more accurate search results becomes a problem which needs to be improved urgently aiming at various different types of multimedia resources related to search terms.

Disclosure of Invention

The present disclosure provides a resource search method, apparatus, computer device, and storage medium to obtain a more accurate search result for a multimedia resource search behavior.

According to an aspect of the embodiments of the present disclosure, there is provided a resource search method, including:

responding to a search request, determining a search word and a first multimedia resource corresponding to the search request, wherein the first multimedia resource is associated with the search word;

acquiring search word characteristics of the search word and first resource characteristics of the first multimedia resource;

determining at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of the plurality of candidate multimedia resources, the second multimedia resource being associated with both the search term and the first multimedia resource;

transmitting the at least one second multimedia resource.

In one possible implementation, the determining, in response to a search request, a search term and a first multimedia resource corresponding to the search request includes:

if the search request is a first search request, the search word and the resource identifier of the first multimedia resource are obtained from the first search request, the first search request is generated based on a trigger operation of a content tag on a playing interface of the first multimedia resource, and the content tag is used as the search word in the first search request.

if the search request is a second search request, obtaining the search word carried by the search request, wherein the second search request is generated based on input operation in a search input box;

and selecting the multimedia resource associated with the search word as the first multimedia resource from the historical browsing records.

In one possible embodiment, the selecting, from the historical browsing records, the multimedia resource associated with the search term as the first multimedia resource includes:

and selecting the multimedia resource with the content tag comprising the search word and the latest timestamp as the first multimedia resource from the historical browsing records.

In one possible implementation, the selecting, from the historical browsing records, the multimedia resource associated with the search term as the first multimedia resource includes:

and selecting the multimedia resource with the highest similarity with the search word from the historical browsing records as the first multimedia resource.

In one possible implementation, the determining at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of the plurality of candidate multimedia resources comprises:

determining at least one third multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature and a second resource feature of the plurality of candidate multimedia resources, wherein the similarity between the third multimedia resource and the search term and the first multimedia resource meets a first target condition;

acquiring a target behavior parameter of the at least one third multimedia resource, wherein the target behavior parameter is used for representing the possibility of target behavior of the account on the third multimedia resource;

and acquiring the third multimedia resource with the target behavior parameter meeting the second target condition from the at least one third multimedia resource as the at least one second multimedia resource.

In one possible implementation, the determining at least one third multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and the second resource feature of the plurality of candidate multimedia resources comprises:

for any candidate multimedia resource in the candidate multimedia resources, acquiring a first similarity between the search term characteristic and a second resource characteristic of the candidate multimedia resource;

obtaining a second similarity between the first resource feature and a second resource feature of the candidate multimedia resource;

weighting the first similarity and the second similarity to obtain a third similarity;

determining the candidate multimedia resource as a third multimedia resource if the third similarity meets the first target condition.

In one possible implementation, the determining that the candidate multimedia resource is a third multimedia resource if the third similarity meets the first target condition includes:

ranking the plurality of candidate multimedia resources based on a descending order of third similarity;

and if the candidate multimedia resource is positioned at the first target position in the sequence, determining the candidate multimedia resource as a third multimedia resource.

and if the third similarity is larger than a similarity threshold value, determining the candidate multimedia resource as a third multimedia resource.

In a possible implementation manner, the obtaining of the target behavior parameter of the at least one third multimedia resource includes:

inputting the search word feature, the first resource feature and a second resource feature of the third multimedia resource into a target behavior model for any third multimedia resource in the at least one third multimedia resource, wherein the target behavior model is used for acquiring a target behavior parameter of the input multimedia resource;

weighting the search term characteristics, the first resource characteristics and the second resource characteristics of the third multimedia resources through the target behavior model to obtain target behavior characteristics;

and carrying out index normalization on the target behavior characteristics to obtain the target behavior parameters.

In a possible implementation manner, the obtaining, from the at least one third multimedia resource, a third multimedia resource whose target behavior parameter meets a second target condition as the at least one second multimedia resource includes:

and sequencing the at least one third multimedia resource based on the sequence of the target behavior parameters from large to small, and selecting the third multimedia resource with the sequence positioned at the front second target position as the at least one second multimedia resource.

and selecting the third multimedia resource with the target behavior parameter larger than the behavior parameter threshold value from the at least one third multimedia resource as the at least one second multimedia resource.

In one possible embodiment, the target behavior parameters include: at least one of a click behavior parameter, a like behavior parameter, a forward behavior parameter, a comment behavior parameter, or a collection behavior parameter.

According to another aspect of the embodiments of the present disclosure, there is provided a resource search method, including:

sending a search request carrying search terms;

receiving at least one second multimedia resource returned based on the search request, wherein the second multimedia resource is associated with both the search word and the first multimedia resource, and the first multimedia resource is associated with the search word;

and displaying the at least one second multimedia resource.

In one possible implementation, the sending the search request carrying the search term includes:

displaying a plurality of content tags of the first multimedia resource in a playing interface of the first multimedia resource;

and responding to the triggering operation of any content label in the plurality of content labels, and sending the search request carrying the content label and the resource identifier of the first multimedia resource.

According to another aspect of the embodiments of the present disclosure, there is provided a resource searching apparatus, including:

the first determining unit is configured to execute the steps of responding to a search request, determining a search word corresponding to the search request and a first multimedia resource, wherein the first multimedia resource is associated with the search word;

an acquisition unit configured to perform acquisition of a search term characteristic of the search term and a first resource characteristic of the first multimedia resource;

a second determining unit configured to perform determining at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of a plurality of candidate multimedia resources, the second multimedia resource being associated with both the search term and the first multimedia resource;

a transmitting unit configured to perform transmitting the at least one second multimedia resource.

In one possible embodiment, the first determining unit is configured to perform:

if the search request is a first search request, the search word and the resource identifier of the first multimedia resource are obtained from the first search request, the first search request is generated based on the triggering operation of the content tag on the playing interface of the first multimedia resource, and the content tag is used as the search word in the first search request.

In one possible implementation, the first determining unit includes:

the first obtaining subunit is configured to, if the search request is a second search request, obtain the search term carried by the search request, where the second search request is generated based on an input operation in a search input box;

a selecting subunit configured to perform selecting a multimedia resource associated with the search term from a historical browsing record as the first multimedia resource.

In one possible embodiment, the selection subunit is configured to perform:

In one possible implementation, the second determining unit includes:

a determining subunit configured to perform determining at least one third multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature and a second resource feature of the plurality of candidate multimedia resources, a similarity between the third multimedia resource and the search term and the first multimedia resource meeting a first target condition;

the second obtaining subunit is configured to perform obtaining of a target behavior parameter of the at least one third multimedia resource, where the target behavior parameter is used to characterize a possibility of a target behavior occurring on the third multimedia resource by the account;

and the third acquiring subunit is configured to acquire, from the at least one third multimedia resource, a third multimedia resource whose target behavior parameter meets a second target condition as the at least one second multimedia resource.

In one possible embodiment, the determining subunit includes:

an obtaining subunit, configured to perform obtaining, for any candidate multimedia resource of the plurality of candidate multimedia resources, a first similarity between the search term feature and a second resource feature of the candidate multimedia resource;

the obtaining subunit is further configured to perform obtaining a second similarity between the first resource characteristic and a second resource characteristic of the candidate multimedia resource;

a weighting subunit configured to perform weighting on the first similarity and the second similarity to obtain a third similarity;

a determining subunit configured to perform determining that the candidate multimedia resource is a third multimedia resource if the third similarity meets the first target condition.

In one possible embodiment, the determining subunit is configured to perform:

In one possible implementation, the second obtaining subunit is configured to perform:

In one possible implementation, the third obtaining subunit is configured to perform:

a transmitting unit configured to perform transmitting a search request carrying a search word;

a receiving unit configured to perform receiving at least one second multimedia resource returned based on the search request, the second multimedia resource being associated with both the search term and a first multimedia resource, the first multimedia resource being associated with the search term;

a display unit configured to perform displaying the at least one second multimedia asset.

In one possible embodiment, the sending unit is configured to perform:

and responding to the triggering operation of any content tag in the plurality of content tags, and sending the search request carrying the content tag and the resource identifier of the first multimedia resource.

According to another aspect of the embodiments of the present disclosure, there is provided a computer apparatus including:

one or more processors;

one or more memories for storing the one or more processor-executable instructions;

wherein the one or more processors are configured to perform the resource searching method of any one of the possible implementations of the above-described aspect.

According to another aspect of embodiments of the present disclosure, there is provided a computer-readable storage medium, wherein at least one instruction of the computer-readable storage medium, when executed by one or more processors of a computer device, enables the computer device to perform the resource search method in any one of the possible implementations of the above-described one aspect.

According to another aspect of embodiments of the present disclosure, there is provided a computer program product comprising one or more instructions executable by one or more processors of a computer device to enable the computer device to perform the resource search method of any one of the possible implementations of the above-mentioned one aspect.

