Information processing method and related device

Abstract

Provided in the embodiments of the present application are an information processing method and a related device. The method comprises: displaying an information retrieval interface, wherein the information retrieval interface is used for performing information retrieval on the basis of a query; displaying a subscription field, wherein the subscription field indicates a configured query; and outputting attention information related to the subscription field. The embodiments of the present application can enrich information retrieval modes, improve the flexibility of information retrieval, and can effectively adapt to a scenario in which an object continuously pays attention to information.

Description

Information processing method and related equipment

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on August 23, 2023, with application number 202311070020.1 and application name “A Method for Information Processing and Related Equipment”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of Internet technology, and specifically to an information processing method and related equipment, and in particular to an information processing method, an information processing device, a computer device, a computer-readable storage medium, and a computer program product. Background Art

[0003] Currently, subjects (e.g., users) can enter a search field (query) through the information retrieval interface provided by the client to retrieve information related to that search field. Furthermore, the information retrieval interface can record the subject's search history. When the subject re-enters the information retrieval interface, the subject can display the historical search fields from recent searches (e.g., the last few days). The subject can click on a historical search field to initiate a re-search of that historical search field to obtain business recall results. In practice, existing information retrieval systems have been found to be too simplistic and inflexible, failing to adapt to scenarios where subjects continuously seek information.

[0004] Summary of the Invention

[0005] The embodiments of the present application provide an information processing method and related equipment, which can enrich information retrieval methods, improve the flexibility of information retrieval, and better adapt to scenarios where the subject continuously pays attention to information.

[0006] In one aspect, an embodiment of the present application provides an information processing method, the method comprising:

[0007] Displaying an information retrieval interface, which is used to retrieve information based on a search field;

[0008] Displays the subscription field, which refers to the configured search field;

[0009] Outputs the follow information related to the subscription field.

[0010] In one aspect, an embodiment of the present application provides an information processing device, comprising:

[0011] A display unit, used to display an information retrieval interface, which is used to retrieve information based on a search field;

[0012] The display unit is further used to display a subscription field, where the subscription field refers to a configured search field;

[0013] The processing unit is configured to output attention information related to the subscription field.

[0014] In one aspect, an embodiment of the present application provides a computer device, comprising:

[0015] a processor suitable for executing a computer program;

[0016] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned information processing method is implemented.

[0017] On the one hand, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and the computer program is loaded by a processor and executes the above-mentioned information processing method.

[0018] On the one hand, an embodiment of the present application provides a computer program product, which includes a computer program or computer instructions, and when the computer program or computer instructions are executed by a processor, the above-mentioned information processing method is implemented.

[0019] In the embodiment of the present disclosure, an information retrieval interface is displayed, which is used to retrieve information based on the search field. It can be seen that information retrieval based on the search field can be realized by displaying the information retrieval interface. A subscription field is displayed, which refers to a configured search field. It can be seen that the object can configure the search field as a subscription field according to its own continuous attention needs. By displaying the subscription field, the configured search field that needs to be continuously paid attention to can be presented, and the search field that the object is continuously interested in can be learned. The attention information related to the subscription field is output. In this way, information that the object is continuously interested in can be continuously retrieved and output through the subscription field. While enriching the information retrieval method, the flexibility of information retrieval is improved, so that it can better adapt to the scenario where the object continuously pays attention to information. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] FIG1 is an architectural diagram of an information processing system provided by an exemplary embodiment of the present application;

[0021] FIG2a is a schematic diagram of an information processing flow provided by an exemplary embodiment of the present application;

[0022] FIG2 b is a schematic diagram of a process of subscribing to update notifications provided by an exemplary embodiment of the present application;

[0023] FIG3a is a flow chart of an information processing method provided by an exemplary embodiment of the present application;

[0024] FIG3 b is a schematic diagram of outputting attention information related to a subscription field provided by an exemplary embodiment of the present application;

[0025] FIG4a is a schematic diagram showing N subscription fields provided by an exemplary embodiment of the present application;

[0026] FIG4 b is a schematic diagram showing N subscription fields according to another exemplary embodiment of the present application;

[0027] FIG4c is a schematic diagram showing N subscription fields provided by another exemplary embodiment of the present application;

[0028] FIG4 d is a schematic diagram of a folded display of N subscription fields provided by an exemplary embodiment of the present application;

