CN110895467A

CN110895467A - Method and device for updating search model, storage medium and electronic device

Info

Publication number: CN110895467A
Application number: CN201811070483.7A
Authority: CN
Inventors: 乔宝琛; 张晓泉; 程昊; 詹焯扬; 蔡健; 袁子斌; 李文文; 邬龙; 江涛; 杨妤卿
Original assignee: Shenzhen Blue Lantern Fish Intelligent Technology Co ltd
Current assignee: Beijing Blue lantern fish Intelligent Technology Co.,Ltd.
Priority date: 2018-09-13
Filing date: 2018-09-13
Publication date: 2020-03-20

Abstract

The invention discloses a method and a device for updating a retrieval model, a storage medium and an electronic device. Wherein, the method comprises the following steps: acquiring a retrieval request, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database; acquiring a retrieval result corresponding to the retrieval request; acquiring operation information generated by executing target operation on a retrieval result, wherein the operation information is used for indicating the similarity between the retrieval result and a retrieval request; and sending the retrieval request, the retrieval result and the operation information with the corresponding relation to the server so that the server updates the retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result. The invention solves the technical problem of low updating efficiency of the retrieval model in the related technology.

Description

Method and device for updating search model, storage medium and electronic device

Technical Field

The invention relates to the field of computers, in particular to a method and a device for updating a retrieval model, a storage medium and an electronic device.

Background

The existing search engine cannot fully and effectively utilize data fed back by a user, firstly, the existing technology is high in cost, a large amount of manpower and material resources are spent to obtain experience feedback of the user in a questionnaire form, and the questionnaire form is difficult to reflect real ideas of the user. On the other hand, the random format feedback data cannot directly reflect the defects on a certain query result, but is a macroscopic overview, so that an algorithm engineer is difficult to find the defects of the artificial intelligence algorithm engine according to the user feedback of the overview.

In view of the above problems, no effective solution has been proposed.

Disclosure of Invention

The embodiment of the invention provides a method and a device for updating a retrieval model, a storage medium and an electronic device, which are used for at least solving the technical problem of low updating efficiency of the retrieval model in the related technology.

According to an aspect of the embodiments of the present invention, there is provided a method for updating a search model, including: acquiring a retrieval request, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database; acquiring a retrieval result corresponding to the retrieval request; acquiring operation information generated by executing target operation on the retrieval result, wherein the operation information is used for indicating the similarity between the retrieval result and the retrieval request; and sending the retrieval request, the retrieval result and the operation information with the corresponding relation to a server so that the server updates a retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result.

Optionally, the obtaining operation information generated by performing the target operation on the retrieval result includes: detecting the target operation executed on the retrieval result; and acquiring the operation information generated by the target operation.

Optionally, the target operation performed on the retrieval result is detected to include at least one of: displaying one or more indication information corresponding to the retrieval result, wherein the one or more indication information is used for indicating the satisfaction degree of the account sending the retrieval request to the retrieval result; detecting a first selected operation performed on target indication information in the one or more indication information; detecting an input operation performed on the retrieval result, wherein the input operation is used for inputting the similarity between the retrieval result and the retrieval request; detecting a second selected operation executed on the retrieval result; detecting a sorting operation executed on a plurality of retrieval data included in the retrieval result, wherein the sorting operation is used for indicating the display sequence of the plurality of retrieval data.

Optionally, the obtaining the operation information generated by the target operation includes: in the case of detecting a first selected operation performed on target indication information in the one or more indication information, acquiring target satisfaction indicated by the target indication information as the operation information in response to the first selected operation; under the condition that the input operation executed on the retrieval result is detected, acquiring the target similarity corresponding to the input operation as the operation information; under the condition that a second selected operation executed on the retrieval result is detected, responding to the second selected operation to display a result page corresponding to the retrieval result; acquiring display time of the result page and/or a selection rate of the retrieval result as the operation information, wherein the selection rate is used for indicating the number of times that the retrieval result is executed with the second selection operation; in a case where a sorting operation performed on a plurality of pieces of retrieval data included in the retrieval result is detected, displaying the plurality of retrieval results in the display order in response to the sorting operation; and acquiring an identification list of the plurality of retrieval data arranged according to the display sequence as the operation information.

Optionally, the search result includes one or more search data, where obtaining operation information generated by performing a target operation on the search result includes: detecting the target operation performed on each of the one or more search data; acquiring target operation information corresponding to each target operation; determining each of the retrieval data and each of the target operation information having a correspondence relationship as the operation information.

Optionally, after sending the retrieval request, the retrieval result, and the operation information having the corresponding relationship to a server, the method further includes: acquiring the retrieval request again; sending the retrieval request to the server so that the server obtains a target retrieval result corresponding to the retrieval request by using the target retrieval model; and acquiring and displaying the target retrieval result.

According to another aspect of the embodiments of the present invention, there is also provided a method for updating a search model, including: inputting a received retrieval request into a retrieval model to obtain a retrieval result, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database, and the retrieval model is obtained by training an initial model by using a request sample and a result sample which have a corresponding relation; responding to the retrieval request and sending the retrieval result to the client; receiving the retrieval request, the retrieval result and the operation information which are sent by the client and have corresponding relations; and updating the retrieval model by using the retrieval request, the retrieval result and the operation information which have the corresponding relation to obtain a target retrieval model.

Optionally, after updating the retrieval model by using the retrieval request, the retrieval result, and the operation information with a corresponding relationship, to obtain a target retrieval model, the method further includes: receiving the retrieval request again; inputting the retrieval request into the target retrieval model to obtain a target retrieval result; and responding to the retrieval request and sending the target retrieval result to the client.

