Information recommendation method and device, electronic equipment, storage medium and product
Through a variety of recall methods and caching technologies, information is recalled and stored in batches, which solves the problems of poor system performance and paging requirements, and achieves system performance improvement and user-friendly information recommendation experience.
Patent Information
- Application Number
- CN202410383181.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-29
- Publication Date
- 2025-09-30
AI Technical Summary
The information recommendation method in the existing technology has poor system performance due to the one-time recall of a large amount of information, and multi-batch recall cannot meet the user's paging requirements.
Adopt multiple recall methods, recall candidate information in multiple recall batches, and use cache technology to store the recalled and integrated information to ensure improved system performance and support paging query.
By recalling information in batches, the amount of information at a time is reduced, system pressure is reduced, system performance is improved, and caching technology is used to support users' free page turning needs.
Smart Images
Figure CN120723967A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of Internet technology, and in particular to an information recommendation method, device, electronic device, storage medium and product. Background Art
[0002] With the development of Internet technology, the amount of information on the Internet has exploded. Faced with the massive amount of information on the Internet, it is crucial to recommend information that meets user needs to users.
[0003] In related technologies, based on the search request sent by the user account, a large amount of information, such as 10,000 pieces, is recalled at one time, and then the recalled information is stored in the cache. When a query request for a certain page is received from the user account, the information corresponding to the page is obtained from the cache, and then the information corresponding to the page is recommended to the user account.
[0004] The above information recommendation method can meet the arbitrary page turning needs of user accounts, but due to recalling a large amount of information at one time, the system performance is poor. Summary of the Invention
[0005] The embodiments of the present application provide an information recommendation method, device, electronic device, storage medium, and product, which employ multiple recall methods to recall candidate information in multiple recall batches, thereby avoiding recalling a large amount of information at one time and improving system performance. The technical solution is as follows:
[0006] In a first aspect, an information recommendation method is provided, the method comprising:
[0007] Receiving a query request sent by a user account, wherein the query request includes page information of a target page requested to be displayed;
[0008] Determining a recall batch for each recall method based on page information of the target page, information stored in the cache, and a recall quantity corresponding to each recall method;
[0009] Based on the recall batches of each recall method, perform information recall operations to obtain candidate information recalled using each recall method;
[0010] Determining target information corresponding to the target page based on candidate information recalled using multiple recall methods;
[0011] The target information is sent to the user account.
[0012] In a second aspect, an information recommendation device is provided, the device comprising:
[0013] A receiving module, configured to receive a query request sent by a user account, wherein the query request includes page information of a target page requested to be displayed;
[0014] A first determining module is configured to determine a recall batch for each recall method based on page information of the target page, information stored in the cache, and a recall quantity corresponding to each recall method;
[0015] A recall module is used to perform information recall operations based on the recall batches of each recall method, and obtain candidate information to be recalled using each recall method;
[0016] A second determining module is configured to determine target information corresponding to the target page based on candidate information recalled using multiple recall methods;
[0017] A sending module is used to send the target information to the user account.
[0018] In a third aspect, an electronic device is provided, comprising a processor and a memory; the memory stores at least one program code; the at least one program code is used to be called and executed by the processor to implement the information recommendation method as described in the first aspect.
[0019] In a fourth aspect, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, and when the at least one computer program is executed by a processor, the information recommendation method as described in the first aspect can be implemented.
[0020] In a fifth aspect, a computer program product is provided, comprising a computer program, which, when executed by a processor, can implement the information recommendation method as described in the first aspect.