[0029] FIG5a is a schematic diagram of outputting multiple documents matching a subscription field in a fixed area according to an exemplary embodiment of the present application;

[0030] FIG5 b is a schematic diagram of outputting a content update notification in a floating area according to an exemplary embodiment of the present application;

[0031] FIG6a is a schematic diagram of setting a subscription field according to an exemplary embodiment of the present application;

[0032] FIG6 b is a schematic diagram of setting a subscription field provided by another exemplary embodiment of the present application;

[0033] FIG6c is a schematic diagram of setting a subscription field provided by yet another exemplary embodiment of the present application;

[0034] FIG6 d is a schematic diagram of setting a subscription field provided by another exemplary embodiment of the present application;

[0035] FIG6e is a schematic diagram of setting a subscription field provided by an exemplary embodiment of the present application;

[0036] FIG7a is a schematic diagram of deleting a subscription field word provided by an exemplary embodiment of the present application;

[0037] FIG7 b is a schematic diagram showing multiple documents matching a subscription, provided by an exemplary embodiment of the present application;

[0038] FIG7c is a schematic diagram of arranging and displaying multiple documents provided by an exemplary embodiment of the present application;

[0039] FIG7 d is a schematic diagram of clustering and displaying multiple documents provided by an exemplary embodiment of the present application;

[0040] FIG7e is a schematic diagram of displaying browsed documents and unbrowsed documents in different areas of an information retrieval interface provided by an exemplary embodiment of the present application;

[0041] FIG7f is a schematic diagram showing browsed documents and unbrowsed documents provided by an exemplary embodiment of the present application;

[0042] FIG8a is a schematic diagram of a content update notification provided by an exemplary embodiment of the present application;

[0043] FIG8b is a schematic diagram of a content update notification provided by another exemplary embodiment of the present application;

[0044] FIG9 is a schematic structural diagram of an information processing device provided by an exemplary embodiment of the present application;

[0045] FIG10 is a schematic structural diagram of a computer device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0046] First, the technical terms involved in the embodiments of this application are introduced.

[0047] 1. Search Fields and Subscription Fields

[0048] The search field (query) refers to a character string used for information retrieval. The search field may include keywords, such as the search field may include a person's name, food, game name, etc., or the search field may include a sentence, such as the search field may include "What food is there in XX city?" The embodiment of the present application does not limit this.

[0049] A subscription field is a configured search field that requires continuous attention (i.e., attention over a certain period of time). Configuring a subscription field can be understood as applying attention to a search field that requires continuous attention. By configuring a search field as a subscription field, you can trigger continuous attention to information related to the subscription field, thereby providing a continuous or regularly updated information retrieval service for the object.

[0050] 2. Pay attention to information

[0051] The information of interest refers to some information related to the subscription field. The information of interest can be understood as information that is obtained by continuously (real-time or scheduled) searching for information through the subscription field and that the object is continuously interested in. The information of interest includes but is not limited to: business recall results obtained by searching for information based on the subscription field, content update notifications related to the subscription field, etc. Among them, the business recall result refers to the result of taking measures (such as information retrieval) to recall or extract certain documents based on specific business needs (such as information retrieval needs for the subscription field). The business recall result may include multiple documents that match the subscription field. The matching here means that the content of the document has a high match with the subscription field (such as higher than a preset threshold). These multiple documents may include retrieved documents and / or recommended documents. The so-called retrieved document refers to the document that matches the subscription field retrieved by searching the recall link; the so-called recommended document refers to the document corresponding to the subject or tag mapped to the subscription field retrieved by the recommendation recall link. In the embodiment of the present application, the function of the search recall link is mainly to directly retrieve information based on the subscription field. The function of the recommendation recall link is mainly to first map and match the subscription field, obtain the subject or label mapped to the subscription field, and then perform information retrieval based on the mapped subject or label. The content update notification related to the subscription field is used to prompt that the information related to the subscription field (i.e., the information of interest) has been updated. The content update notification may include at least one of the following: the number of updated documents and the reason for recommendation. The number of updated documents refers to the number of updated documents, for example: the number of updated documents is 3, which means that there are 3 updated documents. The reason for recommendation may include: the document content type of the updated document. Schematically, if the subscription field includes "model", then the document content type of the updated document can be the model technical principle, model introduction, etc. The reason for recommendation may be: "There is a document update that describes the model technical principle"; or the reason for recommendation may be: "There is a document update that describes the new model and related introductions."

