CN117421466A

CN117421466A - Search method, search device, electronic equipment and storage medium

Info

Publication number: CN117421466A
Application number: CN202311301071.0A
Authority: CN
Inventors: 吴开放; 祖明; 高洁璇; 刘瑞龙; 高扬; 黄春明; 万文奇; 张燕蓟; 陈晓雯; 陈典
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2023-10-09
Filing date: 2023-10-09
Publication date: 2024-01-19

Abstract

The disclosure relates to the field of computers, in particular to the technical fields of artificial intelligence, big data, intelligent search and the like, and particularly relates to a search method, a search device, electronic equipment and a storage medium. The specific implementation scheme is as follows: responding to the current search operation, and initiating a search request carrying the current search word to a server; the display server utilizes the initial search result corresponding to the current search word obtained by the search engine; generating user behavior information in response to a browsing operation for an initial search result so as to initiate a result generation request carrying the user behavior information to a server; the display server utilizes the generated search results corresponding to the current search word obtained by the large language model based on the user behavior information. By adopting the method and the device, the accuracy of the search result can be improved.

Description

Search method, search device, electronic equipment and storage medium

Technical Field

The disclosure relates to the field of computers, in particular to the technical fields of artificial intelligence, big data, intelligent search and the like, and particularly relates to a search method, a search device, electronic equipment and a storage medium.

Background

The search engine is a search technology which uses a specific strategy to search out appointed information from the Internet and feeds the appointed information back to the user according to the user demand and a certain algorithm. With the progress and iteration of computer technology, the search efficiency of a search engine is improved, but accurate search results cannot be provided for users.

Disclosure of Invention

The disclosure provides a search method, a search device, electronic equipment and a storage medium device.

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

responding to the current search operation, and initiating a search request carrying the current search word to a server;

the display server utilizes the initial search result corresponding to the current search word obtained by the search engine;

generating user behavior information in response to a browsing operation for an initial search result so as to initiate a result generation request carrying the user behavior information to a server;

the display server utilizes the generated search results corresponding to the current search word obtained by the large language model based on the user behavior information.

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

receiving a search request carrying a current search word initiated by the terminal equipment in response to the current search operation;

obtaining initial search results aiming at the current search word by utilizing a search engine so as to conveniently send the initial search results to terminal equipment;

after receiving user behavior information generated by the terminal equipment in response to browsing operation aiming at an initial search result, initiating a result generation request carrying the user behavior information;

Based on the user behavior information, the generated search results for the current search word are obtained by using a large language model so as to be transmitted to the terminal equipment.

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

the first request initiating unit is used for responding to the current search operation and initiating a search request carrying the current search word to the server;

the first display unit is used for displaying initial search results corresponding to the current search word, which are obtained by the server through the search engine;

the second request initiating unit is used for responding to the browsing operation aiming at the initial search result and generating user behavior information so as to initiate a result generating request carrying the user behavior information to the server;

and the second display unit is used for displaying the generated search results corresponding to the current search word, which are obtained by the server through the large language model based on the user behavior information.

the first request receiving unit is used for receiving a search request carrying a current search word initiated by the terminal equipment in response to the current search operation;

the searching unit is used for obtaining initial searching results aiming at the current searching words by utilizing a searching engine so as to send the initial searching results to the terminal equipment;

The second request receiving unit is used for receiving a result generation request which is initiated to carry the user behavior information after the terminal equipment generates the user behavior information in response to the browsing operation aiming at the initial search result;

and the result generating unit is used for obtaining the generated search result aiming at the current search word by using the large language model based on the user behavior information so as to send the generated search result to the terminal equipment.

According to another aspect of the present disclosure, there is provided an electronic device including:

at least one processor;

a memory communicatively coupled to the at least one processor;

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the embodiments of the present disclosure.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform a method according to any one of the embodiments of the present disclosure.

According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a method according to any of the embodiments of the present disclosure.

By adopting the method and the device, the accuracy of the search result can be improved.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

fig. 1 is a schematic flow chart of a search method (terminal device side) provided in an embodiment of the disclosure;

fig. 2 is a schematic diagram of a display effect of a search page according to an embodiment of the present disclosure;

FIG. 3 is an illustration of an auxiliary display of initial search results provided by an embodiment of the present disclosure;

FIGS. 4A and 4B are diagrams illustrating a manner in which initial search results may be folded, as provided by embodiments of the present disclosure;

fig. 5 is a schematic flow chart of a search method (server side) provided in an embodiment of the disclosure;

FIG. 6 is an integrity flow auxiliary explanatory diagram of a search method according to an embodiment of the disclosure;

fig. 7 is a schematic view of a scenario of a search method according to an embodiment of the present disclosure;

FIG. 8 is a schematic block diagram of a search device according to an embodiment of the present disclosure;

FIG. 9 is a schematic block diagram of a search device according to an embodiment of the present disclosure;

fig. 10 is a schematic block diagram of an electronic device according to an embodiment of the disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

Currently, search results are typically obtained by integrating full web results in response to user online demand. In terms of use, the input box is usually combined with a search result list, when a user needs to replace a search word, a new search word is input in the input box to obtain a new search result, the new search result is used for replacing an original search result, and the new search word and the new search result list are reloaded, wherein the new search result is irrelevant to the last search result. In the technology, a user interface is generally constructed by adopting a Web technology, so that a user can conveniently input search words, search related results through a Server index library, determine search results to be recommended based on the related results, and display the search results to the user. When a new search term is input, the new search result is obtained in the form of asynchronous Javascript and XML (Asynchronous Javascript And XML, ajax) by constructing a new uniform resource locator (Uniform Resource Locator, URL) and refreshing the page or by Javascript, the original search result is removed, and finally, the new search result is rendered.