[0021] The beneficial effects of the technical solution provided by the embodiments of the present application are:
[0022] The embodiment of the present application supports multiple recall methods. After receiving the query request sent by the user account, based on the recall quantity of each recall method, the information stored in the cache and the page information of the target page requested to be displayed, the recall batch corresponding to each recall method is determined. Then, based on the recall batch corresponding to each recall method, the corresponding recall method is used to perform the information recall operation, and the candidate information recalled by each recall method is obtained. Then, based on the candidate information recalled by multiple recall methods, the target information on the target page is determined, and the target information is recommended to the user account. Since a one-time recall method is no longer used, but multiple recall methods are used to recall in multiple recall batches, the amount of information recalled in a single time is reduced, the pressure on the system is reduced, and the system performance is improved. In addition, the information of each recall will be stored in the cache. By querying the cache, it is possible to avoid recalling duplicate information, further reducing the amount of candidate information, and helping to improve system performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0024] Figure 1 is a schematic diagram of an implementation environment involved in an information recommendation method provided in an embodiment of the present application;
[0025] Figure 2 This is a flow chart of an information recommendation method provided by an embodiment of the present application;
[0026] Figure 3 This is a flowchart of another information recommendation method provided by an embodiment of the present application;
[0027] Figure 4 This is a schematic diagram of the structure of an information recommendation device provided in an embodiment of the present application;
[0028] Figure 5 A structural block diagram of an electronic device 500 provided by an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION
[0029] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.
[0030] It should be understood that the terms "each," "plurality," and "any" used in the embodiments of this application include two or more, "each" refers to each of the corresponding plurality, and "any" refers to any one of the corresponding plurality. For example, if a plurality of words includes 10 words, "each" refers to each of the 10 words, and "any" refers to any one of the 10 words.
[0031] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.
[0032] First, before executing the embodiments of the present application, the terms involved in the present application are explained.
[0033] Soutui is the abbreviation of search recommendation, which means that the search engine recommends relevant search results to users based on the user's search history, search behavior, browsing history and other information. It can help users find the information they want faster and improve the efficiency of information search.
[0034] Caching refers to temporarily storing data in memory to improve the efficiency of data access. In search recommendation systems, cache can be used to store candidate sets to improve recall efficiency.
[0035] Multi-way recall refers to using multiple recall strategies to recall candidate sets separately, and then fusing the recall results of different recall strategies to obtain the final candidate set. Using multi-way recall can improve the accuracy and diversity of recall information.
[0036] Batching refers to dividing the candidate set into multiple batches and then recalling them separately. Batching can improve the efficiency of recall, especially when the candidate set is large.
[0037] Main recall refers to the main recall strategy used. Main recall usually adopts simple strategies, such as text recall strategy, to improve recall efficiency.
[0038] A side recall is a recall strategy used to supplement the main recall. Side recalls often employ complex strategies, such as vector recall, to improve recall accuracy and diversity.
[0039] Multi-way fusion refers to fusing multiple recall results to obtain the final candidate set. Multi-way fusion can improve the accuracy and diversity of recall results.
[0040] In the search and push scenarios, there are currently two main recall fusion solutions:
[0041] The first one: One-time recall and fusion solution
[0042] This solution uses a large-scale recall and fusion approach to retrieve all the information needed for a user's current search or recommendation. This solution typically recalls a large amount of information, for example, 10,000 items. The recalled information is typically stored in a cache, allowing the user to subsequently query it by flipping through pages. This solution can meet the user's query needs for arbitrary page flipping, but the large amount of information retrieved at once places high demands on system processing performance and consumes considerable time.
[0043] The second type: waterfall recall solution
[0044] This solution can use a batch recall and fusion method to recall the information required by the user for this search or recommendation in multiple batches, and use a waterfall browsing method to continuously generate subsequent information for the user's current search or recommendation, thereby reducing the amount of information in each recall batch. Generally speaking, the amount of information recalled in each recall batch can be greater than the amount of information displayed in a single waterfall flow, thereby avoiding blank areas during the user's browsing process and improving the user's browsing experience. Of course, it is also possible to recall a certain amount of information at a time, and then store the recalled information in a cache to meet the user's browsing needs for continuous waterfall flows. This solution solves the performance and time-consuming problems of the one-time recall and fusion solution, but because it uses a waterfall-style browsing method, the total amount of information cannot be accurately known, and the user cannot manually turn the page, so it cannot meet the user's usage needs.
[0045] In order to solve the problem in the above-mentioned search and push scenario that the amount of information required for recall and fusion of a single batch is large, the time consumption is long, and the system performance requirements are high, and at the same time solve the problem that multi-batch recalls cannot be paginated, an embodiment of the present application provides an information recommendation method, which adopts a multi-batch recall method, and recalls information in batches on demand, thereby reducing the amount of information recalled in a single time, improving system performance, and reducing the time consumption of the recall fusion process. In addition, caching technology is used to write the recalled and fused information into the cache to facilitate subsequent paging queries, reduce the number of query requests sent, improve system performance, and reduce the time consumed in the recall fusion process. In addition, by setting the recall and fusion strategies, it is ensured that the recalled candidate information can be paginated normally.