[0052] 3. Document topics and tags

[0053] Documents may include articles, videos, audios, images, web pages, etc., and this application does not impose any restrictions on this. The subject of a document is used to describe the main content of the entire document. For example, the subject of a document may be the title of an article, the name of a video, etc. The tag of a document is used to describe the key information of the document, which may include type, characters, time, place, etc. The type here may be, for example, games, food, models, etc. For example, taking the document as a video, for the video "Have you eaten XX type of food in XX place", the subject of the video may be "Have you eaten XX type of food in XX place", and the tags of the video may be, for example, XX place, XX type of food, etc. For another example, for the article "Xiao Ming is good at playing XX games", the subject of the article may be "Xiao Ming is good at playing XX type of games", and the tags of the article may include: Xiao Ming, XX type of game.

[0054] It should be noted that in this application, the relevant data involved in the information processing process, for example: the object's behavior log (such as historical search fields, document browsing history, accounts followed by the object, etc.), the object's attribute information (such as nicknames, regions), etc. When the embodiments of this application are applied to specific products or technologies, the object's permission or consent must be obtained, and the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and comply with the principles of legality, legitimacy and necessity, and does not involve the acquisition of data types prohibited or restricted by laws and regulations. In some optional embodiments, the relevant data involved in the embodiments of this application are obtained after the object has been separately authorized. In addition, when obtaining the object's separate authorization, the purpose of the relevant data involved is indicated to the object.

[0055] The embodiment of the present application provides an information processing solution, the general principle of which is as follows: displaying an information retrieval interface, which is used to retrieve information based on a search field; displaying a subscription field, which refers to a configured search field; and outputting attention information related to the subscription field. Through the above method, not only can information retrieval be performed based on the search field in the information retrieval interface, but the object can also configure the search field as a subscription field according to its own continuous attention needs. By displaying the subscription field, the configured search field that needs to be continuously followed can be presented, and then the search field that the object is continuously interested in can be known; information can also be continuously retrieved through the subscription field to obtain and output information that the object is continuously interested in. While enriching the information retrieval method, the flexibility of information retrieval is improved, and it can better adapt to the scenario where the object continuously pays attention to information.

[0056] Next, the information processing system provided in the embodiments of the present application will be described.

[0057] Please refer to Figure 1, which provides an architectural diagram of an information processing system for an exemplary embodiment of the present application. The information processing system includes a terminal 101, a terminal 102... and more terminals. The present application does not limit the number of terminals. The information processing system also includes a server 103. Of course, the number of servers can also be multiple, and the present application still does not limit this. Among them, the terminal can be understood as the terminal used by the object browsing the document. Among them, any terminal in the information processing system can be directly or indirectly connected to the server 103 through wired or wireless communication, and any two terminals can exchange information through the server 103. The following is explained using any terminal 101 as an example.

[0058] The terminal 101 may provide an information retrieval interface, and any object using the terminal 101 may enter a search field in the information retrieval interface to perform information retrieval, and information matching the search field may be output through the information retrieval interface, and the matching information may be the retrieval result obtained by performing information retrieval through the search field. In addition, in an embodiment of the present application, an object may also configure a subscription field in the information retrieval interface, and the subscription field refers to a configured search field that requires continuous attention. Then, the terminal 101 may output attention information related to the subscription field. The terminal 101 may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart car-mounted device, a smart wearable device, etc., and the embodiment of the present application does not limit this.

[0059] Server 103 is used to provide information processing services for terminal 101. The information processing services here may include, but are not limited to: information retrieval based on search fields, providing customized services for subscription fields for objects, information retrieval based on subscription fields, and other services. Server 103 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be 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, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0060] Please refer to Figure 2a, which is a schematic diagram of an information processing flow provided in an embodiment of the present application; the information processing flow can be executed by server 103, and the information processing flow includes: 1) offline behavior log and document content understanding, which includes the recommendation calculation of candidate search fields, object information calculation, and document content understanding. 2) Subscription field recommendation and customization. 3) Subscription field search (i.e., information retrieval based on subscription fields). 4) Feedback on subscription search results. 5) Content update notification. The above steps are described in detail below.