Optionally, updating the retrieval model by using the retrieval request, the retrieval result, and the operation information having the corresponding relationship, and obtaining a target retrieval model includes: converting the retrieval request, the retrieval result and the operation information which have corresponding relations into training data in a target format, wherein the target format is the format of an input parameter of the retrieval model; and training the retrieval model by using the training data to obtain the target retrieval model.

Optionally, the retrieval result includes a plurality of retrieval data, the operation information includes an identification list of the plurality of retrieval data, the identification list is used to indicate the plurality of retrieval data arranged in a display order indicated by a sorting operation detected by the client, and after receiving the retrieval request, the retrieval result and the operation information with corresponding relationship sent by the client, the method further includes: and training a ranking model by using the retrieval request, the retrieval result and the identification list of the plurality of retrieval data with the corresponding relation to obtain a target ranking model, wherein the target ranking model is used for ranking the results output by the target retrieval model.

According to another aspect of the embodiments of the present invention, there is also provided an updating apparatus for a search model, including: the system comprises a first acquisition module, a first storage module and a second acquisition module, wherein the first acquisition module is used for acquiring a retrieval request which is used for requesting to retrieve data meeting target conditions from a database; the second acquisition module is used for acquiring a retrieval result corresponding to the retrieval request; a third obtaining module, configured to obtain operation information generated by performing a target operation on the search result, where the operation information is used to indicate a similarity between the search result and the search request; the first sending module is used for sending the retrieval request, the retrieval result and the operation information with the corresponding relation to a server so that the server updates a retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result.

According to another aspect of the embodiments of the present invention, there is also provided an updating apparatus for a search model, including: the system comprises a first input module, a retrieval module and a second input module, wherein the first input module is used for inputting a received retrieval request into a retrieval model to obtain a retrieval result, the retrieval request is used for requesting to retrieve data meeting target conditions from a database, and the retrieval model is obtained by training an initial model by using a request sample and a result sample which have a corresponding relation; the third sending module is used for responding to the retrieval request and sending the retrieval result to the client; the first receiving module is used for receiving the retrieval request, the retrieval result and the operation information which are sent by the client and have corresponding relations; and the updating module is used for updating the retrieval model by utilizing the retrieval request, the retrieval result and the operation information which have the corresponding relation to obtain a target retrieval model.

According to another aspect of the embodiments of the present invention, there is also provided a storage medium, characterized in that the storage medium stores therein a computer program, wherein the computer program is configured to execute the method described in any one of the above when executed.

According to another aspect of the embodiments of the present invention, there is also provided an electronic apparatus, including a memory and a processor, wherein the memory stores therein a computer program, and the processor is configured to execute the method described in any one of the above through the computer program.

In the embodiment of the invention, a retrieval request is obtained, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database; acquiring a retrieval result corresponding to the retrieval request; acquiring operation information generated by executing target operation on a retrieval result, wherein the operation information is used for indicating the similarity between the retrieval result and a retrieval request; sending the search request, the search result and the operation information with the corresponding relation to the server so that the server updates the search model by using the corresponding relation to obtain a target search model, wherein, the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result, obtaining a retrieval result according to the retrieval request, obtaining operation information generated by executing a target operation on the retrieval result, associating the operation information with the retrieval request and the retrieval result and sending the association to the server, updating the retrieval model by the server by using the corresponding relation, thereby obtaining an updated more optimized target retrieval model by using the operation information generated on the retrieval result, therefore, the technical effect of improving the updating efficiency of the retrieval model is achieved, and the technical problem of low updating efficiency of the retrieval model in the related technology is solved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

FIG. 1 is a schematic diagram of an alternative method for updating a search model according to an embodiment of the invention;

FIG. 2 is a schematic diagram of an alternative method for updating a search model according to an embodiment of the invention;

FIG. 3 is a schematic diagram of an alternative apparatus for updating a search model according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of an alternative updating apparatus for search models according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of an application scenario of an alternative method for updating a search model according to an embodiment of the present invention; and

FIG. 6 is a schematic diagram of an alternative electronic device according to an embodiment of the invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above 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 invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

According to an aspect of an embodiment of the present invention, there is provided a method for updating a search model, as shown in fig. 1, the method including:

s102, acquiring a retrieval request, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database;

s104, acquiring a retrieval result corresponding to the retrieval request;

s106, obtaining operation information generated by executing target operation on the retrieval result, wherein the operation information is used for indicating the similarity between the retrieval result and the retrieval request;

and S108, sending the retrieval request, the retrieval result and the operation information with the corresponding relation to the server so that the server updates the retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result.

Optionally, in this embodiment, the above method for updating a search model may be, but is not limited to, applied in a scenario in which a search model is updated during a search process. The search model may be, but is not limited to, a model for search in various types of applications or websites, such as an online education application or website, an instant messaging application or website, a community space application or website, a game application or website, a shopping application or website, a browser application or website, a financial application or website, a multimedia application or website, a live broadcast application or website, an intellectual property search application or website, and the like. Specifically, the method can be applied to, but not limited to, a scene in which the retrieval model is updated in the process of retrieving a patent or a trademark on the intellectual property retrieval website, or can also be applied to, but not limited to, a scene in which the retrieval model is updated in the process of retrieving a video on the multimedia application or the website, so as to improve the updating efficiency of the retrieval model. The above is only an example, and this is not limited in this embodiment.