In the above way, on one hand, the search result is obtained by integrating the whole network result, so that the search range is wider; on the other hand, when the user inputs the search word, the user usually needs to input the search word with definite semantic meaning and needs to have certain search word editing skills to obtain a relatively accurate search result, but in general, the user often does not have the capabilities, so the conventional search method cannot provide the accurate search result for the user, and particularly cannot provide the accurate search result for the user when meeting the complex search requirement or the middle-long tail search requirement.

Based on the above study, the embodiments of the present disclosure provide a search method that can be applied to an electronic device, for example, a terminal device. A search method provided by the embodiment of the present disclosure will be described below with reference to a flowchart shown in fig. 1. It should be noted that although a logical order is illustrated in the flowchart, in some cases, the steps illustrated or described may be performed in other orders.

Step S101, responding to the current searching operation, and initiating a searching request carrying the current searching word to a server;

Step S102, the display server utilizes the initial search result corresponding to the current search word obtained by the search engine;

step S103, generating user behavior information in response to a browsing operation for an initial search result so as to initiate a result generation request carrying the user behavior information to a server;

step S104, the display server uses the large language model to obtain the generated search result corresponding to the current search word based on the user behavior information.

The current search operation may be a search operation triggered by a user inputting a current search term in a search interface. In a specific example, the current search operation may be a search operation that the user enters a current search term in a search session area loaded in the search interface and triggers. Wherein the search term may be a word, phrase, sentence, etc. After triggering the current search operation, the terminal device may generate a search request carrying the current search word in response to the current search operation, and initiate the search request to the server.

After the server receives the search request, the initial search result corresponding to the current search word carried in the search request can be obtained by utilizing the search engine, and the initial search result is sent to the terminal equipment so as to be convenient for the terminal equipment to display. The search engine may be a conventional search engine such as a full text search engine, a catalog search engine, a meta search engine, etc., and is used for retrieving a search result corresponding to a current search word from the internet as an initial search result. Wherein the initial search result may include a plurality of search sub-results.

After the terminal device displays the initial search result, the user may browse any of the plurality of search sub-results included in the initial search result, and the terminal device may generate user behavior information for characterizing the user's intent to determine content of interest to the user in response to the browse operation for the initial search result. Thereafter, the terminal device may generate a result generation request carrying user behavior information, and initiate the result generation request to the server.

After the server receives the result generation request, based on the user behavior information carried in the result generation request, a generated search result corresponding to the current search word is obtained by using the large language model, and the generated search result is sent to the terminal equipment so as to be displayed by the terminal equipment. The large language model is pre-trained, has general language knowledge, world knowledge, professional knowledge in various fields and the like, is stored in the model in a parameter form, and has a strong language interaction function. In a specific example, the large language model may be an autoregressive generation model of a transducer architecture. Wherein, the generated search result may be generated as multi-modal information such as plain text, table, picture, video, etc., which is not particularly limited by the embodiments of the present disclosure.

By adopting the searching method provided by the embodiment of the disclosure, the searching request carrying the current searching word can be initiated to the server in response to the current searching operation; the display server utilizes the initial search result corresponding to the current search word obtained by the search engine; generating user behavior information in response to a browsing operation for an initial search result so as to initiate a result generation request carrying the user behavior information to a server; the display server utilizes the generated search results corresponding to the current search word obtained by the large language model based on the user behavior information. The method comprises the steps that a generated search result is used as a final search result aiming at a current search word, the generated search result is not obtained based on a traditional search engine, after a server obtains an initial search result corresponding to the current search word based on the traditional search engine, user behavior information is generated at a terminal device in response to a browsing operation aiming at the initial search result, and the server obtains the generated search result corresponding to the current search word by using a large language model based on the user behavior information. The user behavior information is used for representing the user intention to determine the interesting content of the user, so that the generated type search result corresponding to the current search word is obtained by utilizing the large language model based on the user behavior information, the personalized matching degree of the generated type search result and the user can be improved, and the accuracy of the generated type search result is improved.

In addition, as previously described, in the embodiments of the present disclosure, the initial search result corresponding to the current search term may include a plurality of search sub-results, which may view more search sub-results corresponding to the current search term by infinitely dropping, scrolling, or loading, the page-turning manner for convenience of user screening.

In some alternative embodiments, "generating user behavior information in response to a browsing operation for initial search results" may include the steps of:

determining a current landing page in response to a browsing operation for the initial search result;

in the case that the search session also comprises a history search word except the current search word, acquiring a history search result corresponding to the history search word and a history landing page corresponding to the history search result;

user behavior information is generated based on the current landing page and the historical landing page.

The current landing page can be used for representing a page corresponding to any search sub-result in a plurality of search sub-results included in the initial search result when a user browses the search sub-result. For example, if the user browses the search sub-result A1 in the initial search result, the current landing page includes a page corresponding to the search sub-result A1; for another example, if the user browses the search sub-results A1 and a search sub-result A2 in the initial search results, the current landing page includes a page corresponding to the search sub-result A1 and a page corresponding to the search sub-result A2.