[0061] 1) Offline behavior logs and document content understanding

[0062] 1.1 Recommendation calculation of candidate search fields

[0063] ① In the offline state, collect the subject's behavior logs by time period. The time period can be flexibly set. For example, the time period can be set by day, such as one day, two days, etc. Another example is the time period can be set by hour, such as one hour, two hours, etc. Another example is the time period can be set by week, such as one week, two weeks, etc. The offline state refers to the state in which the subject's terminal is disconnected from the internet or network. The subject's behavior logs include, but are not limited to: information retrieval history (e.g., the subject entered a historical search field to search for information within a historical time period), video playback history (e.g., videos the subject played historically), followed accounts (e.g., accounts the subject followed that posted a certain article or video), likes (e.g., the subject liked a certain article, video, or video account, etc.), and sharing and forwarding (e.g., the subject shared and forwarded a certain article or video). After obtaining the subject's behavior logs, content tag clustering technology can be used to calculate candidate search fields in the behavior logs to obtain one or more candidate search fields. Content tag clustering is a commonly used technique in information retrieval, used to cluster documents with similar topics or content, thereby better organizing and displaying the content of large numbers of documents. As an implementation, the processing logic for using content tag clustering includes: performing word segmentation on the content in the subject's behavior log (such as historical search fields, titles of historically browsed documents, etc.) to obtain one or more word segments; vectorizing the one or more word segments to obtain a vector for each word segment; performing cluster calculations on each word segment vector to obtain a label corresponding to each word segment, and using the label corresponding to each word segment as a candidate search field.

[0064] ② After obtaining the candidate search field, the candidate search field can be bound to the object and saved to the database. Binding the candidate search field to the object means associating the candidate search field with the object's object identifier (e.g., saving the association to the database). Schematically, the database may include a key-value database (KV database for short). In this case, the object's object identifier can be used as the key, and the candidate search field can be used as the value and stored in the database. The object identifier here can be a nickname, account number, etc.

[0065] It should be understood that embodiments of the present application can collect behavior logs for each object, determine candidate search fields corresponding to each object in the manner described above, and bind each object to the corresponding candidate search field. Then, when an object configures a subscription field, the server can directly recommend its corresponding candidate search field for selection. This improves the efficiency of subscription field configuration while providing personalized subscription services for each object based on the search field.

[0066] 1.2 Object Information Calculation

[0067] ① As described in 1.1, content tag clustering technology is used to perform content tag clustering processing on the collected behavior logs of the objects, and one or more tags can be obtained; then, the one or more tags can be used as object information. The so-called object information refers to information used to outline the image and characteristics of the object. It is a typical object role virtualized (abstracted) based on the real behavior of the object, and can be used to describe the common characteristics of a class of objects. The common characteristics may include basic attributes (such as region), interests and hobbies, etc. Schematically, if the time period is set at the daily level, that is, the behavior logs of the objects are collected at the daily level (such as one day), then the content tag clustering technology is used to perform content tag clustering processing on the collected behavior logs of the objects, and one or more tags can be obtained. The one or more tags can be determined as the object information of the object at the daily level.

[0068] ② Collect the behavior logs of the object in multiple time periods (such as multiple days), and generate the object information in multiple time periods according to ①. For example, if the time period is one day, the behavior logs of the object collected for multiple days (such as 100 days) are clustered by content tags to obtain the object information in these multiple days (such as 100 days). The object information in multiple time periods is integrated with the object attribute information to obtain the long-term object information of the object. The so-called long-term object information refers to the object information that will not change within a certain period of time (such as one year or two years). Object attribute information may include but is not limited to: gender, region, nickname, etc.

[0069] ③ Obtain the object information and behavior log of the object within the target time period, and perform content label clustering processing on the object information and behavior log within the target time period to obtain the real-time interest needs of the object. The real-time interest needs are used to indicate the focus direction of the object within the target time period. The target time period can be the last three days, the last three hours, etc., and the embodiment of the present application does not impose any restrictions; the "last" here refers to the system time when the object information and behavior log of the object within the target time period is obtained, that is, the last three days refer to the three days before the system time, and the last three hours refer to the three hours before the system time. Schematically, the object information and behavior log of the object within the last three days are obtained. The behavior log, for example, records which entertainment news the object has read within these three days; the object information and behavior log of the object within these three days are clustered with content labels to obtain the real-time interest needs of the object. The real-time interest needs are used to indicate that the object has recently paid more attention to entertainment information.