[0046] Please refer to Figure 1 , which shows the implementation environment involved in the information recommendation method provided in the embodiment of the present application, see Figure 1 The implementation environment includes: a terminal 101 and a server 102. The terminal 101 and the server 102 can communicate directly or indirectly through a network 103, and the network 103 can be a wired network or a wireless network.
[0047] The terminal 101 may be a smartphone, tablet computer, laptop computer, desktop computer, smart speaker, wearable device, etc. The terminal 101 has at least one search and push application installed, such as a shopping application. Based on the at least one search and push application, the terminal 101 may send a search request or query request to the server 102 through a logged-in user account to obtain required information.
[0048] Server 102 is the backend server for any search and push application in terminal 101. Server 102 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Server 102 can receive search requests or query requests sent by terminal 101. In response to the search request or query request, it can use multiple recall methods to recall candidate information in multiple recall batches, merge the candidate information recalled by the multiple recall methods, and recommend the merged information to terminal 101.
[0049] This application embodiment provides an information recommendation method, see Figure 2 , the server 102 executes the embodiment of the present application, and the method flow provided by the embodiment of the present application includes:
[0050] 201. Receive a query request sent by a user account.
[0051] The query request may include page information of the target page requested for display, which may include the page number. Generally, the page number requested by the user account is greater than or equal to 1. In the search and push scenario, to obtain the required information, the user account can enter search criteria in the search and push application, which may include search terms, etc., and then send a query request to the server based on the search criteria. In addition to the page information of the target page, the query request also includes the search criteria. In this case, the target page may be the first page.
[0052] Of course, the information and page numbers recalled based on the search conditions can be displayed on the Soupui application interface. When browsing the information on the Soupui application interface, the user account can perform page turning operations to generate a query request for the target page. At this time, the target page can be the page selected by the user account on the Soupui application interface, or it can be the page corresponding to the page number entered by the user account on the Soupui application interface.
[0053] 202. Determine a recall batch for each recall method based on the page information of the target page, the information stored in the cache, and the recall quantity corresponding to each recall method.
[0054] The embodiment of the present application supports multiple recall methods, including main recall methods and bypass recall methods. The main recall can be text recall (TextRecall), that is, the recall is performed directly based on the search terms input by the user account; the bypass recall can be vector recall (VectorRecall), that is, based on the search terms input by the user account, a corresponding feature vector is generated, and then the recall is performed based on the generated feature vector. In the embodiment of the present application, the number of main recall paths can be one, and the number of bypass recall paths can be one or multiple. For multi-path recall fusion, the total number of search results is based on the main recall, and the bypass recall is auxiliary. Therefore, it is necessary to ensure that the minimum number of information recalled by the main recall method is greater than or equal to the number of information carried by the target page. To facilitate analysis and calculation, the embodiment of the present application sets the same amount of information carried on each page. In addition, considering that the bypass recall method is based on the similarity of feature vectors, the accuracy of the recalled information is worse than that of the main recall method. Therefore, to ensure the accuracy of the recall information, the recall batch of the main recall method can be unlimited, while the recall batch of the bypass recall method can be limited, such as 1 time. In addition, the embodiment of the present application can also set corresponding recall quantities for different recall batches of the main recall method. Specifically, the recall quantities corresponding to different recall batches of the main recall method can be set to be the same, or the recall quantities corresponding to different recall batches of the main recall method can be set to be different. For example, the recall quantity of the first recall batch of the main recall method can be set to x, and the recall quantity corresponding to the second recall batch can be set to y, where x and y are positive integers, and x and y can be the same or different. Similarly, the embodiment of the present application can also set corresponding recall quantities for different recall batches of the bypass recall method. Specifically, the recall quantity corresponding to different recall batches of the bypass recall method can be set to be the same, or the recall quantity corresponding to different recall batches of the bypass recall method can be set to be different. For example, the recall quantity of the first recall batch in the bypass recall method can be set to z pieces.