The technical scheme provided by the embodiment of the disclosure at least brings the following beneficial effects:

when a search request is received, a first multimedia resource associated with a search word is determined, and first resource characteristics of the first multimedia resource are introduced into the process of obtaining a search result, so that each second multimedia resource serving as the search result is not only associated with the search word, but also associated with the first multimedia resource, and the multimedia resources associated with the search word but not associated with the first multimedia resource are prevented from appearing in the search result, so that a more accurate search result is obtained, and the search accuracy of a search behavior is improved.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure and are not to be construed as limiting the disclosure.

FIG. 1 is a schematic diagram of an implementation environment of a resource search method according to an example embodiment;

FIG. 2 is a flow diagram illustrating a method of resource searching in accordance with an exemplary embodiment;

FIG. 3 is an interaction flow diagram illustrating a method of resource searching in accordance with an exemplary embodiment;

fig. 4 is a flowchart of determining a second multimedia resource according to an embodiment of the present disclosure;

FIG. 5 is a schematic diagram illustrating a resource searching method according to an embodiment of the present disclosure;

FIG. 6 is a block diagram illustrating a logical structure of a resource search apparatus in accordance with an illustrative embodiment;

FIG. 7 is a block diagram illustrating a logical structure of a resource search apparatus in accordance with an exemplary embodiment;

fig. 8 is a block diagram illustrating a structure of a terminal according to an exemplary embodiment of the present disclosure;

fig. 9 is a schematic structural diagram of a server according to an embodiment of the present disclosure.

Detailed Description

In order to make the technical solutions of the present disclosure better understood, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.

It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the foregoing drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein are capable of operation in sequences other than those illustrated or otherwise described herein. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the disclosure, as detailed in the appended claims.

The user information to which the present disclosure relates may be information authorized by the user or sufficiently authorized by each party.

Hereinafter, terms related to the embodiments of the present disclosure are explained.

Vertical search: also called vertical search, vertical search for short. Vertical search is a product form widely existing in an APP (Application program) of a mobile terminal on a current smart phone, and indicates a search behavior for querying a certain vertical classification and a corresponding search result page thereof as indicated by a literal meaning. Illustratively, taking a short video search scene as an example, a user takes a content tag of a certain short video as a search word and initiates a vertical search, and short videos (or called works) related to the content tag, that is, a vertical category, will be displayed in a result page, where the result page is also referred to as a detail page of the vertical category or a landing page of the search behavior at this time.

Unlike the conventional search behavior, the vertical search behavior initiated by the user carries an attribute of "shopping and strolling", that is, the user may only think about what short videos related to the vertical category are but does not have an explicit search intention, so that each short video in the vertical search result page should carry a recommendation attribute. Generally, although each short video in the result page belongs to the vertical category, for example, each short video in the content vertical category on the short video platform is recalled by the vertical search algorithm after the video publisher itself hits the corresponding content tag, because there are behaviors such as "degree of smearing", that is, the video publisher hits the content tag which is not related but has high degree of popularity to the short video uploaded by itself, it is difficult to ensure whether each short video presented in the result page is really related to the vertical category.

In an exemplary scenario, an entry where a user initiates a vertical search is a short video carrying a certain content tag, for example, in the process of browsing a current short video, the user creates curiosity about the content tag carried by the short video, and when the user clicks the content tag, the user jumps to a result page of the content tag, so as to view other short videos also carrying the content tag.

In view of this, the embodiment of the present disclosure provides a resource searching method, which can introduce the characteristics of a multimedia resource (i.e., a first multimedia resource) where a user has a click behavior into a resource sorting and recall algorithm of a result page, that is, on the premise that the user is interested in the first multimedia resource, promote a second multimedia resource with a higher similarity to the first multimedia resource in the result page to increase exposure, so that the user can browse more second multimedia resources with possible interest in the result page, thereby improving the searching accuracy of each multimedia resource in a vertical search result page, and increasing the consumption and experience of the user on a vertical search behavior, where the first multimedia resource may also be referred to as a flow-through multimedia resource, a flow-through work, and the like.

Fig. 1 is a schematic diagram of an implementation environment of a resource search method according to an exemplary embodiment, and referring to fig. 1, at least one terminal 101 and a server 102 may be included in the implementation environment, which is described in detail below.

The at least one terminal 101 is used for browsing multimedia resources, each of the at least one terminal 101 may have an application installed thereon, the application may be any client capable of providing a multimedia resource browsing service, a user may browse the multimedia resources by starting the application, the application may be at least one of a short video application, an audio and video application, a shopping application, a take-out application, a travel application, a game application or a social application, and the multimedia resources may include at least one of a video resource, an audio resource, a picture resource, a text resource or a web page resource.

At least one terminal 101 may be directly or indirectly connected to the server 102 through wired or wireless communication, which is not limited in the embodiment of the present disclosure.

The server 102 is a computer device for providing a multimedia resource search service to the at least one terminal 101. The server 102 may include at least one of a server, a plurality of servers, a cloud computing platform, or a virtualization center. Alternatively, the server 102 may undertake primary computational work and the at least one terminal 101 may undertake secondary computational work; alternatively, the server 102 may undertake secondary computing work and the at least one terminal 101 may undertake primary computing work; alternatively, the server 102 and the at least one terminal 101 perform cooperative computing by using a distributed computing architecture.

It should be noted that the device type of any one of the at least one terminal 101 may include: at least one of a smart phone, a tablet computer, an e-book reader, an MP3 (Moving Picture Experts Group Audio Layer III, motion Picture Experts compression standard Audio Layer 3) player, an MP4 (Moving Picture Experts Group Audio Layer IV, motion Picture Experts compression standard Audio Layer 4) player, a laptop portable computer, or a desktop computer. For example, the any terminal may be a smartphone, or other hand-held portable electronic device. The following embodiments are illustrated with the terminal comprising a smartphone.

Those skilled in the art will appreciate that the number of terminals described above may be greater or fewer. For example, the number of the terminals may be only one, or several tens or hundreds of the terminals, or more. The number of terminals and the type of the device are not limited in the embodiments of the present disclosure.

Fig. 2 is a flowchart illustrating a resource searching method according to an exemplary embodiment, and referring to fig. 2, the resource searching method is applied to a computer device, and the following description takes the computer device as an example as a server.

In step 201, in response to a search request, a search word and a first multimedia resource corresponding to the search request are determined, and the first multimedia resource is associated with the search word.

In step 202, a search term characteristic of the search term and a first resource characteristic of the first multimedia resource are obtained.

In step 203, at least one second multimedia resource is determined from the candidate multimedia resources based on the search term feature, the first resource feature and a second resource feature of the candidate multimedia resources, wherein the second multimedia resource is associated with both the search term and the first multimedia resource.

In step 204, the at least one second multimedia asset is transmitted.

According to the method provided by the embodiment of the disclosure, when a search request is received, the first multimedia resource associated with the search word is determined, and the first resource characteristic of the first multimedia resource is introduced into the process of obtaining the search result, so that each second multimedia resource serving as the search result is not only associated with the search word, but also associated with the first multimedia resource, and the multimedia resources associated with the search word but not associated with the first multimedia resource are prevented from appearing in the search result, so that a more accurate search result is obtained, and the search accuracy of the search behavior is improved.

In one possible implementation, in response to a search request, determining a search term and a first multimedia resource corresponding to the search request includes:

and selecting the multimedia resource associated with the search word from the historical browsing records as the first multimedia resource.

In one possible implementation, selecting the multimedia resource associated with the search term from the historical browsing records as the first multimedia resource includes:

and selecting the multimedia resource with the content tag including the search word and the latest timestamp as the first multimedia resource from the historical browsing records.

In one possible embodiment, selecting the multimedia resource associated with the search term from the historical browsing records as the first multimedia resource comprises:

In one possible implementation, determining at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and second resource features of the plurality of candidate multimedia resources comprises:

acquiring a target behavior parameter of the at least one third multimedia resource, wherein the target behavior parameter is used for representing the possibility of the account number for generating a target behavior on the third multimedia resource;

In one possible embodiment, determining at least one third multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and the second resource feature of the plurality of candidate multimedia resources comprises:

acquiring a second similarity between the first resource characteristic and a second resource characteristic of the candidate multimedia resource;

and under the condition that the third similarity meets the first target condition, determining the candidate multimedia resource as a third multimedia resource.

In one possible implementation manner, in the case that the third similarity meets the first target condition, determining the candidate multimedia resource as a third multimedia resource includes:

ranking the plurality of candidate multimedia resources based on the order of the third similarity from large to small;

if the candidate multimedia resource is located at the first target position in the sequence, determining the candidate multimedia resource as a third multimedia resource.

and if the third similarity is larger than the similarity threshold value, determining the candidate multimedia resource as a third multimedia resource.

In a possible implementation, the obtaining of the target behavior parameter of the at least one third multimedia resource comprises:

weighting the search term feature, the first resource feature and the second resource feature of the third multimedia resource through the target behavior model to obtain a target behavior feature;

In one possible embodiment, obtaining, from the at least one third multimedia resource, a third multimedia resource whose target behavior parameter meets the second target condition as the at least one second multimedia resource includes:

and sequencing the at least one third multimedia resource according to the sequence of the target behavior parameters from large to small, and selecting the third multimedia resource with the sequence positioned at the front second target position as the at least one second multimedia resource.