The current search word may be one search word in the search session, that is, the search session may include one or more search words, and in the case that the search session includes a plurality of search words, the plurality of search words may be issued in series, where the issued sequence number is the round number of the search word. For example, when the first search word is issued, the step S101 to step S104 are executed with the first search word as the current search word; when a second search word is sent out, the second search word is used as a current search word, and steps S101-S104 are executed, wherein the first search word is a historical search word except the current search word (the second search word) in the search session, and an initial search result and a generated search result corresponding to the first search word are both historical search results corresponding to the historical search word; if the third search word is sent out continuously, the third search word is used as the current search word, and steps S101 to S104 are executed, at this time, the first search word and the second search word are historical search words except the current search word (third search word) in the search session, and the initial search result and the generated search result corresponding to the first search word are both historical search results corresponding to the historical search word.

Also, it is understood that in embodiments of the present disclosure, a history landing page may be used to characterize a page corresponding to each of a plurality of search sub-results included in a history search result as the user browses the page. For example, when the user browses the search sub-result B1 in the history search results, the history landing page includes a page corresponding to the search sub-result B1; for another example, if the user browses the search sub-results B1 and B2 in the history search results, the history landing page includes a page corresponding to the search sub-result B1 and a page corresponding to the search sub-result B2.

Through the steps, in the embodiment of the disclosure, after the current landing page is determined and the historical landing page corresponding to the historical search result is obtained, the user behavior information is generated based on the current landing page and the historical landing page, so that the characterizability of the user behavior information is improved, the personalized matching degree of the generated search result and the user is further improved, and the accuracy of the generated search result is improved.

In some alternative embodiments, "generating user behavior information based on the current landing page and the historical landing page" may include the steps of:

Acquiring page related information of a target landing page;

user behavior information is generated based on the page related information.

Wherein the target landing page is any one of a current landing page and a historical landing page, that is, acquiring page related information of the target landing page includes: and respectively taking the current landing page and the historical landing page as target landing pages, and acquiring page related information of the target landing pages.

Wherein the page-related information includes at least one of a content title, content information, and page source of the target landing page.

In a specific example, the content information may be page content of a target landing page; in another specific example, the content information may be brief information generated after summarizing page content of the target landing page.

In a particular example, the page source may be a web page address.

Through the steps, in the embodiment of the disclosure, the page related information of the target landing page can be obtained, and the user behavior information is generated based on the page related information. Because the target landing page is any one of the current landing page and the historical landing page, and the page related information comprises at least one of the content title, the content information and the page source of the target landing page, the comprehensiveness of the page related information can be improved, so that the personalized matching degree of the generated search result and the user can be further improved, and the accuracy of the generated search result can be improved.

In some alternative embodiments, the search method may further include the steps of:

the display server utilizes one or more predictive search terms obtained by the large language model based on the user behavior information.

That is, after generating the user behavior information, the server may further obtain one or more predicted search words using the large language model based on the user behavior information and transmit the one or more predicted search words to the terminal device for display by the terminal device. Thereafter, the user may directly click on any predicted search word of the one or more predicted search words, and perform steps S101 to S104 using the predicted search word as a current search word. The predicted search term may be a search term related to the user behavior information, and is a predicted search term of the next round that the user may send out.

Through the steps, in the embodiment of the disclosure, one or more predicted search words obtained by the server based on the user behavior information by using the large language model can be directly displayed on the terminal device. The user can directly click on any predicted search word in the one or more predicted search words, and step S101 to step S104 are executed by taking the predicted search word as the current search word, so that the operation flow of the user is simplified, and the availability and the popularization of the search method are improved.

Referring to fig. 2, in a specific example, after a user inputs a current search word in a search page on a terminal device, first, an initial search result corresponding to the current search word may be displayed in a full-network search result display area in the search page, and then, a generated search result corresponding to the current search word is displayed in a personalized result generation area in the search page, and at the same time, one or more predicted search words corresponding to the current search word are displayed in a search recommendation area in the search page. Wherein the generated search results corresponding to the current search word may be displayed in front of the initial search results corresponding to the current search word and the one or more predicted search words corresponding to the current search word, i.e., the generated search results corresponding to the current search word will be displayed as initial display results for prompt browsing by the user.

In some alternative embodiments, the "initial search results corresponding to the current search term obtained by the display server using the search engine" may include the steps of:

creating a root sandbox corresponding to the current search term;

creating a child sandbox in the root sandbox for each search child result included in the initial search result;

For each search sub-result included in the initial search result, display logic of the search sub-result is executed in a sub-sandbox corresponding to the search sub-result to display the search sub-result.

Wherein, for the current search word, a root sandbox can be created, and the root sandbox is bound with the round number of the current search word, so as to complete the step of creating the root sandbox corresponding to the current search word.

Thereafter, a child sandbox may be created in the root sandbox for each search child result included in the initial search result. In the process, a corresponding application program interface (Application Programming Interface, API) may be created for each child sandbox to facilitate the communication needs of the child sandbox with the external environment. Wherein, each child sandbox creates a corresponding API may include Window API, document API, element API, and the like. Finally, for each search sub-result included in the initial search result, display logic of the search sub-result may be executed in a sub-sandbox corresponding to the search sub-result to display the search sub-result.