[0055] Assume that the recall batch of the main recall method is i (where i is greater than or equal to 2), the recall quantity of the first recall batch of the main recall method is x, and the recall quantity of the second recall batch and subsequent recall batches is y. Then, when the main recall method is used for recall, the ranges of the number of information recalled in different recall batches are: the range of the number of information recalled in the first recall batch is [0, x); the range of the number of information recalled in the second recall batch is [x, x+y); the range of the number of information recalled in the i-th recall batch is [y*(i-2)+x, y*(i-1)+x).
[0056] Assume that the recall batch of the bypass recall method is 1, and the recall quantity of the first recall batch of the bypass recall method is z. Then, when the bypass recall method is used for recall, the number of information recalled in the first recall batch is in the range of [0, z).
[0057] In an embodiment of the present application, based on the page information of the target page, the information stored in the cache, and the recall quantity corresponding to each recall method, the recall batch for each recall method may be determined using the following method:
[0058] 2011. Based on the number of pages and the amount of information displayed on each page, determine the information carried by the target page.
[0059] In the embodiment of the application, the number of pages of the target page is set to m, and the amount of information displayed on each page is n. Then, the information quantity interval corresponding to the information carried by the target page is [n*(m-1), nm).
[0060] For example, if the target page has 10 pages and each page displays 10 pieces of information, the information quantity range corresponding to the information carried by the target page is [90, 100).
[0061] 2012. Based on the information carried by the target page, the information stored in the cache, and the recall quantity corresponding to each recall method, determine the recall batch for each recall method.
[0062] In one possible implementation, based on the information carried by the target page, the information stored in the cache, and the recall quantity corresponding to each recall method, the recall batch for each recall method may be determined using the following method:
[0063] 201211. Check whether the information carried by the target page is stored in the cache.
[0064] In an embodiment of the present application, information recalled based on search terms of a user account may be stored in the cache. To avoid repeated recall of information, it may be possible to query whether the information carried by the target page is stored in the cache.
[0065] 201212. When the information carried by the target page is stored in the cache, the amount of information displayed on each page is subtracted from the amount of information carried by the target page stored in the cache to obtain the amount of information to be recalled.
[0066] When the cache stores information carried by the target page, the amount of information carried by the target page stored in the cache can be obtained, and the amount of information displayed on each page can be obtained. Then, the amount of information displayed on each page is subtracted from the amount of information carried by the target page stored in the cache to obtain the amount of information to be recalled.
[0067] The information carried by the target page stored in the cache may include two situations:
[0068] The first case is that all the information carried by the target page is stored in the cache
[0069] In the first case, since all the information on the target page is in the cache, there's no need to recall it; the information can be retrieved directly from the cache. In this case, the amount of information to be recalled is 0. The specific analysis is as follows: Since the amount of information on the target page is equal to the amount of information displayed per page, and since all the information on the target page is in the cache, that is, the amount of information on the target page stored in the cache is equal to the amount of information displayed per page, subtracting the amount of information on the target page stored in the cache from the amount of information displayed per page yields a total of 0 information to be recalled.
[0070] The second case is that part of the information carried by the target page is stored in the cache
[0071] For the second case, the information stored in the cache can be directly obtained from the cache, and the information not stored in the cache needs to be recalled.
[0072] 201213. Based on the recall quantity corresponding to each recall method and the amount of information to be recalled, determine the recall batches for each recall method.
[0073] In the embodiment of the present application, the recall quantity corresponding to each recall method is fixed, while the recall batch of the bypass recall method is limited. Therefore, the recall batch corresponding to the bypass recall method can be pre-set, and then the recall batch of the main recall method can be determined based on the recall quantity corresponding to each recall method and the number of information to be recalled, as well as the recall batch corresponding to the bypass recall method. For example, the number of information to be recalled is 100, the recall batch of the bypass recall method is 1, the recall quantity corresponding to the bypass recall method is 10, the recall quantity corresponding to the first recall batch of the main recall method is 50, and the recall quantity corresponding to the second recall batch and subsequent recall batches of the main recall method is 30. The number of information to be recalled, 100, is subtracted from the 10 recalled by the bypass recall method. The remaining 90 information needs to be recalled using the main recall method. For 90 information, the main recall method needs to be recalled 3 times, so it can be determined that the recall batch of the main recall method is 3.