Optionally, in this embodiment, the above updating method of the search model may be, but is not limited to be, applied to a terminal device. The terminal device may include, but is not limited to, a mobile phone, a tablet computer, a notebook computer, a PC computer, an intelligent wearable device, an intelligent home device, and the like.

Optionally, in this embodiment, the retrieval request may be, but is not limited to, obtained in a form of a retrieval formula, such as: there may be a search formula for representing keywords and relationships between keywords, such as: keyword 1 and keyword 2 and keyword 3, keyword 1 or keyword 2 or keyword 3, and so on.

Optionally, in this embodiment, the obtained retrieval request may be a keyword, a relationship between keywords, a phrase, a sentence, and the like, which are input by the user in the search box, and the user may input only the keyword, which may be but is not limited to a relationship between default keywords and (and), or may also be but is not limited to a relationship between default or (or). The user may also enter relationships between keywords while entering the keywords. Or the user can also input phrases or sentences, and the retrieval model extracts keywords from the received phrases or sentences for retrieval.

Optionally, in this embodiment, the client installed on the terminal device may, but is not limited to, send the acquired search request to the server, and the server searches the search request by using the search model to obtain a search result, and then returns the search result to the client, so as to acquire a search result corresponding to the search request.

Optionally, in the present embodiment, the target operation may include, but is not limited to, any operation performed on the retrieval result. Such as: click operation, drag operation, browse operation, evaluation operation, and the like performed by the user on the search result.

In an optional embodiment, taking patent retrieval as an example, a patent retrieval client is installed on a terminal device, a user opens the client, inputs content to be retrieved in a search box, the client generates a retrieval request according to the received user input and sends the retrieval request to a server, the server inputs the retrieval request to a retrieval model to obtain a retrieval result, the retrieval result is returned to the client, the client obtains the retrieval result and displays the retrieval result on a retrieval result interface, at this time, the user can operate the searched retrieval result, the client obtains operation information generated by target operation performed on the retrieval result, establishes a corresponding relation among the retrieval request, the retrieval result and the operation information, sends the corresponding relation to the server, the server updates the retrieval model by using the corresponding relation to obtain a target retrieval model, thereby realizing the optimization of the retrieval model.

Therefore, through the steps, the retrieval result is obtained according to the retrieval request, the operation information generated by executing the target operation on the retrieval result is obtained, the operation information is associated with the retrieval request and the retrieval result and is sent to the server, the server updates the retrieval model by using the corresponding relation, and the updated more optimized target retrieval model can be obtained by using the operation information generated on the retrieval result, so that the technical effect of improving the updating efficiency of the retrieval model is realized, and the technical problem of lower updating efficiency of the retrieval model in the related technology is solved.

As an alternative, the obtaining operation information generated by performing the target operation on the search result includes:

s1, detecting the target operation executed on the retrieval result;

s2, operation information generated by the target operation is acquired.

Optionally, in this embodiment, the target operation performed on the retrieval result may be, but is not limited to, detected by at least one of the following ways:

the method comprises the steps that one or more pieces of indication information used for indicating the satisfaction degree of an account sending a retrieval request on a retrieval result is displayed at the position of each retrieval result displayed by a client, a user can select the indication information to indicate the satisfaction degree of the user on the retrieval result, the client detects a first selection operation executed on the one or more pieces of indication information, and obtains operation information generated by the first selection operation. For example: displaying one or more indication information corresponding to the retrieval result, wherein the one or more indication information is used for indicating the satisfaction degree of the account sending the retrieval request to the retrieval result; a first selected operation performed on a target indication information of the one or more indication information is detected.

In an alternative embodiment, two pieces of indication information, "step" and "like" are displayed at the position of each retrieval result, wherein the indication information "step" is used for indicating that the satisfaction degree of the account sending the retrieval request on the retrieval result is low, and the indication information "like" is used for indicating that the satisfaction degree of the account sending the retrieval request on the retrieval result is high. The satisfaction of "tramp" and "like" can be quantified, for example: "stepping" corresponds to a satisfaction of 0 or 0% or 10%, etc., and "like" corresponds to a satisfaction of 1 or 100% or 90%, etc.

In another alternative embodiment, a drop-down list of indication information is displayed at the position of each search result, and the list includes satisfaction values from 1% to 100%.

In the second mode, an input box for inputting the similarity is displayed at the position of each search result displayed by the client, and the detection of the target operation is determined by detecting the input operation performed on the input box. The user can input the similarity between the search result and the search request which the user thinks in the input box. For example: detecting an input operation executed on a retrieval result, wherein the input operation is used for inputting the similarity between the retrieval result and a retrieval request;

and the client side can acquire operation information generated by the second selected operation by detecting the second selected operation of the user on the retrieval result. For example: the detection of the target operation may be determined upon detection of a second selected operation performed on the retrieval result.

In a fourth mode, the user may reorder the search results according to the similarity between the search results and the search request, and obtain the operation information by detecting the operation of ordering the search data included in the search results. For example: and detecting a sorting operation executed on a plurality of retrieval data included in the retrieval result, wherein the sorting operation is used for indicating the display sequence of the plurality of retrieval data.

Optionally, in this embodiment, when a different target operation is detected, the operation information may be obtained, but is not limited to, by:

in a case where a first selected operation performed on target indication information of the one or more indication information is detected, a target satisfaction indicated by the target indication information is acquired as operation information in response to the first selected operation.

And acquiring the target similarity corresponding to the input operation as operation information when the input operation executed on the retrieval result is detected.