[0070] It should be understood that both long-term object information and real-time interest requirements can be applied in the process of information retrieval based on subscription fields as a basis for filtering and sorting the retrieved information (such as retrieved documents).

[0071] 1.3. Understanding the Document Content

[0072] This part includes: establishing the corresponding relationship between browsed documents and objects (ie generating historical browsing records of objects), building a search document library and a recall recommendation document library.

[0073] 1.3.1) Generate historical browsing records of the object.

[0074] An object's behavior log may include a document browsing history. A document browsing history refers to a record of documents that an object has browsed. For example, a document browsing history records which documents an object has browsed and when. These previously browsed documents are referred to as viewed documents. That is, the document browsing history includes viewed documents. In embodiments of the present application, fingerprint calculation can be performed on the viewed documents to obtain fingerprint features corresponding to the viewed documents. Fingerprint calculation is a technique used to determine the uniqueness of document content. By performing fingerprint calculation on a document, a unique identifier (fingerprint feature) can be generated to represent the document's content. That is, the document's fingerprint feature is used to uniquely identify a document. The corresponding viewed document can be found using the fingerprint features of the viewed documents. In one implementation, the process of performing fingerprint calculation on the viewed documents to obtain fingerprint features corresponding to the viewed documents may include extracting multiple content features from the viewed documents (the content features herein may include, but are not limited to, document keywords, sentences within the document, etc.), performing hash calculation on the multiple content features using a hash algorithm to obtain hash values ​​corresponding to each content feature, and then merging the hash values ​​corresponding to each content feature to obtain the fingerprint features of the viewed documents. It should be noted that for subsequent fingerprint calculations, please refer to the fingerprint calculation process in the browsed document.

[0075] After obtaining the fingerprint features corresponding to the browsed documents, as an implementation method, the browsed documents and the corresponding fingerprint features can be used as the historical browsing records of the object, and a browsed document filter can be constructed based on the historical browsing records; schematically, the browsed document filter can be a Bloom filter with the length of the browsed document list as the size. As another implementation method, the browsed documents and the corresponding fingerprint features can be written into the database as the historical browsing records of the object. Schematically, if the database is a KV database, the object identifier of the object can be used as the key, and the browsed document and its corresponding fingerprint features can be used as the value and written into the database. The following two examples explain this in detail:

[0076] Example 1: Assume that browsed documents include browsed articles, and that browsed document filters include browsed article filters. The process of generating a historical browsing record for a browsed article object may include: performing fingerprint calculation on the body of the browsed article, and constructing a browsed article filter based on the fingerprint features corresponding to the browsed article (for example, the browsed article filter may be a Bloom filter using the length of the browsed article list as the size). Alternatively, assume that the database includes a browsed article KV database; the process of generating a historical browsing record for a browsed article object may include: constructing a browsed article KV database using the object identifier of the object as the key and the fingerprint features of the browsed article as the value.

[0077] Example 2: Assume that the viewed documents include viewed videos, and the viewed document filter includes a viewed video filter. The process of generating a historical browsing record for the viewed video object may include: extracting frames of the viewed video at a preset rate (the preset rate can be set according to demand, such as extracting one frame every 2 seconds), performing fingerprint calculation on the extracted video frames, obtaining fingerprint features corresponding to the video frames, and constructing a viewed video filter based on the fingerprint features corresponding to the video frames (illustratively, the viewed video filter can use a Bloom filter with the length of the viewed video list as the size). Alternatively, assume that the database includes a viewed video KV database; the process of generating a historical browsing record for the viewed video object may include: constructing a viewed video KV database with the object identifier of the object as the key and the fingerprint features of the viewed video as the value.

[0078] It should be understood that the historical browsing records of the generated object can be used as a basis for filtering documents that match the subscription field during information retrieval based on the subscription field. This can prevent the object from receiving the browsed document again, avoid redundant recommendations, and reduce the information interaction pressure between the server and the terminal.

[0079] 1.3.2) Build a searchable document library.