[0074] The following two points require clarification. First, considering that the embodiments of the present application employ a multi-channel recall method for recall, the information recalled by the multi-channel recall method may have a certain degree of overlap. If the information is recalled according to the number of information to be recalled, after the information recalled by the multi-channel recall method is merged, the amount of fused information plus the amount of information stored in the cache may be less than the amount of information required for the target page, resulting in blank areas on the target page and a poor user experience. To improve the user's browsing experience and ensure that the target page can be displayed normally, the total amount of information recalled using each recall method should be greater than the amount of information to be recalled.
[0075] Secondly, the more recall batches there are, the more information is recalled, and the more it can meet the paging requirements. However, as the number of recall batches increases, the amount of information stored in the cache increases, and the system performance becomes worse. Therefore, when determining the recall batch size for each recall method, it should not be too large.
[0076] In another possible implementation, based on the information quantity interval, the information stored in the cache, and the recall quantity corresponding to each recall method, the recall batch for each recall method may be determined using the following method:
[0077] 201221. Based on the amount of information displayed on each page and the recall quantity corresponding to each recall method, determine the reference recall batch for each recall method.
[0078] 201222. Check whether the information carried by the target page is stored in the cache.
[0079] 201221. When the information carried by the target page is stored in the cache, the number of information displayed on each page is subtracted from the number of information carried by the target page stored in the cache to obtain the number of information to be recalled;
[0080] 201222. Based on the amount of information to be recalled, the reference recall batch for each recall method is revised to obtain the recall batch for each recall method.
[0081] For this step, since the reference recall batch for each recall method is determined based on the amount of information displayed on the target page, and part of the target page's information is stored in the cache, if the recall is performed according to the reference recall batch for each recall method, some duplicate information will be recalled, which will not only waste network resources, but also extend the recall time and reduce system performance. Therefore, when the actual amount of information that needs to be recalled is determined, the reference recall batch for each recall method can be revised based on the actual amount of information that needs to be recalled, that is, the amount of information to be recalled. When revising the reference recall batch corresponding to each recall method, the recall batch corresponding to the bypass recall method can be reduced while ensuring that the main recall method and the bypass recall method are recalled in parallel.
[0082] In another embodiment of the present application, before determining the recall batch for each recall method based on the page information of the target page, the information stored in the cache, and the recall quantity corresponding to each recall method, the server can query whether the shared information context of the user account is stored in the cache. The shared information context is the information shared when the user account performs multiple query operations on the candidate information searched for the same search term. The shared information can be the total number of candidate information searched for the same search term based on the user account, etc. When the shared information context is stored in the cache, the shared information context is obtained, and then based on the shared information context and the number of information displayed on each page, the server can query whether the shared information context of the user account is stored in the cache. amount, determine the maximum number of pages of information in the system, and then determine whether the page number of the target page exceeds the maximum number of pages. If the number of pages included in the page information of the target page is less than the maximum number of pages, it can be determined that the page number of the target page does not exceed the page limit, and then perform operations based on the page information of the target page, the information stored in the cache and the recall quantity corresponding to each recall method, and determine the recall batch operation for each recall method; if the number of pages included in the page information of the target page is greater than the maximum number of pages, it can be determined that the page number of the target page exceeds the page limit, and then send a prompt message to the user account to prompt the user account to modify the page information of the target page to be displayed.
[0083] Furthermore, when the shared information context is not stored in the cache, a shared information context can be generated for the user account after the information recall operation is completed, and then the shared information context is stored in the cache so that the shared information context can be obtained from the cache the next time a query request for the user account is received.
[0084] 203. Based on the recall batch of each recall method, perform an information recall operation to obtain candidate information recalled using each recall method.
[0085] In an embodiment of the present application, after determining the recall batches for each recall method, the recall method corresponding to each recall method can be used to perform an information recall operation, thereby obtaining candidate information for each recall method in each recall batch. For example, if the recall batch for the main recall method is 2 and the recall batch for the bypass recall method is 1, then two main recalls and one bypass recall can be performed. During each main recall, candidate information can be recalled according to the recall quantity corresponding to each recall batch, thereby obtaining candidate information for two recall batches corresponding to the main recall method and candidate information for one recall batch for the bypass recall method.