Under the condition that a second selection operation executed on the retrieval result is detected, responding to the second selection operation and displaying a result page corresponding to the retrieval result; and acquiring display time of the result page and/or a selection rate of the retrieval result as operation information, wherein the selection rate is used for indicating the times of executing second selection operation on the retrieval result.

In a case where a sorting operation performed on a plurality of pieces of retrieval data included in the retrieval result is detected, displaying the plurality of retrieval results in a display order in response to the sorting operation; an identification list of a plurality of search data arranged in a display order is acquired as operation information.

As an optional solution, the retrieval result includes one or more retrieval data, and acquiring operation information generated by performing the target operation on the retrieval result includes:

s1, detecting a target operation executed on each retrieval data in the one or more retrieval data;

s2, acquiring target operation information corresponding to each target operation;

s3, each of the search data and each of the target operation information having the correspondence relationship are determined as operation information.

Alternatively, in this embodiment, the retrieval result may include one or more pieces of retrieval data, the user may perform different target operations on each piece of retrieval data to generate target operation information, and the operation information may be determined by establishing a correspondence between each piece of retrieval data and each piece of target operation information.

As an optional scheme, after sending the retrieval request, the retrieval result, and the operation information having the corresponding relationship to the server, the method further includes:

s1, acquiring the retrieval request again;

s2, sending the search request to the server, so that the server obtains a target search result corresponding to the search request by using the target search model;

and S3, acquiring and displaying the target retrieval result.

Optionally, in this embodiment, after the retrieval request, the retrieval result, and the operation information having the corresponding relationship are sent to the server, the server updates the retrieval model by using the corresponding relationship, and when the retrieval request is obtained again, the retrieval request is sent to the server, so that the server can use the updated target retrieval model to perform a retrieval operation on the retrieval request, and obtain a more optimized target retrieval result that is more similar to the retrieval request and better meets the user satisfaction, thereby achieving an effect of optimizing the retrieval result by using the operation of the user on the retrieval result.

According to another aspect of the embodiment of the present invention, there is also provided another method for updating a search model, as shown in fig. 2, the method including:

s202, inputting a received retrieval request into a retrieval model to obtain a retrieval result, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database, and the retrieval model is obtained by training an initial model by using a request sample and a result sample which have a corresponding relation;

s204, responding to the retrieval request and sending the retrieval result to the client;

s206, receiving a retrieval request, a retrieval result and operation information which are sent by a client and have corresponding relations;

and S208, updating the retrieval model by using the retrieval request, the retrieval result and the operation information with the corresponding relation to obtain a target retrieval model.

Optionally, in this embodiment, the above-mentioned updating method of the search model may be, but is not limited to be, applied to a server. The server receives a retrieval request sent by the client, inputs the received retrieval request into the retrieval model to obtain a retrieval result, responds to the retrieval request and sends the retrieval result to the client, receives the retrieval request, the retrieval result and the operation information which are sent by the client and have corresponding relations, and updates the retrieval model by utilizing the retrieval request, the retrieval result and the operation information which have corresponding relations to obtain a target retrieval model.

Optionally, in this embodiment, the client installed on the terminal device may, but is not limited to, send the acquired search request to the server, and the server searches the search request by using the search model to obtain a search result and then returns the search result to the client.

In an optional embodiment, taking patent retrieval as an example, a patent retrieval client is installed on a terminal device, a user opens the client, inputs content to be retrieved in a search box, the client generates a retrieval request according to received user input and sends the retrieval request to a server, the server inputs the retrieval request to a retrieval model to obtain a retrieval result, the retrieval result is returned to the client, the client obtains the retrieval request, the retrieval result and operation information with corresponding relations, the corresponding relations are sent to the server, and the server updates the retrieval model by using the corresponding relations to obtain a target retrieval model, so that optimization of the retrieval model is realized.

As an optional scheme, after the retrieval model is updated by using the retrieval request, the retrieval result, and the operation information with the corresponding relationship, and a target retrieval model is obtained, the method further includes:

s1, receiving the search request again;

s2, inputting the retrieval request into the target retrieval model to obtain a target retrieval result;

and S3, responding to the retrieval request and sending the target retrieval result to the client.

Optionally, in this embodiment, when the search request is received again, the search request is input into the updated target search model to obtain a target search result, and the target search result is returned to the client, so that a target search result more optimized than the previous search result can be obtained after the search request is obtained again, and optimization of the search result by the user operation is also achieved.

As an optional solution, updating the retrieval model by using the retrieval request, the retrieval result, and the operation information having the corresponding relationship, and obtaining the target retrieval model includes:

s1, converting the retrieval request, the retrieval result and the operation information with the corresponding relation into training data with a target format, wherein the target format is the format of the input parameters of the retrieval model;

and S2, training the retrieval model by using the training data to obtain the target retrieval model.

Optionally, in this embodiment, the received search request, the search result, and the operation information having the corresponding relationship may be converted into labeled data conforming to a format of the search model, and the labeled data is used as training data to perform optimization training on the search model.

Optionally, in the present embodiment, the search model may be trained, but not limited to, by metric learning or interactive machine learning.

As an optional scheme, the retrieval result includes a plurality of retrieval data, the operation information includes an identification list of the plurality of retrieval data, the identification list is used for indicating a plurality of retrieval data arranged in a display order indicated by a sorting operation detected by the client, and after receiving a retrieval request, the retrieval result and the operation information with a corresponding relationship sent by the client, the method further includes:

and S1, training the ranking model by using the retrieval request, the retrieval result and the identification list of the plurality of retrieval data with the corresponding relation to obtain a target ranking model, wherein the target ranking model is used for ranking the results output by the target retrieval model.