All the above optional technical solutions may be combined arbitrarily to form the optional embodiments of the present disclosure, and are not described herein again.

Fig. 3 is an interaction flowchart illustrating a resource search method according to an exemplary embodiment, where, as shown in fig. 3, the resource search method is used in an interaction process between a terminal and a server, and the terminal and the server are exemplary illustrations of computer devices, and the embodiment includes the following steps.

In step 301, the terminal sends a search request carrying a search term to the server.

The terminal is provided with and runs with an application program supporting multimedia resources, and a user starts the application program on the terminal and can browse each multimedia resource in the application program. Optionally, a multimedia resource currently browsed by the user is referred to as a first multimedia resource, and the user may trigger generation of the search request through a content tag of the first multimedia resource in a process of browsing the first multimedia resource. Optionally, the user may also directly click the search tag in the application program, trigger and display the search input box and the search confirmation option, directly input the search word in the search input box, click the search confirmation option, and trigger and generate the search request, which is not specifically limited in the embodiment of the present application.

In some embodiments, the search request may be divided into a first search request and a second search request, depending on how the search request is triggered. The first search request is generated based on the triggering operation of the content label on the playing interface of the first multimedia resource, and the second search request is generated based on the input operation of the user in the search input box.

In some embodiments, since the first search request is generated based on a trigger operation on a content tag on a playing interface of the first multimedia resource, meaning that the first search request is formed by streaming of the first multimedia resource, in the case of browsing the first multimedia resource, a user performs a trigger operation on the content tag of the first multimedia resource, and the terminal sends the first search request to the server.

In some embodiments, the terminal displays a plurality of content tags of the first multimedia resource in a playing interface of the first multimedia resource; and sending a search request carrying the content tag and the resource identifier of the first multimedia resource in response to the triggering operation of the user on any content tag in the plurality of content tags. The content tag is a search word, and the search request is a first search request.

Optionally, the triggering operation includes, but is not limited to: the touch control method includes click operation, touch operation, double-click operation, long-press operation, drag operation, voice instruction, gesture instruction and the like.

In the process, the terminal generates the search request through the triggering of the content tag of the first multimedia resource, so that the content tag is used as a search word, and the resource identifier of the first multimedia resource is directly added in the search request to represent that the content tag is associated with the first multimedia resource, thereby facilitating the determination of the search word and the first multimedia resource by the server, and improving the calculation efficiency of the server.

In some embodiments, since the second search request is generated based on the user's input operation in the search input box, it means that the second search request is not guided by some certain determined first multimedia resource, but is triggered by a search action actively and spontaneously initiated by the user in the search input box, at which time the terminal sends the second search request to the server.

In some embodiments, the terminal displays a search tag in a main interface of an application program, a user can display a search input box and a search confirmation option in the main interface by performing a trigger operation on the search tag, then, the user can perform an input operation in the search input box and perform a trigger operation on the search confirmation option after the input is completed, the terminal obtains information input in the search input box as a search word, and sends a search request carrying the search word, where the search request is a second search request.

Optionally, the input operation includes, but is not limited to: the input method is used for manual input, the automatic voice recognition technology is used for voice input and converting the voice input into corresponding text information, and the like, and the embodiment of the disclosure does not specifically limit the operation mode of input operation.

In the process, the terminal triggers and generates the search request through the input operation in the search input box, so that in a scene without explicit diversion multimedia resources, the server can select the first multimedia resources according to the historical browsing records of the account of the login terminal under the condition, and each second multimedia resource can still be acquired by using the resource search mode provided by the embodiment of the disclosure, so that the accuracy of the search result is improved.

In step 302, the server determines a search word and a first multimedia resource corresponding to the search request in response to the search request, wherein the first multimedia resource is associated with the search word.

In some embodiments, the server may determine the search term and the first multimedia resource in different ways according to the type of the search request. The following discussion will be directed to the first search request and the second search request, respectively.

In some embodiments, if the search request is a first search request, the server obtains the search word and the resource identifier of the first multimedia resource from the first search request, where the first search request is generated based on a trigger operation on a content tag on a play interface of the first multimedia resource, and the first search request takes the content tag as the search word. For example, the content tag is encapsulated into a search term field in the first search request.

Optionally, the server receives the search request, analyzes a header field in the search request, determines that the search request is the first search request if the header field carries the type identifier of the first search request, and at this time, may directly analyze a search word field of the first search request to obtain the search word, analyze a resource identifier field of the first search request to obtain the resource identifier, and determine that the multimedia resource indicated by the resource identifier is the first multimedia resource.

In the process, the server can directly analyze the first search request to obtain the search terms and the resource identifier of the first multimedia resource without additional processing, so that the computing resource of the server is saved, and the computing efficiency of the server is improved.

In some embodiments, if the search request is a second search request, the search term carried by the search request is obtained, and the second search request is generated based on an input operation in a search input box; and selecting the multimedia resource associated with the search word from the historical browsing records as the first multimedia resource.

Optionally, the server receives the search request, parses a header field in the search request, determines that the search request is the second search request if the header field carries a type identifier of the second search request, and then may directly parse a search term field of the second search request to obtain the search term, further, parses a user field of the second search request to obtain an account identifier of an account logging in the terminal corresponding to the search request, then obtains a historical browsing record corresponding to the account identifier, and queries, from the historical browsing record, a multimedia resource associated with the search term as the first multimedia resource.

In the process, the server can configure the corresponding first multimedia resource according to the historical browsing record of the account for the second search request even if the server does not carry a clear resource identifier to indicate the corresponding first multimedia resource, so that the universality of the resource search method provided by the embodiment of the disclosure can be improved, and the search accuracy under various scenes can be integrally improved.

In some embodiments, when the server selects the first multimedia resource from the historical browsing history, the server may select a multimedia resource whose content tag includes the search term and whose timestamp is the latest from the historical browsing history as the first multimedia resource.

Optionally, the server obtains a tag set formed by all content tags carried by all multimedia resources in the history browsing record, and queries whether any content tag is hit in the index in the tag set by using the search word as the index, and if any content tag is hit in the index, the server indicates that the search word is included in any content tag. The above process is repeatedly executed, one or more content tags hit by the index can be determined, at this time, one or more multimedia resources carrying the one or more content tags are obtained from the historical browsing record, and the multimedia resource with the latest timestamp in the one or more multimedia resources is selected as the first multimedia resource.

Optionally, the latest meaning of the timestamp may be that the timestamp of the last browsing of the corresponding multimedia resource by the account is latest, or the timestamp of the video publisher publishing the corresponding multimedia resource is latest, or the timestamp of any account in the platform reviewing the corresponding multimedia resource is latest, which is not specifically limited in this embodiment of the disclosure.

In the above process, for the case that the search term has high correlation with the content tag, the server selects, as the first multimedia resource, the multimedia resource having the content tag and the latest timestamp from the historical browsing record of the account, so as to improve the search accuracy, because when the user inputs a certain content tag in the search input box, the user usually browses the multimedia resource carrying the content tag once and curiously wants to view other similar multimedia resources, so that the first multimedia resource is determined from the historical browsing record, and information supplement can be performed for the scene missing the first multimedia resource.

In some embodiments, if the index misses all content tags in the tag set, in this case, the server may select a multimedia resource with the highest similarity to the search term from the historical browsing records as the first multimedia resource.

Optionally, the server may first perform the operation of performing a matching query on the tag set based on the index, select, as the first multimedia resource, a multimedia resource having any content tag and a latest timestamp if the index can hit any content tag, and if the index cannot hit all content tags, may further obtain a similarity between the search term and each multimedia resource in the history browsing record, and select, as the first multimedia resource, a multimedia resource with a highest similarity, so as to improve a direct association between the first multimedia resource and the search term.

Optionally, the server may also not perform the above operation of performing a matching query on the tab set based on the index, but directly obtain the similarity between the search term and each multimedia resource in the historical browsing record, and select the multimedia resource with the highest similarity as the first multimedia resource, so that the processing flow of obtaining the first multimedia resource can be simplified.

Optionally, when the server obtains the similarity between the search term and each multimedia resource in the history browsing record, for any multimedia resource in the history browsing record, the server may extract an Embedding (Embedding) feature as a search term feature for the search term, extract an Embedding feature as a resource feature for the detail information of the multimedia resource, then obtain a euclidean distance between the search term feature and the resource feature, and obtain the similarity between the search term and the multimedia resource based on the euclidean distance, where the similarity is negatively correlated with the euclidean distance.

Optionally, when the server obtains the similarity between the search word and each multimedia resource in the history browsing record, for any multimedia resource in the history browsing record, the server may extract an embedded feature as a search word feature for the search word, extract an embedded feature as a resource feature for the detailed information of the multimedia resource, and then obtain a cosine distance between the search word feature and the resource feature as the similarity between the search word and the multimedia resource.

In some embodiments, in addition to the above-mentioned embedded feature for extracting the detail information of the multimedia resource as the resource feature, the multimedia resource may be subjected to feature extraction by using a machine learning model, which is not limited to: convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), Multi-Layer perceptrons (MLPs), and the like, which are not specifically limited in this embodiment of the present disclosure.