[0086] 204. Determine target information corresponding to the target page based on the candidate information recalled using multiple recall methods.
[0087] Specifically, based on the candidate information recalled using multiple recall methods, the target information corresponding to the target page can be determined using the following methods:
[0088] 2041. The candidate information recalled by multiple recall methods in the same recall batch is integrated to obtain the integration results of multiple recall batches.
[0089] For candidate information recalled using different recall methods, the candidate information recalled within the same recall batch can be fused to obtain the fusion results for each recall batch. When fusion is performed on the candidate information recalled within the same recall batch, the information recalled within the recall batch using the main recall method and the information recalled within the recall batch using the bypass recall method can be deduplicated, the deduplicated information can be sorted, and the sorted information can be stored in a cache for easy access after the recall is completed.
[0090] Two points need to be explained here. First, the fusion operation in this step can be performed during the information recall process, that is, each time a batch of candidate information is recalled by the main recall method and the bypass recall method, the information recalled in the same recall batch is fused, and then the fusion result is stored in the cache so that it can be obtained from the cache after the fusion of this information recall is completed.
[0091] Second, the recall batches corresponding to the main recall method and the recall batches corresponding to the bypass recall method may be different. When merging the candidate information recalled by the main recall method with the candidate information recalled by the bypass recall method, only the candidate information recalled in the same recall batch is merged. For a recall batch with only one line, it can be directly stored in the cache.
[0092] 2042. Based on the amount of information displayed on each page, obtain a certain amount of information from the fusion results of multiple recall batches as target information corresponding to the target page.
[0093] The target information is the information displayed on the target page. Since the information stored in the cache is sorted by the search terms entered by the user account and the attribute information of the user account, the higher the ranking, the more it meets the user's needs. Therefore, when retrieving the target information corresponding to the target page from the cache, the information stored in the cache can be prioritized, and then the information after the current recall is obtained. When the amount of information obtained reaches the amount of information displayed on the target page, no further information is obtained.
[0094] 205. Send the target information to the user account.
[0095] After the target information is obtained, the target information may be sent to the user account so as to be displayed on the terminal that logs into the user account, thereby satisfying the user account's browsing requirements for the target page.
[0096] The embodiments of the present application, based on caching technology, employ a multi-batch recall method for information recall, resolving the issues of poor system performance and long time consumption associated with large-scale information recalls in multi-channel recalls. By predicting the batches of this recall, it ensures that after the recall is written to the cache, the data in the cache can satisfy the paging request, thereby providing the user with the ability to freely flip pages, in line with user usage habits. Furthermore, the method provided in the embodiments of the present application recalls all information in the database, while a one-time recall fusion solution can only recall a portion of the data in the database, resulting in a less rich recall of information.
[0097] Figure 3 The overall flow chart of the information recommendation method provided by the embodiment of the present application is shown in FIG. Figure 3 The whole process includes the following steps:
[0098] 1. Upon receiving a query request from a user account, the server queries the shared context information of the user account;
[0099] 2. The server obtains the shared context information of the user account.
[0100] 3. The server determines the recall batches for the main recall and the recall batches for the bypass recall based on the recall quantity for the main recall method, the recall quantity for the bypass recall method, and the page information of the target page to be displayed.
[0101] 4. Check whether the information that needs to be recalled is stored in the cache.
[0102] 5. If the cache stores the information that needs to be recalled, filter out the information that is not in the cache and needs to be recalled.
[0103] 6. Recalls are conducted using the main recall method and the bypass recall method.
[0104] 7. Fuse the information of the same recall batch in the main recall mode and the bypass recall mode, and store the fusion result in the cache.
[0105] 8. If this recall operation is the first recall operation for the user account, generate shared context information for the user account and store the shared context information in the cache.
[0106] 9-10. According to the requested paging information, obtain the corresponding information from the cache and obtain the paging result.
[0107] All of the above optional technical solutions can be combined in any way to form optional embodiments of the present application, and will not be described in detail here.