Referring to fig. 3, in a specific example, the initial search results corresponding to the current search term include a search sub-result C1, a search sub-result C2, and a search sub-result C3, and then after creating the root sandbox D1 corresponding to the current search term, create sub-sandboxes, namely, a sub-sandbox D11, a sub-sandbox D12, and a sub-sandbox D13, in the root sandbox D1 for the search sub-result C1, the search sub-result C2, and the search sub-result C3, respectively, so as to execute display logic of the search sub-result C1 in the sub-sandbox D11 to display the search sub-result C1; display logic for searching sub-result C2 is executed in AND sub-sandbox D12 to display search sub-result C2; display logic of the search sub-result C3 is executed in the and sub-sandbox D13 to display the search sub-result C3.

Through the steps, in the embodiment of the disclosure, the root sandbox corresponding to the current search word can be created so as to protect and isolate the display logic of the initial search result corresponding to the current search word, and meanwhile, in the root sandbox corresponding to the current search word, a sub-sandbox is created for each search sub-result included in the initial search result, and the search sub-results included in the initial search result can be prevented from being in conflict with the display logic (for example, conflicts of variables, functions, document object models (Document Object Model, DOM) and the like occur), so that the initial search result is ensured to be displayed normally.

and responding to the next round of search operation, switching the display state of the initial search result from an unfolding state to a folding state, or switching the display state of the search sub-result which is not browsed in the initial search result from the unfolding state to the folding state.

It may be appreciated that, in the embodiment of the present disclosure, the initial search result corresponding to the current search term may be displayed in a manner of folding the supplementary result, that is, the display state of the initial search result may be switched from the unfolded state to the folded state in response to the next round of search operation.

Referring to fig. 4A, in a specific example, after a user inputs a current search word "i want to travel to E region" on a search page on a terminal device, first, initial search results corresponding to the current search word "i want to travel to E region" including a search sub-result F1, a search sub-result F2, and a search sub-result F3 may be displayed in a full-network search result display area, and thereafter, a generated search result corresponding to the current search word "i want to travel to E region" is displayed in a personalized result generation area, and at the same time, one or more predicted search words corresponding to the current search word "i want to travel to E region" are displayed in a search recommendation area. Taking the way of folding the supplementary results as an example, after the user inputs the next round of search words "how local weather is" on the search page on the terminal device, the display state of the initial search results can be switched from the unfolded state to the folded state.

If the search method further includes a step of "displaying one or more predicted search words obtained by the server using the large language model based on the user behavior information", when the initial search result corresponding to the current search word is displayed in a manner of folding the supplementary result, the display state of the one or more predicted search words corresponding to the current search word may also be switched from the unfolded state to the folded state.

It may be appreciated that, in the embodiment of the present disclosure, an inserting and folding manner may also be adopted to display an initial search result corresponding to a current search word, that is, a display state of a search sub-result that is not browsed in the initial search result may be switched from an unfolded state to a folded state in response to a next round of search operation.

Referring to fig. 4B, in another specific example, after the user inputs the current search word "i want to travel to E region" on the search page on the terminal device, first, initial search results corresponding to the current search word "i want to travel to E region" including a search sub-result F1, a search sub-result F2, and a search sub-result F13 may be displayed in the whole network search result display area, and thereafter, a generated search result corresponding to the current search word "i want to travel to E region" is displayed in the personalized result generation area, and at the same time, one or more predicted search words corresponding to the current search word "i want to travel to E region" are displayed in the search recommendation area. Taking the manner of inserting and folding as an example, after the user inputs the next round of search words "how local weather is" on the search page on the terminal device, the display state of the search sub-result (for example, the search sub-result F13) which is not browsed in the initial search result may be switched from the unfolded state to the folded state.

If the search method further includes a step of "displaying one or more predicted search words obtained by the server using the large language model based on the user behavior information", the display state of the one or more predicted search words corresponding to the current search word may be switched from the expanded state to the collapsed state when the initial search result corresponding to the current search word is displayed in a mid-insert collapse manner.

Through the above steps, in the embodiments of the present disclosure, in response to the next round of search operation, the display state of the initial search result may be switched from the unfolded state to the folded state, or the display state of the search sub-result that is not browsed in the initial search result may be switched from the unfolded state to the folded state. Therefore, under the condition that the core search results (the generated search results corresponding to the current search word) are displayed, the browsing space can be saved, and the search interface is simpler.

The embodiment of the disclosure provides a search method which can be applied to electronic equipment, such as a server. A search method provided by the embodiment of the present disclosure will be described below with reference to a flowchart shown in fig. 5. It should be noted that although a logical order is illustrated in the flowchart, in some cases, the steps illustrated or described may be performed in other orders.

Step S501, receiving a search request carrying a current search word initiated by a terminal device in response to a current search operation;

step S502, obtaining initial search results aiming at the current search word by utilizing a search engine so as to send the initial search results to terminal equipment;

step S503, after receiving the user behavior information generated by the terminal equipment in response to the browsing operation for the initial search result, initiating a result generation request carrying the user behavior information;

step S504, based on the user behavior information, the generated search results for the current search word are obtained by using the large language model so as to send the generated search results to the terminal equipment.

After the server receives the search request, the initial search result corresponding to the current search word carried in the search request can be obtained by utilizing the search engine, and the initial search result is sent to the terminal equipment so as to be convenient for the terminal equipment to display. The search engine may be a conventional search engine such as a full text search engine, a catalog search engine, a meta search engine, etc., which is used to retrieve initial search results corresponding to the current search word from the internet. Wherein the initial search result may include a plurality of search sub-results.