In the above process, for the case that the search term does not have high correlation with the content tag, the server selects, from the historical browsing records of the account, the multimedia resource with the highest similarity with the search term as the first multimedia resource, so as to improve the search accuracy, because when the user inputs a certain search term in the search input box, the user may browse the multimedia resource related to the search term once and cause curiosity to view other similar multimedia resources, so that the first multimedia resource is determined from the historical browsing records, and information supplement can be performed for the scene missing the first multimedia resource.

In step 303, the server obtains a search term characteristic of the search term and a first resource characteristic of the first multimedia resource.

In some embodiments, the server may perform embedding processing on the search term by using a term vector model to obtain an embedded feature of the search term, and use the embedded feature of the search term as the search term feature, thereby improving the expression capability of the search term.

In some embodiments, the server may further perform One-hot (One-hot) encoding on the search term to obtain a One-hot vector of the search term, and use the One-hot vector of the search term as the search term feature, so that the computation complexity in obtaining the search term feature may be simplified.

In some embodiments, the server may obtain the detail information of the first multimedia asset, which may be text formed by at least one of a title, a brief description, a summary, a publisher nickname, a publisher account number, or a content tag of the first multimedia asset. Next, the detail information is subjected to embedding processing to obtain the embedding feature of the detail information, and the embedding feature of the detail information is used as the first resource feature, so that the complexity in acquiring the first resource feature can be simplified.

In some embodiments, the server may further input the detail information of the first multimedia resource and the key frame of the first multimedia resource into a feature extraction model, extract a fusion feature between the detail information and the key frame through the feature extraction model, and use the fusion feature as the first resource feature, so as to improve the expressive power of the first resource feature. Optionally, the key frame may be a cover of the first multimedia resource, or may be one or more key video frames of the first multimedia resource, where the key video frames may be set by a video publisher, or may be intelligently identified and extracted by a server, and the embodiment of the present disclosure is not specifically limited to this.

Optionally, the feature extraction model may be a CNN, DNN, MLP, or the like, and may also be a machine learning model with multi-modal processing capability, which is not specifically limited in the embodiment of the present disclosure. Illustratively, the feature extraction model may include a word vector submodel and a DNN submodel, the embedding feature of the detail information is extracted through the word vector submodel, the image feature of the key frame is extracted through the DNN submodel, the embedding feature and the image feature are spliced and input into the full-link layer, and the fusion feature is extracted through the full-link layer.

In step 304, the server determines at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of the plurality of candidate multimedia resources, the second multimedia resource being associated with both the search term and the first multimedia resource.

In some embodiments, the server may obtain all multimedia resources from the multimedia resource library as the plurality of candidate multimedia resources, or the server may randomly sample a target number of multimedia resources from the multimedia resource library as the plurality of candidate multimedia resources, and the like, which is not specifically limited in the embodiments of the present disclosure.

In some embodiments, for each candidate multimedia resource, the server may obtain detail information for the candidate multimedia resource, which may be text formed from at least one of a title, a summary, a publisher nickname, a publisher account number, or a content tag of the candidate multimedia resource. Then, the detail information is embedded to obtain the embedded feature of the detail information, and the embedded feature of the detail information is used as the second resource feature, so that the complexity of acquiring the second resource feature can be simplified.

In some embodiments, the server may further input the detail information of the candidate multimedia resource and the key frame of the candidate multimedia resource into a feature extraction model, extract a fusion feature between the detail information and the key frame through the feature extraction model, and use the fusion feature as the second resource feature, so as to improve the expressive ability of the second resource feature. Optionally, the key frame may be a cover page of the candidate multimedia resource, or may be one or more key video frames of the candidate multimedia resource, where the key video frames may be set by a video publisher, or may be intelligently identified and extracted by a server, and the embodiment of the present disclosure is not specifically limited to this.

Optionally, the feature extraction model may be CNN, DNN, MLP, or the like, and may also be a machine learning model with multi-modal processing capability, which is not specifically limited by the embodiments of the present disclosure. Illustratively, the feature extraction model may include a word vector submodel and a DNN submodel, the embedding feature of the detail information is extracted through the word vector submodel, the image feature of the key frame is extracted through the DNN submodel, the embedding feature and the image feature are spliced and input into the full-link layer, and the fusion feature is extracted through the full-link layer.

In some embodiments, after the search term feature, the first resource feature, and the second resource feature of each candidate multimedia resource are extracted, fig. 4 is a flowchart of determining the second multimedia resource according to an embodiment of the disclosure, as shown in fig. 4, the server may determine each second multimedia resource through the following

steps

3041 and 3043, which are described in detail below.

In step 3041, the server determines at least one third multimedia resource from the candidate multimedia resources based on the search term feature, the first resource feature and the second resource feature of the candidate multimedia resources, wherein the similarity between the third multimedia resource and the search term and the first multimedia resource meets the first target condition.

In some embodiments, the server, in determining the respective third multimedia resource, may perform the following operations for any candidate multimedia resource of the plurality of candidate multimedia resources: acquiring a first similarity between the search term characteristic and a second resource characteristic of the candidate multimedia resource; acquiring a second similarity between the first resource characteristic and a second resource characteristic of the candidate multimedia resource; weighting the first similarity and the second similarity to obtain a third similarity; and under the condition that the third similarity meets the first target condition, determining the candidate multimedia resource as a third multimedia resource.

In some embodiments, the first similarity may be a cosine distance between the search word feature and the second resource feature of the candidate multimedia resource, or an inverse euclidean distance between the search word feature and the second resource feature of the candidate multimedia resource, or other parameters for characterizing the similarity degree, such that the similarity degree between the search word and the candidate multimedia resource is higher when the first similarity degree is larger, and conversely, the similarity degree between the search word and the candidate multimedia resource is lower when the first similarity degree is smaller.

In some embodiments, the second similarity may be a cosine distance between the first resource feature and a second resource feature of the candidate multimedia resource, an inverse euclidean distance between the first resource feature and the second resource feature of the candidate multimedia resource, or other parameters for representing the degree of similarity, such that the greater the second similarity, the higher the degree of similarity between the first multimedia resource and the candidate multimedia resource, and vice versa, the smaller the second similarity, the lower the degree of similarity between the first multimedia resource and the candidate multimedia resource.

In some embodiments, when obtaining the third similarity, an arithmetic mean value between the first similarity and the second similarity may be used as the third similarity, or a harmonic mean value between the first similarity and the second similarity may be used as the third similarity, or a first coefficient and a second coefficient may be obtained, a sum of the first coefficient and the second coefficient is equal to 1, the first coefficient is multiplied by the first similarity to obtain a first value, the second coefficient is multiplied by the second similarity to obtain a second value, the first value is added to the second value to obtain the third similarity, wherein the first coefficient represents a specific gravity of the first similarity in the third similarity, and the second coefficient represents a specific gravity of the second similarity in the third similarity. The first coefficient and the second coefficient are both numerical values greater than or equal to 0 and less than or equal to 1.

The server executes the above operation on each candidate multimedia resource, so as to obtain a third similarity of each candidate multimedia resource, and then selects a candidate multimedia resource with the third similarity meeting the first target condition as the third multimedia resource.

In some embodiments, in determining whether the third similarity meets the first target condition, the server may perform the following operations: sorting the plurality of candidate multimedia resources in order from large to small based on the third similarity; if the candidate multimedia resource is located at the first target position in the sequence, determining the candidate multimedia resource as a third multimedia resource. Wherein the first target bit is any integer number of bits greater than or equal to 1.

In the above process, the first target position candidate multimedia resources with the largest third similarity among all the candidate multimedia resources are selected as the third multimedia resources, which is equivalent to recall the third multimedia resources related to the search term and the first multimedia resources from the massive candidate multimedia resources, so that the final second multimedia resources are determined by sequencing the third multimedia resources through the following

steps

3042 and 3043, and the search accuracy of the second multimedia resources can be improved.

In some embodiments, in determining whether the third similarity meets the first target condition, the server may perform the following operations: and if the third similarity is larger than the similarity threshold value, determining the candidate multimedia resource as a third multimedia resource. Wherein the similarity threshold is any value greater than or equal to 0.

In the process, the candidate multimedia resources with the third similarity larger than the similarity threshold value are selected from all the candidate multimedia resources as the third multimedia resources, so that the number of the third multimedia resources determined in the recall stage is not limited, the omission of some candidate multimedia resources with higher similarity degrees is avoided, and the searching accuracy of the second multimedia resources is improved.

In step 3042, the server obtains a target behavior parameter of the at least one third multimedia resource, where the target behavior parameter is used to characterize a possibility that the account performs a target behavior on the third multimedia resource.

In some embodiments, the target behavior parameters include: at least one of a click behavior parameter, a like behavior parameter, a forward behavior parameter, a comment behavior parameter, or a collection behavior parameter. For example, the target behavior parameter is a click behavior parameter, that is, an estimated click rate, which represents a predicted possibility that the account clicks the third multimedia resource, and for example, the target behavior parameter is a praise behavior parameter, that is, an estimated praise rate, which represents a predicted possibility that the account praise the third multimedia resource, and the like.