[0108] Please refer to Figure 4, which shows a schematic diagram of the structure of an information recommendation device provided in an embodiment of the present application. The device can be implemented through software, hardware, or a combination of both, and becomes all or part of an electronic device. The device includes:
[0109] Receiving module 401, configured to receive a query request sent by a user account, the query request including page information of a target page requested to be displayed;
[0110] A first determining module 402 is configured to determine a recall batch for each recall method based on page information of the target page, information stored in the cache, and the recall quantity corresponding to each recall method;
[0111] The recall module 403 is used to perform an information recall operation based on the recall batch of each recall method to obtain candidate information recalled using each recall method;
[0112] A second determining module 404 is configured to determine target information corresponding to a target page based on candidate information retrieved using multiple retrieval methods;
[0113] The sending module 405 is used to send the target information to the user account.
[0114] In another embodiment of the present application, the page information includes the number of pages, and the first determination module is used to determine the information carried by the target page based on the number of pages and the amount of information displayed on each page; and determine the recall batch for each recall method based on the information carried by the target page, the information stored in the cache, and the recall quantity corresponding to each recall method.
[0115] In another embodiment of the present application, a first determination module is used to query whether the information carried by the target page is stored in the cache; when the information carried by the target page is stored in the cache, the number of information displayed on each page is subtracted from the number of information carried by the target page stored in the cache to obtain the number of information to be recalled; based on the recall number corresponding to each recall method and the number of information to be recalled, the recall batch for each recall method is determined.
[0116] In another embodiment of the present application, a first determination module is used to determine a reference recall batch for each recall method based on the amount of information displayed on each page and the recall amount corresponding to each recall method; query whether the information carried by the target page is stored in the cache; when the information carried by the target page is stored in the cache, subtract the amount of information carried by the target page stored in the cache from the amount of information displayed on each page to obtain the amount of information to be recalled; based on the amount of information to be recalled, correct the reference recall batch for each recall method to obtain a recall batch for each recall method.
[0117] In another embodiment of the present application, the device further comprises:
[0118] A query module is used to query whether the shared information context of the user account is stored in the cache. The shared information context is the information shared when the user account performs multiple query operations on candidate information searched for the same search term;
[0119] The acquisition module is used to obtain the shared information context when the shared information context is stored in the cache;
[0120] The first determination module is used to determine the recall batch for each recall method based on the page information of the target page, the information stored in the cache and the recall quantity corresponding to each recall method when the number of target pages does not exceed the page limit based on the shared information context.
[0121] In another embodiment of the present application, the device further comprises:
[0122] A generating module, configured to generate a shared information context when no shared information context is stored in the cache;
[0123] The storage module is used to store the shared information context in the cache so that the shared information context can be obtained from the cache next time a query request for a user account is received.
[0124] In another embodiment of the present application, the second determination module is used to fuse the candidate information recalled by multiple recall methods in the same recall batch to obtain the fusion results of multiple recall batches; based on the amount of information displayed on each page, a number of information is obtained from the fusion results of multiple recall batches as the target information corresponding to the target page.
[0125] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0126] Figure 5 FIG2 shows a block diagram of an electronic device 500 according to an exemplary embodiment of the present application. Generally, the electronic device 500 includes a processor 501 and a memory 502 .
[0127] The processor 501 can be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 501 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state; the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 501 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 501 may also include an artificial intelligence processor, which is used to process computing operations related to machine learning.
[0128] The memory 502 may include one or more computer-readable storage media, which may be non-transitory computer-readable storage media, such as CD-ROMs (Compact Disc Read-Only Memory), ROMs, RAMs (Random Access Memory), magnetic tapes, floppy disks, and optical data storage devices. The computer-readable storage media may store at least one computer program, which, when executed, can implement the information recommendation method.
[0129] Of course, the electronic device described above may also include other components, such as input / output interfaces and communication components. The input / output interface provides an interface between the processor and a peripheral interface module, which may be an output device, an input device, etc. The communication component is configured to facilitate wired or wireless communication between the electronic device and other devices.
[0130] Those skilled in the art will understand that Figure 5 The structure shown in the figure does not constitute a limitation on the electronic device 500, and the electronic device 500 may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.
[0131] An embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one computer program, and when the at least one computer program is executed by a processor, the above-mentioned information recommendation method can be implemented.