After the server receives the result generation request, based on the user behavior information carried in the result generation request, a generated search result corresponding to the current search word is obtained by using the large language model, and the generated search result is sent to the terminal equipment so as to be displayed by the terminal equipment. The large language model is pre-trained, has general language knowledge, world knowledge, professional knowledge in various fields and the like, is stored in the model in a parameter form, and has a strong language interaction function. In a specific example, the large language model may be an autoregressive generation model of a transducer architecture.

By adopting the searching method provided by the embodiment of the disclosure, the searching request carrying the current searching word initiated by the terminal equipment in response to the current searching operation can be received; obtaining initial search results aiming at the current search word by utilizing a search engine so as to conveniently send the initial search results to terminal equipment; after receiving user behavior information generated by the terminal equipment in response to browsing operation aiming at an initial search result, initiating a result generation request carrying the user behavior information; based on the user behavior information, the generated search results for the current search word are obtained by using a large language model so as to be transmitted to the terminal equipment. The method comprises the steps that a generated search result is used as a final search result aiming at a current search word, the generated search result is not obtained based on a traditional search engine, after a server obtains an initial search result corresponding to the current search word based on the traditional search engine, user behavior information is generated at a terminal device in response to a browsing operation aiming at the initial search result, and the server obtains the generated search result corresponding to the current search word by using a large language model based on the user behavior information. The user behavior information is used for representing the user intention to determine the interesting content of the user, so that the generated type search result corresponding to the current search word is obtained by utilizing the large language model based on the user behavior information, the personalized matching degree of the generated type search result and the user can be improved, and the accuracy of the generated type search result is improved.

In some alternative embodiments, "obtaining initial search results for the current search term using a search engine to facilitate sending the initial search results to the terminal device" may include the steps of:

in the case that the search session also includes a history search word other than the current search word, obtaining a target search word based on the history search word and the current search word;

obtaining search results for the target search term using a search engine; wherein the search results for the target search term are initial search results;

and sending the initial search result to the terminal equipment.

The current search word may be one search word in the search session, that is, the search session may include one or more search words, and in the case that the search session includes a plurality of search words, the plurality of search words may be issued in series, where the issued sequence number is the round number of the search word. For example, when the first search word is issued, the step S101 to step S104 are executed with the first search word as the current search word; thereafter, when the second search word is issued, the second search word is used as the current search word, and steps S101 to S104 are executed, where the first search word is a historical search word except the current search word (second search word) in the search session; if the third search word is sent out continuously, the third search word is used as the current search word, and steps S101-S104 are executed, and at this time, the first search word and the second search word are historical search words except the current search word (third search word) in the search session.

The historical search terms may be stored in a server, so that the historical search terms are obtained from the server directly when the steps are executed, or may be carried in a search request and sent to the server, which is not particularly limited in the embodiment of the disclosure.

In a specific example, "obtaining a target search term based on a historical search term and a current search term" may include: and based on the historical search words, rewriting the current search words to obtain target search words. For example, based on the historical search terms, the current search term is rewritten using a Query rewrite model to obtain the target search term. Illustratively, the current search word is "how local weather is," and the history search word is "how me wants to travel to E region," then based on the history search word "how me wants to travel to E region," the current search word "how local weather is" is rewritten, and the obtained target search word may be "how weather is in E region.

Through the steps, in the embodiment of the disclosure, when the search session further includes a history search word other than the current search word, the target search word is obtained based on the history search word and the current search word, and the search engine is utilized to obtain the search result for the target search word, and the search result is used as the initial search result corresponding to the current search word and is sent to the terminal device. Because the target search word is obtained based on the historical search word and the current search word, i.e. is obtained in combination with the context, the method has definite semantics, and even if the user does not have the search word editing skill, the reliability of the initial search result can be improved, so that the accuracy of the generated search result is improved.

In some alternative embodiments, "based on user behavior information, obtaining the generated search results for the current search term using a large language model to facilitate sending the generated search results to the terminal device" may include the steps of:

generating prompt information based on the user behavior information;

inputting the prompt information into a large language model to obtain a generated search result which is output by the large language model and aims at the current search word;

and sending the generated search result to the terminal equipment.

For further description of the user behavior information, reference may be made to the description related to the search method embodiment applied to the terminal device, which is not described herein.

Based on the above steps, it may be understood that in the embodiment of the present disclosure, after the user behavior information is obtained, prompt information (also referred to as Prompt information of a large language model) may be generated based on the user behavior information, and the Prompt information is input into the large language model, so as to obtain a generated search result for the current search word output by the large language model.

Through the steps, in the embodiment of the disclosure, the prompt information can be generated based on the user behavior information, the large language model is guided to output the generated search result aiming at the current search word according to the prompt information, and the personalized matching degree of the generated search result and the user can be improved, so that the accuracy of the generated search result is improved, and the search efficiency can be further improved based on a simplified search flow.

In some alternative embodiments, "generating hint information based on user behavior information" may include the steps of:

in the case that the search session also comprises historical search words except the current search word, acquiring historical search results corresponding to the historical search words;

the hint information is generated based on the search session, the historical search results, the initial search results, and the user behavior information.