In some embodiments, for any one of the at least one third multimedia resource, the server may perform the following operations when obtaining the target behavior parameter: inputting the search word feature, the first resource feature and a second resource feature of the third multimedia resource into a target behavior model, wherein the target behavior model is used for acquiring a target behavior parameter of the input multimedia resource; weighting the search term characteristics, the first resource characteristics and the second resource characteristics of the third multimedia resources through the target behavior model to obtain target behavior characteristics; and carrying out index normalization on the target behavior characteristics to obtain the target behavior parameters.

Illustratively, the target behavior model is DNN, the DNN includes at least one hidden layer and an index normalization layer, the server concatenates (Concat) the search term feature, the first resource feature and the second resource feature of the third multimedia resource to obtain a concatenated feature, inputs the concatenated feature into at least one hidden layer of the DNN, weights the concatenated feature by the at least one hidden layer, and outputs the target behavior feature from the last hidden layer, where the at least one hidden layer is connected in series, that is, the feature output from the previous hidden layer serves as the feature input from the next hidden layer. Then, the target behavior characteristics are input into the index normalization layer, an index normalization function is called in the index normalization layer to carry out index normalization on the target behavior characteristics, and the target behavior parameters are output.

In the process, the target behavior parameters are predicted by calling the target behavior model, the prediction accuracy of the target behavior parameters can be improved, different target behavior models can be trained for different target behavior parameters, for example, if the target behavior parameters are the predicted click rate, the click rate model can be trained, and if the target behavior parameters are the predicted click rate, the click rate model can be trained, so that different target behavior models can be flexibly configured for different target behavior parameters.

In step 3043, the server obtains, from the at least one third multimedia resource, the third multimedia resource whose target behavior parameter meets the second target condition as the at least one second multimedia resource.

In some embodiments, the server, in determining whether the target behavior parameter meets the second target condition, may perform the following operations: and sequencing the at least one third multimedia resource according to the sequence of the target behavior parameters from large to small, and selecting the third multimedia resource with the sequence positioned at the front second target position as the at least one second multimedia resource. Wherein the second target bit is any integer number of bits greater than or equal to 1.

In the process, the second target position third multimedia resource with the largest target behavior parameter is selected from all the third multimedia resources as the second multimedia resource, which is equivalent to the second multimedia resource which is most likely to be interested by the user can be screened from the recalled third multimedia resources, so that the second multimedia resource which is more interested by the user is recommended according to the search behavior of the user, and the search accuracy in the resource search process is improved.

In some embodiments, the server, in determining whether the target behavior parameter meets the second target condition, may perform the following: and selecting the third multimedia resource with the target behavior parameter larger than the behavior parameter threshold value from the at least one third multimedia resource as the at least one second multimedia resource.

In the process, the third multimedia resource with the target behavior parameter larger than the behavior parameter threshold value is selected as the second multimedia resource from all the third multimedia resources, so that the number of the second multimedia resources determined in the fine ranking stage is not limited, the second multimedia resources which are possibly interested by some users are avoided being omitted, and the searching accuracy of the second multimedia resources is improved.

In some embodiments, since the terminal usually displays only a fixed number of multimedia resources in the search result page, after the server determines the second multimedia resource in the above two manners, the server may also sample the fixed number of second multimedia resources in a random sampling manner, and perform the following step 305, so as to flexibly control the number of the returned second multimedia resources.

In the

above step

3041 and 3043, the stage of recalling the third multimedia resource from the candidate multimedia resources may be regarded as a coarse screening process, the coarse screening process can ensure that the recalled third multimedia resource is not only related to the search term, but also related to the guided first multimedia resource, the stage of determining the second multimedia resource from the third multimedia resource may be regarded as a fine ranking process, the fine ranking process can ensure that the screened second multimedia resource is a resource more in line with the interest of the user, and the two-stage screening process can improve the search accuracy of the finally determined second multimedia resource.

In step 305, the server transmits the at least one second multimedia asset to the terminal.

In some embodiments, the server may send, to the terminal, related information of the at least one second multimedia resource, where the related information carries at least a resource identifier and a sequencing sequence number of the at least one second multimedia resource, and may also carry at least one of the following: a cover, a title, a brief introduction, a summary, a content tag, a nickname of a video publisher, etc. of the at least one second multimedia asset, which is not particularly limited by the embodiments of the present disclosure.

In some embodiments, the server may encapsulate the related information of the at least one second multimedia resource by using a data Transmission Protocol to obtain a search result message corresponding to the search request, where the data Transmission Protocol may be a Transmission Control Protocol (TCP), a User Datagram Protocol (UDP), an Internet Protocol (IP), and the like, and this is not specifically limited in this embodiment of the disclosure.

In some embodiments, before sending the search result message, the server may further compress and encrypt the search result message, and the compression algorithm used for compression and the encryption algorithm used for encryption are not specifically limited in the embodiments of the present disclosure.

In step 306, the terminal receives at least one second multimedia resource returned by the server based on the search request.

In some embodiments, the terminal may receive the search result message returned by the server, and if the search result message is compressed and encrypted, the search result message may be decrypted and decompressed, where an encryption algorithm used for decryption matches an encryption algorithm used for encryption, and a decompression algorithm used for decompression matches a compression algorithm used for compression. Further, after the preprocessing, the search result message is analyzed to obtain the related information of the at least one second multimedia resource, where the related information at least carries the resource identifier and the sequencing sequence number of the at least one second multimedia resource, and in addition, may also carry at least one of the following: a cover, a title, a brief introduction, a summary, a content tag, a nickname of a video publisher, etc. of the at least one second multimedia asset, which is not particularly limited by the embodiments of the present disclosure.

In step 307, the terminal displays the at least one second multimedia asset.

In some embodiments, the terminal may display the result page of the search, and display the at least one second multimedia resource in the result page, for example, display a cover of the at least one second multimedia resource in the result page based on the ranking number of the at least one second multimedia resource, and display a respective title under each cover of the at least one second multimedia resource, so that the user may click on the cover or title of any second multimedia resource, trigger the terminal to load the any second multimedia resource from the server, and jump to a playing interface of the any second multimedia resource to play the any second multimedia resource.

Fig. 5 is a schematic diagram of a resource searching method provided by an embodiment of the present disclosure, and as shown in fig. 5, when a user browses a first multimedia resource, a terminal provides a content tab 511 of the first multimedia resource in a playing interface 510 of the first multimedia resource, where the first multimedia resource may be referred to as a work to be guided, and the content tab 511 may be referred to as a tab button. The terminal responds to the click operation of the user on the content tag 511, sends a search request to the server, and receives 6 second multimedia resources returned by the server. Then, the user jumps from the playing interface 510 to the result page 520 for displaying the search, the cover 521 and 526 of the 6 second multimedia resources are displayed in the result page 520, and the user can click the cover of any second multimedia resource, thereby jumping from the result page 520 to the playing interface of any second multimedia resource.

In the embodiment of the present disclosure, when each second multimedia resource in the result page 520 is processed, a feature of a guided work (i.e., a first resource feature) in the play interface 510 is introduced into the recall and sorting algorithm, so that a work (i.e., a second multimedia resource) more similar to the guided work can be arranged in front in the result page 520 to increase exposure, so that when a user browses each work in the result page 520, each work and the guided work have a stronger content similarity, and the user's visual inertia is used to increase the user's consumption of the result page 520, in other words, the sorting problem of the vertically searched result page 520 can be converted into a recommendation problem, so as to recommend more works meeting the user's consumption inertia.

Furthermore, in the process of testing by applying the resource searching method provided by the embodiment of the disclosure, the site tag is used for testing, by introducing the features of the work of flow guidance into the ranking model, i.e., the target behavior model, the user is enabled to select, from among different works of flow guidance, even if the same place tag is clicked on, the ranking of each work presented in the search result page returned by each is different, the reason is that the sequencing model ranks the more similar works in the content of the guide works in the front, so that the user can find out that more similar and related works with the guide works are displayed in the result page during consumption, the visual inertia of the user consumption is met, the consumption amount of the user is increased, meanwhile, the overall similarity between the works in the result page and the guide works is improved, so that the correlation between the works in the result page and the searching behaviors is improved.

When a Point of Interest (POI) tag is taken as an example for testing, by introducing a first resource feature of a first multimedia resource, which is a diversion work clicked by a user, into a target behavior model (hereinafter, this model is abbreviated as "this model"), the correlation is greatly improved when the user performs manual gsb (good-same-bad) evaluation, wherein gsb is a shape like a: b: and c, comparing the search results returned by different models aiming at the same search word after inputting a plurality of same search words into the model and the traditional model, and comparing the advantages and the disadvantages of the performance (such as indexes of accuracy, recall rate and the like), wherein the value a represents the number of the search words with improved performance compared with the traditional model, the value b represents the number of the search words with the same performance compared with the existing model, the value c represents the number of the search words with deteriorated performance compared with the existing model, and when the value a is greater than the value c, the performance of the model is represented by the traditional model, and the larger the difference between the value a and the value c is, the more the performance of the model is represented. In the experiment, gsb for the high heat POI word was 89: 9: 9, gsb for a random POI, 57: 46: 27, the experimental results show that the posterior indexes are also greatly improved, including the click rate improvement of +2.8%, the exposure duration improvement of +4.15%, and the like. Where POI refers to any non-geographically meaningful point on the map: such as shops, bars, gas stations, hospitals, stations, etc.