[0080] In one implementation, the full amount of documents in the search document library can be directly obtained, and the full amount of documents refers to all documents in the search document library; fingerprint calculation is performed on each obtained document to obtain the fingerprint features of each document; each obtained document is used as a candidate retrieval document, and the fingerprint features of each document are written into the search document library as the document features of the corresponding candidate retrieval document. Here, the search document library refers to a database that provides data support for information retrieval, that is, the embodiment of the present application performs information retrieval in the search document library. The search document library can include various types of documents, including but not limited to pre-set documents, documents uploaded by received objects, or documents obtained from the Internet, etc., and the search document library supports continuous updates.

[0081] In another implementation, the search document library can be divided into a full search document library (gob) and a local search document library according to the document publication time, that is, the search document library includes a full search document library (gob) and a local search document library. The document publication time refers to the time point when the document is officially released or announced. The full search document library contains all the documents in the search document library and the fingerprint features of each document. At this time, the construction method of the full search document library includes: obtaining all the documents in the search document library, and performing fingerprint calculation on each obtained document to obtain the fingerprint features of each document, and using each obtained document as a candidate retrieval document, and using the fingerprint features of each document as the document features of the corresponding candidate retrieval document; based on the candidate retrieval documents and the document features of the candidate retrieval documents, constructing the full search document library.

[0082] The local search document library is used to store local documents. Local documents are a concept relative to the full amount of documents, that is, local documents refer to a part of the full amount of documents. In this embodiment, local documents refer to: documents in the search document library whose publication time is within the second preset time period. The second preset time period can be a time period corresponding to the week level, hour level, or day level. Schematically, local documents can refer to documents published within a certain week, or documents published within a certain two weeks. The construction method of the local search document library includes: obtaining documents whose publication time is within the second preset time period, and performing fingerprint calculation on each document within the second preset time period to obtain the fingerprint features of each document within the second preset time period, and using each document within the second preset time period as a local search document, and using the fingerprint features of each document within the second preset time period as the document features of the corresponding local search document; constructing a local search document library based on the local search document and the document features of the local search document.

[0083] Among them, the local search document library includes one or more of the latest search document library (fob) and the latest search document library (latest ob). In an embodiment of the present application, various types of local search document libraries (immediate new search document library (fob) and latest search document library (latest ob)) can be constructed according to their respective time periods. Schematically, the latest search document library can be constructed with a weekly time period, such as the latest search document library can include documents published in the last week or documents published in the last two weeks. The latest search document library can be constructed with a daily time period or an hourly time period, such as the latest search document library includes documents published in the last day or documents published in the last hour. Various types of ob are constructed according to their respective time periods, which can facilitate the subsequent acquisition of retrieval documents that match the subscription fields from the corresponding ob based on different needs, thereby improving information retrieval efficiency. When the new search document library (fob) is constructed with a weekly time period (such as one week), the documents published in the most recent week can be obtained, and fingerprints of the obtained documents can be calculated. The calculated fingerprint features are used as the document features of the obtained documents, and the obtained documents and corresponding document features are written into the new search document library (fob). When the latest search document library (latest ob) is constructed with a daily time period (such as one day), the documents published in the most recent day can be obtained, and fingerprints of the obtained documents can be calculated. The calculated fingerprint features are used as the document features of the obtained documents, and the obtained documents and corresponding document features are written into the latest search document library (latest ob).

[0084] 1.3.3) Construct a recall recommendation document library: obtain documents from the most recent period of time (such as the last month) to obtain P candidate recommendation documents, where P is a positive integer. Offline score the P candidate recommendation documents according to the recommendation impact factor to obtain the score corresponding to each candidate recommendation document in the P candidate recommendation documents; sort each of the P candidate recommendation documents in descending order according to the score corresponding to each candidate recommendation document to obtain the sorted P candidate recommendation documents, and obtain the top Y candidate recommendation documents from the sorted P candidate recommendation documents, and construct a recall recommendation document library based on the top Y candidate recommendation documents. Here, the recall recommendation document refers to a database that provides data support for information recommendation for an object, that is, the embodiment of the present application performs information retrieval and recommendation based on the subscription field in the recall recommendation document library. The recall recommendation document library can include various high-quality documents, which may include but are not limited to pre-set documents, documents uploaded by the object, or documents obtained from the Internet, etc., and the recall recommendation document library supports continuous updates. The so-called high-quality documents generally refer to documents that are of high quality (the quality of the text is higher than the preset quality threshold), complete, clear, accurate, and reliable (that is, the content source is reliable, has clear basis or data support, and can be relied upon).