As previously described, the current search term may be one of the search terms in the search session, i.e., the search session may include one or more search terms, and in the case where the search session includes multiple search terms, the multiple search terms may be issued in series, with the issued sequence number being the turn number of the search term. For example, when the first search word is issued, the step S101 to step S104 are executed with the first search word as the current search word; when a second search word is sent out, the second search word is used as a current search word, and steps S101-S104 are executed, wherein the first search word is a historical search word except the current search word (the second search word) in the search session, and an initial search result and a generated search result corresponding to the first search word are both historical search results corresponding to the historical search word; if the third search word is sent out continuously, the third search word is used as the current search word, and steps S101 to S104 are executed, at this time, the first search word and the second search word are historical search words except the current search word (third search word) in the search session, and the initial search result and the generated search result corresponding to the first search word are both historical search results corresponding to the historical search word.

In the embodiment of the disclosure, the historical search results and the initial search results for generating the prompt information may specifically include at least one of a content title, content information and an internal source of any search sub-result in the historical search results, and at least one of a content title, content information and an internal source of any search sub-result in the initial search results.

In addition, in the embodiment of the present disclosure, the historical search terms, the historical search results and the initial search results in the search session other than the current search term may be stored in the server, so that the historical search terms, the historical search results and the initial search results may be obtained from the server directly when the above steps are executed, or may be carried in a result generation request and sent to the server, which is not particularly limited in the embodiment of the present disclosure.

Through the steps, in the embodiment of the disclosure, when the search session further includes a history search word other than the current search word, a history search result corresponding to the history search word may be obtained, and prompt information may be generated based on the search session, the history search result, the initial search result, and the user behavior information. Because the prompt information is generated based on the search session, the historical search result, the initial search result and the user behavior information, the prompt information has certain comprehensiveness, and therefore the accuracy of the generated search result can be further improved.

based on the user behavior information, one or more predicted search terms obtained using the large language model;

one or more predictive search terms are transmitted to the terminal device.

Through the steps, in the embodiment of the disclosure, one or more predicted search words obtained by the server based on the user behavior information by using the large language model can be directly displayed on the terminal device. The user may click on any predicted search word of the one or more predicted search words directly, and execute steps S101 to S104 with the predicted search word as a current search word, so as to simplify the user operation, thereby improving the usability of the search method.

It should be noted that, in the embodiment of the present disclosure, when a search session further includes a history search word other than a current search word, one or more predicted search words may be generated in a process of generating a generated search result for the current search word output by the large language model by inputting the prompt information into the large language model after generating the prompt information based on the search session, the history search result corresponding to the history search word, the initial search result corresponding to the current search word, and the user behavior information.

An integrity procedure of a search method provided in an embodiment of the present disclosure will be described below with reference to fig. 6.

(1) After the user enters the search application installed on the terminal device, a search page is displayed on the terminal device.

(2) The user inputs the current search word in the search page, triggers the current search operation and initiates a search request carrying the current search word to the server; and in the case that the search session further comprises a historical search word except the current search word, the search request further carries the historical search word.

(3) The server rewrites the current search word by using a Query rewrite model based on the historical search word to obtain a target search word, obtains a search result aiming at the target search word by using a search engine, and sends the initial search result to the terminal equipment as an initial search result corresponding to the current search word.

(4) Displaying initial search results corresponding to the current search word in a full-network search result display area in a search page; wherein, "displaying initial search results corresponding to the current search term" includes: creating a root sandbox corresponding to the current search term; creating a child sandbox in the root sandbox for each search child result included in the initial search result; for each search sub-result included in the initial search result, display logic of the search sub-result is executed in a sub-sandbox corresponding to the search sub-result to display the search sub-result.

(5) The terminal equipment responds to the browsing operation aiming at the initial search result to generate user behavior information so as to initiate a result generation request carrying the user behavior information to the server; in the case that the search session further includes a history search word other than the current search word, the result generation request further carries the history search word other than the current search word, a history search result corresponding to the history search word, and an initial search result corresponding to the current search word.

(6) Generating prompt information based on a search session, a historical search result corresponding to a historical search word, an initial search result corresponding to a current search word and user behavior information, inputting the prompt information into a large language model, obtaining a generated search result aiming at the current search word output by the large language model, and sending the generated search result to terminal equipment by utilizing a Server-Send Event (SSE) technology; in the process, the large language model can be endowed with wind control intervention capability so as to effectively avoid unreasonable content generation, and meanwhile, the large language model can also output one or more predicted search words corresponding to the current search word and send the predicted search word to the terminal equipment.

(7) The terminal device displays the generated search result for the current search word and one or more predicted search words corresponding to the current search word, and records a landing page corresponding to the generated search result.

(8) And responding to the next round of search operation, switching the display state of the initial search result from an unfolding state to a folding state, or switching the display state of the search sub-result which is not browsed in the initial search result from the unfolding state to the folding state, and executing the next round of search operation.

Fig. 7 is a schematic view of a search method according to an embodiment of the disclosure.

As described above, the search method provided by the embodiment of the present disclosure may be applied to an electronic device. Electronic devices are intended to represent various forms of digital computers, such as servers, or end devices such as desktop computers, laptop computers, mobile terminals, and the like.

The terminal device may be configured to:

The server may be configured to:

The current search operation may be a search operation triggered by a user inputting a current search term in a search interface.

It should be noted that, in the embodiment of the present disclosure, the schematic view of the scenario shown in fig. 7 is merely illustrative and not restrictive, and those skilled in the art may make various obvious changes and/or substitutions based on the example of fig. 7, and the obtained technical solution still falls within the scope of the embodiment of the present disclosure.