Optionally, in this embodiment, the acquired operation information may also be used to train a ranking model, so as to achieve ranking of the retrieval results.

It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required by the invention.

According to another aspect of the embodiments of the present invention, there is also provided an updating apparatus of a search model for implementing the above-mentioned method for updating a search model, as shown in fig. 3, the apparatus including:

a first obtaining module 32, configured to obtain a retrieval request, where the retrieval request is used to request that data meeting a target condition be retrieved from a database;

a second obtaining module 34, configured to obtain a retrieval result corresponding to the retrieval request;

a third obtaining module 36, configured to obtain operation information generated by performing a target operation on the search result, where the operation information is used to indicate a similarity between the search result and the search request;

and a first sending module 38, configured to send the search request, the search result, and the operation information with the corresponding relationship to the server, so that the server updates the search model with the corresponding relationship to obtain a target search model, where the search model is used to search the received search request to obtain the search result.

Optionally, in this embodiment, the above updating apparatus for the search model may be applied, but not limited to, in a scenario where the search model is updated during the search process. The search model may be, but is not limited to, a model for search in various types of applications or websites, such as an online education application or website, an instant messaging application or website, a community space application or website, a game application or website, a shopping application or website, a browser application or website, a financial application or website, a multimedia application or website, a live broadcast application or website, an intellectual property search application or website, and the like. Specifically, the method can be applied to, but not limited to, a scene in which the retrieval model is updated in the process of retrieving a patent or a trademark on the intellectual property retrieval website, or can also be applied to, but not limited to, a scene in which the retrieval model is updated in the process of retrieving a video on the multimedia application or the website, so as to improve the updating efficiency of the retrieval model. The above is only an example, and this is not limited in this embodiment.

Optionally, in this embodiment, the above updating apparatus for the search model may be, but is not limited to be, applied to a terminal device. The terminal device may include, but is not limited to, a mobile phone, a tablet computer, a notebook computer, a PC computer, an intelligent wearable device, an intelligent home device, and the like.

Therefore, by the aid of the device, the retrieval result is obtained according to the retrieval request, the operation information generated by executing the target operation on the retrieval result is obtained, the operation information is associated with the retrieval request and the retrieval result and is sent to the server, the server updates the retrieval model by means of the corresponding relation, and the updated more optimized target retrieval model can be obtained by means of the operation information generated on the retrieval result, so that the technical effect of improving the updating efficiency of the retrieval model is achieved, and the technical problem that the updating efficiency of the retrieval model in the related technology is low is solved.

As an optional scheme, the third obtaining module includes:

a first detection unit configured to detect a target operation performed on a retrieval result;

the first acquisition unit is used for acquiring operation information generated by a target operation.

As an alternative, the first detection unit is used for at least one of:

displaying one or more indication information corresponding to the retrieval result, wherein the one or more indication information is used for indicating the satisfaction degree of the account sending the retrieval request to the retrieval result; detecting a first selected operation executed on target indication information in one or more indication information;

detecting an input operation executed on a retrieval result, wherein the input operation is used for inputting the similarity between the retrieval result and a retrieval request;

detecting a second selected operation executed on the retrieval result;

and detecting a sorting operation executed on a plurality of retrieval data included in the retrieval result, wherein the sorting operation is used for indicating the display sequence of the plurality of retrieval data.

Optionally, in this embodiment, one or more pieces of indication information for indicating satisfaction of the account sending the search request with the search result are displayed at a position of each search result displayed by the client, the user may select the indication information to indicate the satisfaction of the user with the search result, and the client detects a first selection operation performed on the one or more pieces of indication information and acquires operation information generated by the first selection operation.

Alternatively, in the present embodiment, an input box for inputting the similarity is displayed at the position of each search result displayed by the client, and the detection of the target operation is determined by detecting the input operation performed on the input box.

Optionally, in this embodiment, the user may select a search result that the user considers satisfactory to perform operations such as browsing and reading, and the client may obtain operation information generated by a second selection operation of the search result by detecting the second selection operation of the user.

Alternatively, in this embodiment, the user may reorder the search results according to the similarity between the search results and the search request that the user thinks, and acquire the operation information by detecting the sorting operation on the search data included in the search results.

As an optional solution, the first obtaining unit is configured to:

in the case of detecting a first selected operation performed on target indication information in the one or more indication information, acquiring target satisfaction indicated by the target indication information as operation information in response to the first selected operation;

under the condition that the input operation executed on the retrieval result is detected, acquiring the target similarity corresponding to the input operation as operation information;

under the condition that a second selection operation executed on the retrieval result is detected, responding to the second selection operation and displaying a result page corresponding to the retrieval result; acquiring display time of a result page and/or a selection rate of a retrieval result as operation information, wherein the selection rate is used for indicating the number of times that the retrieval result is executed with second selection operation;

As an optional scheme, the retrieval result includes one or more retrieval data, where the third obtaining module includes:

a second detection unit configured to detect a target operation performed on each of the one or more retrieval data;

the second acquisition unit is used for acquiring target operation information corresponding to each target operation;

a determining unit configured to determine each of the retrieval data and each of the target operation information having the correspondence relationship as operation information.

As an optional solution, the apparatus further includes:

the fourth acquisition module is used for acquiring the retrieval request again;

the second sending module is used for sending the retrieval request to the server so that the server obtains a target retrieval result corresponding to the retrieval request by using the target retrieval model;

and the fifth acquisition module is used for acquiring and displaying the target retrieval result.