[0132] An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it can implement the above-mentioned information recommendation method.
[0133] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. An information recommendation method, characterized in that: The method comprises: Receiving a query request sent by a user account, wherein the query request includes page information of a target page requested to be displayed; Determining a recall batch for each recall method based on page information of the target page, information stored in the cache, and a recall quantity corresponding to each recall method; Based on the recall batches of each recall method, perform information recall operations to obtain candidate information recalled using each recall method; Determining target information corresponding to the target page based on candidate information recalled using multiple recall methods; The target information is sent to the user account.
2. The method according to claim 1, characterized in that The page information includes the number of pages. Determining the recall batch for each recall method based on the page information of the target page, the information stored in the cache, and the recall quantity corresponding to each recall method includes: Determining the information carried by the target page based on the number of pages and the amount of information displayed on each page; The recall batch for each recall method is determined based on the information carried by the target page, the information stored in the cache, and the recall quantity corresponding to each recall method.
3. The method according to claim 2, characterized in that The determining of the recall batch for each recall method based on the information carried by the target page, the information stored in the cache, and the recall quantity corresponding to each recall method includes: Querying whether the cache stores information carried by the target page; When the cache stores information carried by the target page, the number of information displayed on each page is subtracted from the number of information carried by the target page stored in the cache to obtain the number of information to be recalled; Based on the recall quantity corresponding to each recall method and the amount of information to be recalled, the recall batch for each recall method is determined.
4. The method according to claim 2, characterized in that The determining of the recall batch for each recall method based on the information carried by the target page, the information stored in the cache, and the recall quantity corresponding to each recall method includes: Based on the amount of information displayed on each page and the recall quantity corresponding to each recall method, determine the reference recall batch for each recall method; Querying whether the cache stores information carried by the target page; When the information carried by the target page is stored in the cache, the amount of information displayed on each page is subtracted from the amount of information carried by the target page stored in the cache to obtain the amount of information to be recalled; Based on the amount of information to be recalled, the reference recall batch for each recall method is corrected to obtain the recall batch for each recall method.
5. The method according to claim 1, wherein Before determining the recall batch for each recall method based on the page information of the target page, the information stored in the cache, and the recall quantity corresponding to each recall method, the method further includes: querying whether the shared information context of the user account is stored in the cache, where the shared information context is information shared when the user account performs multiple query operations on candidate information searched for the same search term; When the shared information context is stored in the cache, acquiring the shared information context; When it is determined based on the shared information context that the number of pages of the target page does not exceed the page limit, the operation of determining the recall batch of each recall method is executed based on the page information of the target page, the information stored in the cache and the recall quantity corresponding to each recall method.
6. The method according to claim 5, characterized in that The method further comprises: When the shared information context is not stored in the cache, the shared information context is generated and stored in the cache, so that the shared information context can be obtained from the cache next time a query request for the user account is received.
7. The method according to claim 1, characterized in that The determining of target information corresponding to the target page based on the candidate information recalled using multiple recall methods includes: The candidate information recalled by multiple recall methods in the same recall batch is integrated to obtain the integration results of multiple recall batches; Based on the amount of information displayed on each page, the amount of information is obtained from the fusion results of multiple recall batches as target information corresponding to the target page.
8. An information recommendation device, characterized in that: The device comprises: A receiving module, configured to receive a query request sent by a user account, wherein the query request includes page information of a target page requested to be displayed; A first determining module is configured to determine a recall batch for each recall method based on page information of the target page, information stored in the cache, and a recall quantity corresponding to each recall method; A recall module is used to perform information recall operations based on the recall batches of each recall method, and obtain candidate information to be recalled using each recall method; A second determining module is configured to determine target information corresponding to the target page based on candidate information recalled using multiple recall methods; A sending module is used to send the target information to the user account.
9. An electronic device, characterized in that: The system comprises a processor and a memory; the memory stores at least one program code; the at least one program code is used to be called and executed by the processor to implement the information recommendation method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores at least one computer program, and when the at least one computer program is executed by a processor, it can implement the information recommendation method according to any one of claims 1 to 7.
11. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the information recommendation method according to any one of claims 1 to 7 can be implemented.