In order to better implement the above searching method applicable to the terminal device, the embodiments of the present disclosure further provide a searching apparatus, which may be integrated in an electronic device, for example, integrated in the terminal device. Hereinafter, a search apparatus 800 provided in the disclosed embodiment will be described with reference to a schematic structural diagram shown in fig. 8.

The search device 800 includes:

a first request initiating unit 801, configured to initiate a search request carrying a current search term to a server in response to a current search operation;

a first display unit 802, configured to display an initial search result corresponding to a current search term obtained by the server using the search engine;

a second request initiating unit 803 for generating user behavior information in response to a browsing operation for an initial search result so as to initiate a result generation request carrying the user behavior information to the server;

and a second display unit 804, configured to display a generated search result corresponding to the current search word, which is obtained by the server using the large language model, based on the user behavior information.

In some alternative embodiments, the second request initiating unit 803 is configured to:

acquiring page related information of a target landing page; the target landing page is any one of a current landing page and a historical landing page, and the page related information comprises at least one of a content title, content information and page source of the target landing page;

user behavior information is generated based on the page related information.

In some alternative embodiments, the search apparatus 800 further includes a third display unit for:

In some alternative embodiments, the first display unit 802 is configured to:

creating a root sandbox corresponding to the current search term;

In some alternative embodiments, the search apparatus 800 further includes a display state managing unit for:

The specific functions and examples of each unit of the search apparatus 800 in the embodiment of the present disclosure may be described with reference to the foregoing related descriptions of corresponding steps in the embodiment of the search method that may be applied to the terminal device, which are not described herein.

In order to better implement the above search method applicable to the server, the embodiments of the present disclosure also provide a search apparatus, which may be integrated in an electronic device, for example, integrated in the server. Hereinafter, a search apparatus 900 according to the disclosed embodiment will be described with reference to a schematic diagram shown in fig. 9.

Search apparatus 900, comprising:

a first request receiving unit 901, configured to receive a search request carrying a current search word initiated by a terminal device in response to a current search operation;

a search unit 902, configured to obtain an initial search result for a current search term by using a search engine, so as to send the initial search result to a terminal device;

A first request receiving unit 903, configured to receive a result generation request carrying user behavior information initiated by a terminal device after generating user behavior information in response to a browsing operation for an initial search result;

the result generating unit 904 is configured to obtain, based on the user behavior information, a generated search result for the current search term using the large language model, so as to send the generated search result to the terminal device.

In some alternative embodiments, the search unit 902 is configured to:

obtaining search results aiming at target search words by using a search engine as initial search results;

and sending the initial search result to the terminal equipment.

In some alternative embodiments, the result generation unit 904 is configured to:

generating prompt information based on the user behavior information;

and sending the generated search result to the terminal equipment.

In some alternative embodiments, the search unit is further configured to:

one or more predictive search terms are transmitted to the terminal device.

The specific functions and examples of the units of the apparatus 900 in the embodiment of the present disclosure may be referred to the foregoing related descriptions of the corresponding steps in the search method embodiment that may be applied to the server, which are not described herein.

In the technical scheme of the disclosure, the acquisition, storage, application and the like of the related user personal information all conform to the regulations of related laws and regulations, and the public sequence is not violated.

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

Fig. 10 shows a schematic block diagram of an example electronic device 1000 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile apparatuses, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing apparatuses. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 10, the apparatus 1000 includes a computing unit 1001 that can perform various appropriate actions and processes according to a computer program stored in a Read-Only Memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access Memory (Random Access Memory, RAM) 1003. In the RAM 1003, various programs and data required for the operation of the device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other by a bus 1004. An Input/Output (I/O) interface 1005 is also connected to bus 1004.

Various components in device 1000 are connected to I/O interface 1005, including: an input unit 1006 such as a keyboard, a mouse, and the like; an output unit 1007, for example, various types of displays, speakers, and the like; a storage unit 1008, such as a magnetic disk, an optical disk, or the like; and a communication unit 1009, such as a network card, modem, wireless communication transceiver, or the like. Communication unit 1009 allows device 1000 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunications networks.

The computing unit 1001 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 1001 include, but are not limited to, a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), various dedicated artificial intelligence (Artificial Intelligence, AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (Digital Signal Process, DSP), and any suitable processors, controllers, microcontrollers, etc. The computing unit 1001 performs the respective methods and processes described above, for example, a search method. For example, in some embodiments, the search method may be implemented as a computer software program tangibly embodied on a machine-readable medium, e.g., the storage unit 1008. In some embodiments, part or all of the computer program may be loaded and/or installed onto device 1000 via ROM 1002 and/or communication unit 1009. When the computer program is loaded into RAM 1003 and executed by computing unit 1001, one or more steps of the search method described above may be performed. Alternatively, in other embodiments, the computing unit 1001 may be configured to perform the search method in any other suitable way (e.g., by means of firmware).

Various implementations of the systems and techniques described here above can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (Field Programmable Gate Array, FPGAs), application specific integrated circuits (Application Specific Integrated Circuit, ASICs), application specific standard products (Application Specific Standard Product, ASSPs), systems On Chip (SOC), load programmable logic devices (Complex Programmable Logic Device, CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a RAM, a ROM, an erasable programmable read-Only Memory (EPROM) or flash Memory, an optical fiber, a portable compact disc read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a Cathode Ray Tube (CRT) display or a liquid crystal display (Liquid Crystal Display, LCD)) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area network (Local Area Network, LAN), wide area network (Wide Area Network, WAN) and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

The disclosed embodiments also provide a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform a search method.