According to another aspect of the embodiments of the present invention, there is also provided another apparatus for updating a search model, for implementing the above method for updating a search model, as shown in fig. 4, the apparatus includes:

a first input module 42, configured to input a received search request into a search model to obtain a search result, where the search request is used to request to search data satisfying a target condition from a database, and the search model is obtained by training an initial model using a request sample and a result sample having a corresponding relationship;

a third sending module 44, configured to send a retrieval result to the client in response to the retrieval request;

a first receiving module 46, configured to receive a retrieval request, a retrieval result, and operation information that are sent by a client and have a corresponding relationship;

and the updating module 48 is used for updating the retrieval model by using the retrieval request, the retrieval result and the operation information which have the corresponding relation to obtain a target retrieval model.

Alternatively, in this embodiment, the above-mentioned updating apparatus for the search model may be, but is not limited to, applied to a server. The server receives a retrieval request sent by the client, inputs the received retrieval request into the retrieval model to obtain a retrieval result, responds to the retrieval request and sends the retrieval result to the client, receives the retrieval request, the retrieval result and the operation information which are sent by the client and have corresponding relations, and updates the retrieval model by utilizing the retrieval request, the retrieval result and the operation information which have corresponding relations to obtain a target retrieval model.

As an optional solution, the apparatus further includes:

the second receiving module is used for receiving the retrieval request again;

the second input module is used for inputting the retrieval request into the target retrieval model to obtain a target retrieval result;

and the fourth sending module is used for responding to the retrieval request and sending the target retrieval result to the client.

Optionally, in this embodiment, when the search request is received again, the search request is input into the updated target search model to obtain a target search result, and the target search result is returned to the client, so that a target search result more optimized than the previous search result can be obtained after the search request is obtained again, and optimization of the search result by the user operation is also achieved

As an optional solution, the update module includes:

the conversion unit is used for converting the retrieval request, the retrieval result and the operation information which have the corresponding relation into training data in a target format, wherein the target format is the format of the input parameters of the retrieval model;

and the training unit is used for training the retrieval model by using the training data to obtain the target retrieval model.

As an alternative, the search result includes a plurality of search data, the operation information includes an identification list of the plurality of search data, the identification list is used to indicate a plurality of search data arranged in a display order indicated by the sorting operation detected by the client, and the apparatus further includes:

and the training module is used for training the ranking model by using the retrieval request, the retrieval result and the identification list of the plurality of retrieval data with the corresponding relation to obtain a target ranking model, wherein the target ranking model is used for ranking the result output by the target retrieval model.

The application environment of the embodiment of the present invention may refer to the application environment in the above embodiments, but is not described herein again. The embodiment of the invention provides an optional specific application example of the connection method for implementing the real-time communication.

As an alternative embodiment, the above-mentioned updating method of the search model can be applied, but not limited, to a scenario in which the search model is optimized by the user's operation on the search result, as shown in fig. 5. In the scene, taking patent and trademark retrieval as an example, a retrieval engine collects feedback data (namely operation information) of a user as original data for next algorithm optimization, in a text list of patent search, the user can approve and step on a retrieval result given by an artificial intelligent algorithm of the retrieval engine according to own judgment, and meanwhile, on a result list interface of the trademark similarity intelligent algorithm, the user can also score the similarity of the trademark result according to own cognitive judgment result similarity, and all the collected user data are collected through a background database and indicate corresponding retrieval requests and retrieval results.

The results fed back by the user after retrieval are mainly used for two aspects: on one hand, the feature extraction algorithm of patents and trademarks is optimized, according to data fed back by a user, a retrieval engine can introduce the data fed back by the user in the processes of construction of a trademark convolutional neural network, construction of a knowledge graph, feature similarity training and the like, and the accuracy of the system is improved through methods such as metric learning and interactive machine learning. On the other hand, in the final result sorting optimization, according to the data fed back by the user, the retrieval engine can further optimize the final retrieved result according to the expected algorithm of the user, so that the user experience is improved.

Through the result feedback of the user to the trademark picture query result and the character search result, the retrieval engine can collect the feedback marking data of each record of stepping or praise of the user to the patent result, and meanwhile, the operation and evaluation of the user are collected in the trademark retrieval result, so that the user requirements can be better understood. The secondary data fed back by the user will be the main data source and reference for the search engine optimization algorithm.

The search engine can improve the model of the existing artificial intelligent search engine by using a metric learning and interactive machine learning method, and simultaneously train a result sorting algorithm by using a neural network according to the use of a large number of users, so that the final sorting of the existing search engine is optimized, and the final experience of the users is optimized.

Optionally, in this embodiment, when the user completes the search task and browses the search result of the search engine through artificial intelligence, and screens out a patent or trademark record that is useful for the user through the user's own cognitive judgment, a corresponding click behavior may be performed at the front end of the search engine website. The search engine optimizes the algorithm by collecting the information generated by these click behaviors.

Optionally, in this embodiment, for the patent retrieval result, the operation of praise and trample on the retrieved document is supported, and the praise and trample behavior data will be introduced into the building of the next stage model, so as to influence the retrieval result and the file ranking when the same file is retrieved next time.

For the trademark retrieval result, after retrieval, user operations (like/step on) are collected, and the ranking of the retrieval result list is influenced finally. Meanwhile, the retrieval engine can optimize the retrieval result by recording the behavior data clicked by the user.

Optionally, in this embodiment, the user may actively click a thumbs-up/thumbs-down button to evaluate a certain search result. The user data may also be recorded passively, and each time the user browses the corresponding search result, the search engine may record the time spent on a certain result and the number of clicks, so as to evaluate the search result.