Fig. 6 is a block diagram illustrating a logical structure of a resource search apparatus according to an exemplary embodiment. Referring to fig. 6, the apparatus includes a first determining unit 601, an acquiring unit 602, a second determining unit 603, and a transmitting unit 604:

a first determining unit 601 configured to perform, in response to a search request, determining a search word corresponding to the search request and a first multimedia resource, the first multimedia resource being associated with the search word;

an obtaining unit 602 configured to perform obtaining a search term feature of the search term and a first resource feature of the first multimedia resource;

a second determining unit 603 configured to perform determining at least one second multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of a plurality of candidate multimedia resources, the second multimedia resource being associated with both the search term and the first multimedia resource;

a sending unit 604 configured to perform sending the at least one second multimedia resource.

According to the device provided by the embodiment of the disclosure, when a search request is received, the first multimedia resources associated with the search terms are determined, and the first resource characteristics of the first multimedia resources are introduced into the process of obtaining the search result, so that each second multimedia resource serving as the search result is not only associated with the search terms, but also associated with the first multimedia resources, and the multimedia resources associated with the search terms but not associated with the first multimedia resources are prevented from appearing in the search result, so that a more accurate search result is obtained, and the search accuracy of the search behavior is improved.

In one possible implementation, the first determining unit 601 is configured to perform:

if the search request is a first search request, the search term and the resource identifier of the first multimedia resource are obtained from the first search request, the first search request is generated based on the triggering operation of the content tag on the playing interface of the first multimedia resource, and the content tag is used as the search term in the first search request.

In a possible implementation, based on the apparatus composition of fig. 6, the first determining unit 601 includes:

the first obtaining subunit is configured to perform, if the search request is a second search request, obtaining the search word carried by the search request, where the second search request is generated based on an input operation in a search input box;

and the selecting subunit is configured to select the multimedia resource associated with the search term from the historical browsing records as the first multimedia resource.

In one possible embodiment, the selection subunit is configured to perform:

and selecting the multimedia resource with the content tag comprising the search word and the latest timestamp as the first multimedia resource from the historical browsing record.

In one possible embodiment, the selection subunit is configured to perform:

In a possible implementation, based on the apparatus composition of fig. 6, the second determining unit 603 includes:

a determining subunit configured to perform determining at least one third multimedia resource from the plurality of candidate multimedia resources based on the search term feature, the first resource feature, and a second resource feature of the plurality of candidate multimedia resources, a similarity between the third multimedia resource and the search term and the first multimedia resource meeting a first target condition;

the second obtaining subunit is configured to perform obtaining of a target behavior parameter of the at least one third multimedia resource, where the target behavior parameter is used to characterize a possibility that the account performs a target behavior on the third multimedia resource;

and the third acquiring subunit is configured to acquire, from the at least one third multimedia resource, a third multimedia resource with a target behavior parameter meeting a second target condition as the at least one second multimedia resource.

In one possible embodiment, based on the apparatus composition of fig. 6, the determining subunit includes:

the acquiring subunit is configured to execute acquiring, for any candidate multimedia resource in the plurality of candidate multimedia resources, a first similarity between the search term feature and a second resource feature of the candidate multimedia resource;

the obtaining subunit is further configured to perform obtaining a second similarity between the first resource feature and a second resource feature of the candidate multimedia resource;

a determining subunit configured to perform, in a case that the third similarity meets the first target condition, determining that the candidate multimedia resource is a third multimedia resource.

In one possible embodiment, the determining subunit is configured to perform:

if the candidate multimedia resource is located at the first target position in the ranking, the candidate multimedia resource is determined to be a third multimedia resource.

In one possible embodiment, the determining subunit is configured to perform:

In one possible embodiment, the second acquiring subunit is configured to perform:

In one possible embodiment, the third obtaining subunit is configured to perform:

With regard to the apparatus in the above-mentioned embodiment, the specific manner in which each unit performs the operation has been described in detail in the embodiment related to the resource searching method, and will not be elaborated herein.

Fig. 7 is a block diagram illustrating a logical structure of a resource search apparatus according to an exemplary embodiment, and referring to fig. 7, the apparatus includes a transmitting unit 701, a receiving unit 702, and a display unit 703.

A transmitting unit 701 configured to perform transmitting a search request carrying a search word;

a receiving unit 702 configured to perform receiving at least one second multimedia resource returned based on the search request, the second multimedia resource being associated with both the search term and a first multimedia resource, the first multimedia resource being associated with the search term;

a display unit 703 configured to perform displaying the at least one second multimedia asset.

According to the device provided by the embodiment of the disclosure, after the search request is sent, each received second multimedia resource is not only associated with the search word, but also associated with the first multimedia resource, so that the multimedia resources which are associated with the search word but not associated with the first multimedia resource are prevented from appearing in the search result, a more accurate search result is obtained, and the search accuracy of the search behavior is improved.

In one possible implementation, the sending unit 701 is configured to perform:

displaying a plurality of content labels of the first multimedia resource in a playing interface of the first multimedia resource;

With regard to the apparatus in the above-described embodiment, the specific manner in which each unit performs the operation has been described in detail in the embodiment related to the resource search method, and will not be elaborated here.

Fig. 8 shows a block diagram of a terminal, which is an exemplary illustration of a computer device, according to an exemplary embodiment of the present disclosure. The terminal 800 may be: a smart phone, a tablet computer, an MP3 player (Moving Picture Experts Group Audio Layer III, motion video Experts compression standard Audio Layer 3), an MP4 player (Moving Picture Experts Group Audio Layer IV, motion video Experts compression standard Audio Layer 4), a notebook computer, or a desktop computer. The terminal 800 may also be referred to by other names such as user equipment, portable terminal, laptop terminal, desktop terminal, etc.

In general, the terminal 800 includes: a processor 801 and a memory 802.

Processor 801 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and so forth. The processor 801 may be implemented in at least one hardware form of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), and a PLA (Programmable Logic Array). The processor 801 may also include a main processor and a coprocessor, where the main processor is a processor for Processing data in an awake state, and is also called a Central Processing Unit (CPU); a coprocessor is a low power processor for processing data in a standby state. In some embodiments, the processor 801 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content required to be displayed on the display screen. In some embodiments, the processor 801 may further include an AI (Artificial Intelligence) processor for processing computing operations related to machine learning.

Memory 802 may include one or more computer-readable storage media, which may be non-transitory. Memory 802 may also include high speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In some embodiments, a non-transitory computer readable storage medium in memory 802 is used to store at least one instruction for execution by processor 801 to implement the resource search methods provided by various embodiments of the present disclosure.

In some embodiments, the terminal 800 may further optionally include: a peripheral interface 803 and at least one peripheral. The processor 801, memory 802, and peripheral interface 803 may be connected by buses or signal lines. Various peripheral devices may be connected to peripheral interface 803 by a bus, signal line, or circuit board. Specifically, the peripheral device includes: at least one of a radio frequency circuit 804, a touch screen display 805, a camera assembly 806, an audio circuit 807, a positioning assembly 808, and a power supply 809.

The peripheral interface 803 may be used to connect at least one peripheral related to I/O (Input/Output) to the processor 801 and the memory 802. In some embodiments, the processor 801, memory 802, and peripheral interface 803 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 801, the memory 802, and the peripheral interface 803 may be implemented on separate chips or circuit boards, which are not limited by this embodiment.

The Radio Frequency circuit 804 is used for receiving and transmitting RF (Radio Frequency) signals, also called electromagnetic signals. The radio frequency circuitry 804 communicates with communication networks and other communication devices via electromagnetic signals. The radio frequency circuit 804 converts an electrical signal into an electromagnetic signal to transmit, or converts a received electromagnetic signal into an electrical signal. Optionally, the radio frequency circuit 804 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a subscriber identity module card, and so forth. The radio frequency circuit 804 may communicate with other terminals via at least one wireless communication protocol. The wireless communication protocols include, but are not limited to: metropolitan area networks, various generation mobile communication networks (2G, 3G, 4G, and 5G), Wireless local area networks, and/or WiFi (Wireless Fidelity) networks. In some embodiments, the radio frequency circuit 804 may also include NFC (Near Field Communication) related circuits, which are not limited by this disclosure.

The display screen 805 is used to display a UI (User Interface). The UI may include graphics, text, icons, video, and any combination thereof. When the display 805 is a touch display, the display 805 also has the ability to capture touch signals on or above the surface of the display 805. The touch signal may be input to the processor 801 as a control signal for processing. At this point, the display 805 may also be used to provide virtual buttons and/or a virtual keyboard, also referred to as soft buttons and/or a soft keyboard. In some embodiments, the display 805 may be one, providing the front panel of the terminal 800; in other embodiments, the display 805 may be at least two, respectively disposed on different surfaces of the terminal 800 or in a folded design; in still other embodiments, the display 805 may be a flexible display disposed on a curved surface or a folded surface of the terminal 800. Even further, the display 805 may be arranged in a non-rectangular irregular pattern, i.e., a shaped screen. The Display 805 can be made of LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), and other materials.