The disclosed embodiments also provide a computer program product comprising a computer program which, when executed by a processor, implements a search method.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein. Moreover, in this disclosure, relational terms such as "first," "second," "third," and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Furthermore, "plurality" in the present disclosure may be understood as at least two.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions, improvements, etc. that are within the principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. A search method, comprising:

displaying initial search results corresponding to the current search word, which are obtained by the server through a search engine;

generating user behavior information in response to a browsing operation for the initial search result so as to initiate a result generation request carrying the user behavior information to the server;

and displaying the generated search results corresponding to the current search word, which are obtained by the server based on the user behavior information by using a large language model.

2. The method of claim 1, wherein the generating user behavior information in response to a browsing operation for the initial search result comprises:

in the case that the search session further comprises a history search word except the current search word, acquiring a history search result corresponding to the history search word and a history landing page corresponding to the history search result;

and generating the user behavior information based on the current landing page and the historical landing page.

3. The method of claim 2, wherein generating user behavior information based on the current landing page and the historical landing page comprises:

acquiring page related information of a target landing page; wherein the target landing page is any one of the current landing page and the history landing page, and the page related information includes at least one of a content title, content information and a page source of the target landing page;

and generating the user behavior information based on the page related information.

4. A method according to any one of claims 1 to 3, further comprising:

displaying one or more predicted search words obtained by the server by using a large language model based on the user behavior information.

5. A method according to any one of claims 1 to 3, wherein said displaying initial search results corresponding to the current search term obtained by the server using a search engine comprises:

creating a root sandbox corresponding to the current search term;

and for each search sub-result included in the initial search result, executing display logic of the search sub-result in a sub-sandbox corresponding to the search sub-result to display the search sub-result.

6. A method according to any one of claims 1 to 3, further comprising:

7. A search method, comprising:

obtaining initial search results aiming at the current search word by utilizing a search engine so as to send the initial search results to the terminal equipment;

receiving a result generation request carrying the user behavior information initiated after the terminal equipment generates the user behavior information in response to the browsing operation aiming at the initial search result;

based on the user behavior information, a large language model is utilized to obtain the generated search result aiming at the current search word so as to send the generated search result to the terminal equipment.

8. The method of claim 7, wherein the obtaining, with a search engine, initial search results for the current search term to facilitate sending the initial search results to the terminal device comprises:

Obtaining a target search word based on the historical search word and the current search word under the condition that the search session also comprises the historical search word except the current search word;

obtaining search results for the target search word by using a search engine as the initial search results;

and sending the initial search result to the terminal equipment.

9. The method of claim 7, wherein the obtaining, based on the user behavior information, the generated search results for the current search term using a large language model to facilitate sending the generated search results to the terminal device comprises:

generating prompt information based on the user behavior information;

inputting the prompt information into the large language model to obtain a generated search result which is output by the large language model and aims at the current search word;

and sending the generated search result to the terminal equipment.

10. The method of claim 9, wherein the generating hint information based on the user behavior information comprises:

acquiring a history search result corresponding to the history search word under the condition that the search session further comprises the history search word except the current search word;

And generating the prompt information based on the search session, the historical search results, the initial search results and the user behavior information.

11. The method of claim 7, further comprising:

based on the user behavior information, one or more predictive search terms obtained using a large language model;

and sending the one or more predicted search words to the terminal equipment.

12. A search apparatus comprising:

the first display unit is used for displaying initial search results corresponding to the current search word, which are obtained by the server through a search engine;

a second request initiating unit, configured to generate user behavior information in response to a browsing operation for the initial search result, so as to initiate a result generating request carrying the user behavior information to the server;

and the second display unit is used for displaying the generated search results corresponding to the current search word, which are obtained by the server based on the user behavior information by using a large language model.

13. The apparatus of claim 12, wherein the second request initiating unit is configured to:

14. The apparatus of claim 13, wherein the second request initiating unit is configured to:

15. The apparatus according to any one of claims 12 to 14, further comprising a third display unit for:

16. The apparatus according to any one of claims 12 to 14, wherein the first display unit is configured to:

Creating a root sandbox corresponding to the current search term;

17. The apparatus according to any one of claims 12 to 14, further comprising a display state managing unit configured to:

18. A search apparatus comprising:

the searching unit is used for obtaining initial searching results aiming at the current searching words by utilizing a searching engine so as to conveniently send the initial searching results to the terminal equipment;

the second request receiving unit is used for receiving a result generation request carrying the user behavior information initiated after the terminal equipment generates the user behavior information in response to the browsing operation aiming at the initial search result;

And the result generating unit is used for obtaining the generated search result aiming at the current search word by using a large language model based on the user behavior information so as to conveniently send the generated search result to the terminal equipment.

19. The apparatus of claim 18, wherein the search unit is configured to:

and sending the initial search result to the terminal equipment.

20. The apparatus of claim 18, wherein the result generation unit is configured to:

generating prompt information based on the user behavior information;

and sending the generated search result to the terminal equipment.

21. The apparatus of claim 20, wherein the result generation unit is configured to:

22. The apparatus of claim 18, further comprising a search term prediction unit to:

and sending the one or more predicted search words to the terminal equipment.

23. An electronic device, comprising:

at least one processor;

a memory communicatively coupled to the at least one processor;

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 11.

24. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-11.

25. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 11.