In an alternative embodiment, the search engine records the click behavior of the user at the front end of the website, specifically including browsing time and browsing times of each search result. In addition, under each retrieval result, the retrieval engine provides a button for pressing and holding, so that the user performs corresponding operation to send corresponding data to the server back-end database.

The website back-end database can determine corresponding data results according to the defined user behaviors, and the feedback results corresponding to the characteristics of each record and the like are used as important original data of the artificial intelligent search engine. The existing search model is then optimized using metric learning or interactive machine learning methods. And finally, on the final result display, the retrieval engine can utilize the marked data to further carry out final sequencing on the retrieval results according to the user requirements, so that the user experience is optimized.

Through the method, the accuracy and the efficiency of the artificial intelligent retrieval engine are adjusted through the measurement learning and interactive machine learning modes, the user experience of trademark and patent retrieval is optimized, and meanwhile, the user participates in the promotion of the algorithm.

According to still another aspect of the embodiments of the present invention, there is also provided an electronic apparatus for implementing the above-described updating of the search model, as shown in fig. 6, the electronic apparatus including: one or more processors 602 (only one of which is shown in the figure), in which a computer program is stored, a memory 604, in which a processor 606 is arranged to carry out the steps of any of the above-described method embodiments by means of the computer program, a sensor 606, an encoder 608 and a transmission means 610.

Optionally, in this embodiment, the electronic apparatus may be located in at least one network device of a plurality of network devices of a computer network.

Optionally, in this embodiment, the processor may be configured to execute the following steps by a computer program:

s1, acquiring a retrieval request, wherein the retrieval request is used for requesting to retrieve data meeting the target conditions from the database;

s2, obtaining a retrieval result corresponding to the retrieval request;

s3, obtaining operation information generated by executing target operation on the retrieval result, wherein the operation information is used for indicating the similarity between the retrieval result and the retrieval request;

and S4, sending the retrieval request, the retrieval result and the operation information with the corresponding relation to the server so that the server updates the retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result.

s1, inputting the received retrieval request into a retrieval model to obtain a retrieval result, wherein the retrieval request is used for requesting to retrieve data meeting the target condition from a database, and the retrieval model is obtained by training an initial model by using a request sample and a result sample which have a corresponding relation;

s2, responding to the retrieval request and sending the retrieval result to the client;

s3, receiving a retrieval request, a retrieval result and operation information which are sent by a client and have corresponding relations;

and S4, updating the retrieval model by using the retrieval request, the retrieval result and the operation information with the corresponding relation to obtain a target retrieval model.

Alternatively, it can be understood by those skilled in the art that the structure shown in fig. 6 is only an illustration, and the electronic device may also be a terminal device such as a smart phone (e.g., an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, a Mobile Internet Device (MID), a PAD, and the like. Fig. 6 is a diagram illustrating a structure of the electronic device. For example, the electronic device may also include more or fewer components (e.g., network interfaces, display devices, etc.) than shown in FIG. 6, or have a different configuration than shown in FIG. 6.

The memory 602 may be used to store software programs and modules, such as program instructions/modules corresponding to the method and apparatus for updating a search model in the embodiment of the present invention, and the processor 604 executes various functional applications and data processing by executing the software programs and modules stored in the memory 602, that is, implementing the control method of the target component described above. The memory 602 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 602 may further include memory located remotely from the processor 604, which may be connected to the terminal over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission device 610 is used for receiving or transmitting data via a network. Examples of the network may include a wired network and a wireless network. In one example, the transmission device 610 includes a network adapter (NIC) that can be connected to a router via a network cable and other network devices to communicate with the internet or a local area network. In one example, the transmission device 610 is a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.

In particular, the memory 602 is used for storing application programs.

Embodiments of the present invention also provide a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the steps of any of the above method embodiments when executed.

Alternatively, in the present embodiment, the storage medium may be configured to store a computer program for executing the steps of:

s2, obtaining a retrieval result corresponding to the retrieval request;

Optionally, in this embodiment, the storage medium may be further configured to store a computer program for executing the following steps:

Optionally, the storage medium is further configured to store a computer program for executing the steps included in the method in the foregoing embodiment, which is not described in detail in this embodiment.

Alternatively, in this embodiment, a person skilled in the art may understand that all or part of the steps in the methods of the foregoing embodiments may be implemented by a program instructing hardware associated with the terminal device, where the program may be stored in a computer-readable storage medium, and the storage medium may include: flash disks, Read-Only memories (ROMs), Random Access Memories (RAMs), magnetic or optical disks, and the like.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.

The integrated unit in the above embodiments, if implemented in the form of a software functional unit and sold or used as a separate product, may be stored in the above computer-readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing one or more computer devices (which may be personal computers, servers, network devices, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the several embodiments provided in the present application, it should be understood that the disclosed client may be implemented in other manners. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one type of division of logical functions, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. A method for updating a search model, comprising:

acquiring a retrieval request, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database;

acquiring a retrieval result corresponding to the retrieval request;

acquiring operation information generated by executing target operation on the retrieval result, wherein the operation information is used for indicating the similarity between the retrieval result and the retrieval request;

and sending the retrieval request, the retrieval result and the operation information with the corresponding relation to a server so that the server updates a retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result.

2. The method of claim 1, wherein obtaining operation information resulting from performing a target operation on the search result comprises:

detecting the target operation executed on the retrieval result;

and acquiring the operation information generated by the target operation.