The camera assembly 806 is used to capture images or video. Optionally, camera assembly 806 includes a front camera and a rear camera. Generally, a front camera is disposed at a front panel of the terminal, and a rear camera is disposed at a rear surface of the terminal. In some embodiments, the number of the rear cameras is at least two, and each rear camera is any one of a main camera, a depth-of-field camera, a wide-angle camera and a telephoto camera, so that the main camera and the depth-of-field camera are fused to realize a background blurring function, and the main camera and the wide-angle camera are fused to realize panoramic shooting and VR (Virtual Reality) shooting functions or other fusion shooting functions. In some embodiments, camera assembly 806 may also include a flash. The flash lamp can be a monochrome temperature flash lamp or a bicolor temperature flash lamp. The double-color-temperature flash lamp is a combination of a warm-light flash lamp and a cold-light flash lamp, and can be used for light compensation at different color temperatures.

The audio circuit 807 may include a microphone and a speaker. The microphone is used for collecting sound waves of a user and the environment, converting the sound waves into electric signals, and inputting the electric signals to the processor 801 for processing or inputting the electric signals to the radio frequency circuit 804 to realize voice communication. The microphones may be provided in a plurality, respectively, at different portions of the terminal 800 for the purpose of stereo sound collection or noise reduction. The microphone may also be an array microphone or an omni-directional pick-up microphone. The speaker is used to convert electrical signals from the processor 801 or the radio frequency circuit 804 into sound waves. The loudspeaker can be a traditional film loudspeaker or a piezoelectric ceramic loudspeaker. When the speaker is a piezoelectric ceramic speaker, the speaker can be used for purposes such as converting an electric signal into a sound wave audible to a human being, or converting an electric signal into a sound wave inaudible to a human being to measure a distance. In some embodiments, the audio circuitry 807 may also include a headphone jack.

The positioning component 808 is used to locate the current geographic position of the terminal 800 for navigation or LBS (Location Based Service). The Positioning component 808 may be a Positioning component based on the GPS (Global Positioning System) in the united states, the beidou System in china, the graves System in russia, or the galileo System in the european union.

Power supply 809 is used to provide power to various components in terminal 800. The power supply 809 can be ac, dc, disposable or rechargeable. When power source 809 comprises a rechargeable battery, the rechargeable battery can support wired charging or wireless charging. The rechargeable battery may also be used to support fast charge technology.

In some embodiments, the terminal 800 also includes one or more sensors 810. The one or more sensors 810 include, but are not limited to: acceleration sensor 811, gyro sensor 812, pressure sensor 813, fingerprint sensor 814, optical sensor 815 and proximity sensor 816.

The acceleration sensor 811 may detect the magnitude of acceleration in three coordinate axes of the coordinate system established with the terminal 800. For example, the acceleration sensor 811 may be used to detect components of the gravitational acceleration in three coordinate axes. The processor 801 may control the touch screen 805 to display the user interface in a landscape view or a portrait view according to the gravitational acceleration signal collected by the acceleration sensor 811. The acceleration sensor 811 may also be used for acquisition of motion data of a game or a user.

The gyro sensor 812 may detect a body direction and a rotation angle of the terminal 800, and the gyro sensor 812 may cooperate with the acceleration sensor 811 to acquire a 3D motion of the user with respect to the terminal 800. From the data collected by the gyro sensor 812, the processor 801 may implement the following functions: motion sensing (such as changing the UI according to a user's tilting operation), image stabilization at the time of photographing, game control, and inertial navigation.

Pressure sensors 813 may be disposed on the side bezel of terminal 800 and/or underneath touch display 805. When the pressure sensor 813 is disposed on the side frame of the terminal 800, the holding signal of the user to the terminal 800 can be detected, and the processor 801 performs left-right hand recognition or shortcut operation according to the holding signal collected by the pressure sensor 813. When the pressure sensor 813 is disposed at the lower layer of the touch display screen 805, the processor 801 controls the operability control on the UI interface according to the pressure operation of the user on the touch display screen 805. The operability control comprises at least one of a button control, a scroll bar control, an icon control, and a menu control.

The fingerprint sensor 814 is used for collecting a fingerprint of the user, and the processor 801 identifies the identity of the user according to the fingerprint collected by the fingerprint sensor 814, or the fingerprint sensor 814 identifies the identity of the user according to the collected fingerprint. Upon identifying that the user's identity is a trusted identity, the processor 801 authorizes the user to perform relevant sensitive operations including unlocking a screen, viewing encrypted information, downloading software, paying for and changing settings, etc. Fingerprint sensor 814 may be disposed on the front, back, or side of terminal 800. When a physical button or a vendor Logo is provided on the terminal 800, the fingerprint sensor 814 may be integrated with the physical button or the vendor Logo.

The optical sensor 815 is used to collect ambient light intensity. In one embodiment, the processor 801 may control the display brightness of the touch screen 805 based on the ambient light intensity collected by the optical sensor 815. Specifically, when the ambient light intensity is high, the display brightness of the touch display screen 805 is increased; when the ambient light intensity is low, the display brightness of the touch display 805 is turned down. In another embodiment, the processor 801 may also dynamically adjust the shooting parameters of the camera assembly 806 according to the ambient light intensity collected by the optical sensor 815.

A proximity sensor 816, also known as a distance sensor, is typically provided on the front panel of the terminal 800. The proximity sensor 816 is used to collect the distance between the user and the front surface of the terminal 800. In one embodiment, when the proximity sensor 816 detects that the distance between the user and the front surface of the terminal 800 gradually decreases, the processor 801 controls the touch display 805 to switch from the bright screen state to the dark screen state; when the proximity sensor 816 detects that the distance between the user and the front surface of the terminal 800 becomes gradually larger, the processor 801 controls the touch display 805 to switch from the screen-on state to the screen-on state.

Those skilled in the art will appreciate that the configuration shown in fig. 8 is not intended to be limiting of terminal 800 and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components may be used.

Fig. 9 is a schematic structural diagram of a server provided in an embodiment of the present disclosure, where the server 900 may have a relatively large difference due to different configurations or performances, and may include one or more processors (CPUs) 901 and one or more memories 902, where the memory 902 stores at least one program code, and the at least one program code is loaded and executed by the processors 901 to implement the resource search method provided in the foregoing embodiments. Certainly, the server 900 may also have components such as a wired or wireless network interface, a keyboard, and an input/output interface, so as to perform input and output, and the server 900 may also include other components for implementing device functions, which are not described herein again.

In an exemplary embodiment, a computer-readable storage medium comprising at least one instruction, e.g., a memory comprising at least one instruction, is also provided, the at least one instruction being executable by a processor in a computer device to perform the resource search method in the above embodiments. Alternatively, the computer-readable storage medium may be a non-transitory computer-readable storage medium, and the non-transitory computer-readable storage medium may include a ROM (Read-Only Memory), a RAM (Random-Access Memory), a CD-ROM (Compact Disc Read-Only Memory), a magnetic tape, a floppy disk, an optical data storage device, and the like, for example.

In an exemplary embodiment, a computer program product is also provided, which includes one or more instructions executable by a processor of a computer device to perform the resource search method provided by the various embodiments described above.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for resource search, comprising:

in response to a first search request generated based on a triggering operation on a content tag, acquiring the content tag from a search word field of the first search request;

acquiring a resource identifier from a resource identifier field of the first search request, and determining a multimedia resource indicated by the resource identifier as a first multimedia resource, wherein the content tag is displayed in a playing interface of the first multimedia resource;

embedding the content tag to obtain search term characteristics;

inputting the detail information of the first multimedia resource and the key frame of the first multimedia resource into a feature extraction model, wherein the feature extraction model comprises a word vector submodel and a Deep Neural Network (DNN) submodel, and the detail information comprises a text consisting of at least one of a title, a brief introduction, a summary, a nickname of a publisher, an account number of the publisher or a content label of the first multimedia resource;

extracting and obtaining the embedded characteristics of the detail information through the word vector submodel;

extracting the image characteristics of the key frame through the DNN submodel;

splicing the embedded feature of the detail information and the image feature of the key frame, inputting the spliced feature into a full-connection layer, and extracting a first resource feature of the first multimedia resource through the full-connection layer, wherein the first resource feature is a fusion feature between the detail information and the key frame;

for any candidate multimedia resource in a plurality of candidate multimedia resources, acquiring a first similarity between the search term characteristic and a second resource characteristic of the candidate multimedia resource;

multiplying the first similarity by a first coefficient to obtain a first numerical value, wherein the first coefficient represents the proportion of the first similarity in a third similarity;

multiplying the second similarity by a second coefficient to obtain a second numerical value, wherein the second coefficient represents the proportion of the second similarity in the third similarity, and the sum of the second coefficient and the first coefficient is equal to 1;

adding the first numerical value and the second numerical value to obtain the third similarity;

determining the candidate multimedia resource as a third multimedia resource under the condition that the third similarity meets a first target condition;

and sending at least one second multimedia resource obtained by screening from at least one third multimedia resource, wherein the second multimedia resource is associated with both the content tag and the first multimedia resource.