3. The method of claim 2, wherein detecting the target operation performed on the search result comprises at least one of:

displaying one or more indication information corresponding to the retrieval result, wherein the one or more indication information is used for indicating the satisfaction degree of the account sending the retrieval request to the retrieval result; detecting a first selected operation performed on target indication information in the one or more indication information;

detecting an input operation performed on the retrieval result, wherein the input operation is used for inputting the similarity between the retrieval result and the retrieval request;

detecting a second selected operation executed on the retrieval result;

detecting a sorting operation executed on a plurality of retrieval data included in the retrieval result, wherein the sorting operation is used for indicating the display sequence of the plurality of retrieval data.

4. The method of claim 3, wherein obtaining the operation information resulting from the target operation comprises:

in the case of detecting a first selected operation performed on target indication information in the one or more indication information, acquiring target satisfaction indicated by the target indication information as the operation information in response to the first selected operation;

under the condition that the input operation executed on the retrieval result is detected, acquiring the target similarity corresponding to the input operation as the operation information;

under the condition that a second selected operation executed on the retrieval result is detected, responding to the second selected operation to display a result page corresponding to the retrieval result; acquiring display time of the result page and/or a selection rate of the retrieval result as the operation information, wherein the selection rate is used for indicating the number of times that the retrieval result is executed with the second selection operation;

in a case where a sorting operation performed on a plurality of pieces of retrieval data included in the retrieval result is detected, displaying the plurality of retrieval results in the display order in response to the sorting operation; and acquiring an identification list of the plurality of retrieval data arranged according to the display sequence as the operation information.

5. The method of claim 1, wherein the search result comprises one or more search data, and wherein obtaining operation information resulting from performing the target operation on the search result comprises:

detecting the target operation performed on each of the one or more search data;

acquiring target operation information corresponding to each target operation;

determining each of the retrieval data and each of the target operation information having a correspondence relationship as the operation information.

6. The method according to claim 1, wherein after the search request, the search result, and the operation information having the correspondence relationship are transmitted to a server, the method further comprises:

acquiring the retrieval request again;

sending the retrieval request to the server so that the server obtains a target retrieval result corresponding to the retrieval request by using the target retrieval model;

and acquiring and displaying the target retrieval result.

7. A method for updating a search model, comprising:

inputting a received retrieval request into a retrieval model to obtain a retrieval result, wherein the retrieval request is used for requesting to retrieve data meeting target conditions from a database, and the retrieval model is obtained by training an initial model by using a request sample and a result sample which have a corresponding relation;

responding to the retrieval request and sending the retrieval result to the client;

receiving the retrieval request, the retrieval result and the operation information which are sent by the client and have corresponding relations;

and updating the retrieval model by using the retrieval request, the retrieval result and the operation information which have the corresponding relation to obtain a target retrieval model.

8. The method according to claim 7, wherein after updating the search model with the search request, the search result, and the operation information having the correspondence relationship, obtaining a target search model, the method further comprises:

receiving the retrieval request again;

inputting the retrieval request into the target retrieval model to obtain a target retrieval result;

and responding to the retrieval request and sending the target retrieval result to the client.

9. The method of claim 7, wherein updating the search model with the search request, the search result, and the operation information having the correspondence relationship, and obtaining a target search model comprises:

converting the retrieval request, the retrieval result and the operation information which have corresponding relations into training data in a target format, wherein the target format is the format of an input parameter of the retrieval model;

and training the retrieval model by using the training data to obtain the target retrieval model.

10. The method according to claim 7, wherein the search result includes a plurality of search data, the operation information includes an identification list of the plurality of search data, the identification list is used to indicate the plurality of search data arranged in a display order indicated by the sorting operation detected by the client, and after receiving the search request, the search result and the operation information with corresponding relationship sent by the client, the method further comprises:

and training a ranking model by using the retrieval request, the retrieval result and the identification list of the plurality of retrieval data with the corresponding relation to obtain a target ranking model, wherein the target ranking model is used for ranking the results output by the target retrieval model.

11. An apparatus for updating a search model, comprising:

the system comprises a first acquisition module, a first storage module and a second acquisition module, wherein the first acquisition module is used for acquiring a retrieval request which is used for requesting to retrieve data meeting target conditions from a database;

the second acquisition module is used for acquiring a retrieval result corresponding to the retrieval request;

a third obtaining module, configured to obtain operation information generated by performing a target operation on the search result, where the operation information is used to indicate a similarity between the search result and the search request;

the first sending module is used for sending the retrieval request, the retrieval result and the operation information with the corresponding relation to a server so that the server updates a retrieval model by using the corresponding relation to obtain a target retrieval model, wherein the retrieval model is used for retrieving the received retrieval request to obtain the retrieval result.

12. An apparatus for updating a search model, comprising:

the system comprises a first input module, a retrieval module and a second input module, wherein the first input module is used for inputting a received retrieval request into a retrieval model to obtain a retrieval result, the retrieval request is used for requesting to retrieve data meeting target conditions from a database, and the retrieval model is obtained by training an initial model by using a request sample and a result sample which have a corresponding relation;

the third sending module is used for responding to the retrieval request and sending the retrieval result to the client;

the first receiving module is used for receiving the retrieval request, the retrieval result and the operation information which are sent by the client and have corresponding relations;

and the updating module is used for updating the retrieval model by utilizing the retrieval request, the retrieval result and the operation information which have the corresponding relation to obtain a target retrieval model.

13. A storage medium, in which a computer program is stored, wherein the computer program is arranged to perform the method of any of claims 1 to 10 when executed.

14. An electronic device comprising a memory and a processor, characterized in that the memory has stored therein a computer program, the processor being arranged to execute the method of any of claims 1 to 10 by means of the computer program.