2. The method of claim 1, further comprising:

responding to a second search request, and acquiring search terms carried by the second search request, wherein the second search request is generated based on input operation in a search input box;

3. The method of claim 2, wherein selecting the multimedia resource associated with the search term from the historical browsing records as the first multimedia resource comprises:

4. The method of claim 2, wherein selecting the multimedia resource associated with the search term from the historical browsing records as the first multimedia resource comprises:

5. The method of claim 1, wherein determining the candidate multimedia resource as a third multimedia resource if the third similarity meets the first target condition comprises:

6. The method of claim 1, wherein the determining the candidate multimedia resource as a third multimedia resource if the third similarity meets the first target condition comprises:

7. The method of claim 1, further comprising:

8. The method according to claim 7, wherein said obtaining the target behavior parameter of the at least one third multimedia resource comprises:

9. The method according to claim 7, wherein the obtaining, from the at least one third multimedia resource, a third multimedia resource whose target behavior parameter meets a second target condition as the at least one second multimedia resource comprises:

10. The method according to claim 7, wherein the obtaining, from the at least one third multimedia resource, a third multimedia resource whose target behavior parameter meets a second target condition as the at least one second multimedia resource comprises:

11. The method of claim 7, wherein the target behavior parameters comprise: at least one of a click behavior parameter, a like behavior parameter, a forward behavior parameter, a comment behavior parameter, or a collection behavior parameter.

12. A method for resource search, comprising:

displaying a plurality of content tags of a first multimedia resource in a playing interface of the first multimedia resource;

responding to a triggering operation of any content label in the plurality of content labels, and sending a first search request carrying the content label and a resource identifier of the first multimedia resource, wherein the content label is packaged in a search word field of the first search request, and the resource identifier is packaged in a resource identifier field of the first search request;

receiving at least one second multimedia resource returned based on the first search request, wherein the second multimedia resource is associated with both the content tag and the first multimedia resource, the second multimedia resource is obtained by screening from at least one third multimedia resource, the third multimedia resource is a candidate multimedia resource with a third similarity meeting a first target condition, the third similarity is obtained by adding a first numerical value and a second numerical value, the first numerical value is a numerical value obtained by multiplying the first similarity by a first coefficient, the second numerical value is a numerical value obtained by multiplying the second similarity by a second coefficient, the first coefficient represents the proportion of the first similarity in the third similarity, the second coefficient represents the proportion of the second similarity in the third similarity, and the sum of the second coefficient and the first coefficient is equal to 1, the first similarity is the similarity between a search word feature and a second resource feature of the candidate multimedia resource, the second similarity is the similarity between a first resource feature and the second resource feature of the candidate multimedia resource, the search word feature is obtained based on embedding processing of the content tag, the first resource feature is a fusion feature between detail information of the first multimedia resource and a key frame of the first multimedia resource, the first resource feature is obtained through feature extraction through a feature extraction model based on the detail information and the key frame, the feature extraction model comprises a word vector sub-model and a deep neural network DNN sub-model, the word vector sub-model is used for extracting the embedding feature of the detail information, the DNN sub-model is used for extracting the image feature of the key frame, and the first resource feature is the embedding feature of a full connection layer on the detail information and the image feature splicing of the key frame Extracting the features, wherein the detailed information comprises a text consisting of at least one of a title, a brief introduction, an abstract, a nickname of a publisher, an account number of the publisher or a content label of the first multimedia resource;

and displaying the at least one second multimedia resource.

13. A resource search apparatus, comprising:

a first determination unit configured to execute, in response to a first search request generated based on a trigger operation on a content tag, acquiring the content tag from a search word field of the first search request; acquiring a resource identifier from a resource identifier field of the first search request, and determining a multimedia resource indicated by the resource identifier as a first multimedia resource, wherein the content tag is displayed in a playing interface of the first multimedia resource;

the acquisition unit is configured to perform embedding processing on the content tag to obtain search term characteristics; inputting the detail information of the first multimedia resource and the key frame of the first multimedia resource into a feature extraction model, wherein the feature extraction model comprises a word vector submodel and a Deep Neural Network (DNN) submodel, and the detail information comprises a text consisting of at least one of a title, a brief introduction, a summary, a publisher nickname, a publisher account number or a content label of the first multimedia resource; extracting the embedded characteristics of the detail information through the word vector submodel; extracting the image characteristics of the key frame through the DNN submodel; splicing the embedded feature of the detail information and the image feature of the key frame, inputting the spliced feature into a full-connection layer, and extracting a first resource feature of the first multimedia resource through the full-connection layer, wherein the first resource feature is a fusion feature between the detail information and the key frame;

the second determining unit comprises a determining subunit, wherein the determining subunit comprises an acquiring subunit, a weighting subunit and a determining subunit;

the obtaining subunit is configured to perform obtaining, for any candidate multimedia resource of a plurality of candidate multimedia resources, a first similarity between the search term feature and a second resource feature of the candidate multimedia resource; obtaining a second similarity between the first resource feature and a second resource feature of the candidate multimedia resource;

the weighting subunit is configured to perform multiplication on the first similarity and a first coefficient to obtain a first numerical value, wherein the first coefficient represents a proportion of the first similarity in a third similarity; multiplying the second similarity by a second coefficient to obtain a second numerical value, wherein the second coefficient represents the proportion of the second similarity in the third similarity, and the sum of the second coefficient and the first coefficient is equal to 1; adding the first numerical value and the second numerical value to obtain the third similarity;

a determining subunit configured to perform, in a case that the third similarity meets a first target condition, determining that the candidate multimedia resource is a third multimedia resource;

a sending unit configured to perform sending of at least one second multimedia resource filtered from at least one third multimedia resource, the second multimedia resource being associated with both the content tag and the first multimedia resource.

14. The apparatus of claim 13, wherein the first determining unit comprises:

the first obtaining subunit is configured to perform, in response to a second search request, obtaining the search word carried in the second search request, where the second search request is generated based on an input operation in a search input box;

15. The apparatus according to claim 14, wherein the selection subunit is configured to perform:

16. The apparatus according to claim 14, wherein the selection subunit is configured to perform:

17. The apparatus of claim 13, wherein the determining subunit is configured to perform:

18. The apparatus of claim 13, wherein the determining subunit is configured to perform:

19. The apparatus of claim 13, wherein the second determining unit further comprises:

20. The apparatus of claim 19, wherein the second obtaining subunit is configured to perform:

21. The apparatus of claim 19, wherein the third obtaining subunit is configured to perform:

22. The apparatus of claim 19, wherein the third obtaining subunit is configured to perform:

23. The apparatus of claim 19, wherein the target behavior parameters comprise: at least one of a click behavior parameter, a like behavior parameter, a forward behavior parameter, a comment behavior parameter, or a collection behavior parameter.

24. A resource search apparatus, comprising:

the device comprises a sending unit, a display unit and a display unit, wherein the sending unit is configured to display a plurality of content labels of a first multimedia resource in a playing interface of the first multimedia resource; responding to a triggering operation of any content label in the plurality of content labels, and sending a first search request carrying the content label and a resource identifier of the first multimedia resource; wherein the content tag is encapsulated in a search term field of the first search request, and the resource identifier is encapsulated in a resource identifier field of the first search request;

a receiving unit configured to perform receiving at least one second multimedia resource returned based on the first search request, the second multimedia resource being associated with both the content tag and the first multimedia resource, the second multimedia resource being obtained by filtering from at least one third multimedia resource, the third multimedia resource being a candidate multimedia resource whose third similarity meets the first target condition, the third similarity being obtained by adding a first value and a second value, the first value being a value obtained by multiplying the first similarity by a first coefficient, the second value being a value obtained by multiplying the second similarity by a second coefficient, the first coefficient representing a proportion of the first similarity in the third similarity, the second coefficient representing a proportion of the second similarity in the third similarity, the sum of the second coefficient and the first coefficient is equal to 1, the first similarity is a similarity between a search word feature and a second resource feature of the candidate multimedia resource, the second similarity is a similarity between a first resource feature and a second resource feature of the candidate multimedia resource, the search word feature is obtained based on the embedding processing of the content tag, the first resource feature is a fusion feature between the detail information of the first multimedia resource and a key frame of the first multimedia resource, the first resource feature is obtained by extracting through a feature extraction model based on the detail information and the key frame, the feature extraction model comprises a word vector sub-model and a Deep Neural Network (DNN) sub-model, the word vector sub-model is used for extracting the embedding feature of the detail information, and the DNN sub-model is used for extracting the image feature of the key frame, the first resource characteristics are obtained by extracting characteristics of the full connection layer after splicing the embedded characteristics of the detail information and the image characteristics of the key frame, and the detail information comprises a text consisting of at least one of a title, a brief introduction, an abstract, a publisher nickname, a publisher account number or a content label of the first multimedia resource;

a display unit configured to perform displaying the at least one second multimedia resource.

25. A computer device, comprising:

one or more processors;

wherein the one or more processors are configured to execute the instructions to implement the resource search method of any one of claim 1 to claim 11 or claim 12.

26. A computer-readable storage medium having at least one instruction thereon that, when executed by one or more processors of a computer device, enable the computer device to perform the resource search method of any one of claims 1-11 or 12.