[0085] Recommendation impact factors may include at least one of the following: document originality, document activity, document likes, document shares, document forwarding, document text quality, document publication time, etc. Recommendation impact factors can help to include some high-quality documents in the recall recommendation documents.

[0086] ① Document originality: refers to an original document that is independently created and not referenced by other documents. In an embodiment of the present application, offline scoring of P candidate recommended documents according to the recommendation impact factor to obtain the score corresponding to each candidate recommended document in the P candidate recommended documents may include: scoring the document originality of the input P candidate recommended documents through the originality determination model to obtain the document originality score of each candidate recommended document in the P candidate recommended documents, and using the document originality score of each candidate recommended document as the score of each candidate recommended document. The originality determination model refers to a model or system used to evaluate whether a work or document is original. The originality determination model usually uses natural language processing technology and machine learning algorithms to determine the degree of originality of the document by comparing the similarity of the document with existing works and possible plagiarism. Indicatively, the originality determination model may include but is not limited to: a model based on TF-DF (word frequency and inverse document frequency), a model based on N-gram (word fragment), a model based on word embedding, a model based on convolutional neural network, etc., which is not limited in the present application. ② Document activity: may refer to the frequency of document browsing. The more frequently a candidate document is viewed, the more active it is, and the higher its score. ③ Document Likes: This refers to the number of times a document has been liked. Liking is a form of social interaction, expressing approval, support, or appreciation for a document's content. The higher the number of likes a candidate document receives, the higher its score. ④ Document Shares: This refers to the number of times a document has been shared. Sharing is the process of disseminating content, information, knowledge, or resources through various channels to more people, allowing them to view, understand, or benefit from it. The higher the number of times a candidate document has been shared, the higher its score. ⑤ Document Forwards: This refers to the number of times a document has been forwarded. Forwarding refers to sending a document to others to expand its reach and allow more people to view and understand its content. The higher the number of forwards a candidate document receives, the higher its score. ⑥ Document Content Quality: This refers to the quality and accuracy of the document's main content. The quality of the document body can be evaluated from the perspectives of content accuracy, completeness, logic, structural clarity, etc. In the embodiment of the present application, the document quality of P candidate recommended documents can be scored by a document quality assessment model to obtain a score for each candidate recommended document. A document quality assessment model refers to a model or method for evaluating document quality. Schematically, a document quality assessment model may include but is not limited to: machine learning models, neural network models, text classification models, etc. ⑦ Document publication time: refers to the time point when the document is officially released or published.The later the candidate recommendation document is published, the higher the score of the candidate recommendation document. That is, the score of the most recently published candidate recommendation document will be higher, so that the object can obtain the latest recommended document during subsequent information retrieval.

[0087] It should be understood that the above-mentioned recommendation influencing factors can be flexibly selected for use alone or in combination. For example, the above-mentioned methods ① and ② can be used in combination. For another example, the above-mentioned methods ① and ③ can also be used in combination, or the above-mentioned ①②⑤ can also be used in combination. Schematically, the proportion of each recommendation influencing factor can be set. For example, the recommendation influencing factors include document originality and document body quality. The proportion of document originality can be set to 30%, and the proportion of document body quality can be set to 70%. The scores of candidate recommended documents are comprehensively determined by setting weights. By ranking candidate recommended documents through a combination of multiple recommendation influencing factors, better and higher-quality candidate recommended documents can be selected as recommended documents for the object.

[0088] As an implementation method, based on the first Y candidate recommended documents, constructing a recalled recommended document library may include: performing document topic calculation or content understanding on the first Y candidate recommended documents to obtain the topic to which each candidate recommended document belongs, and performing label calculation or content understanding on the Y candidate recommended documents to obtain the label to which each candidate recommended document belongs, using the topic and / or label as the tag of each candidate recommended document, using the topic and / or label as the retrieval keyword key, and taking the candidate recommended documents (or candidate recommended document list) that hit the corresponding topic and / or label as the index content to construct a recalled recommended document library, which includes at least one of the following: candidate recommended documents corresponding to different topics, candidate recommended documents corresponding to different labels; illustratively, the value of Y can be in the millions or tens of millions, for example: the value of Y can be one million or ten million.

[0089] 1.3.4) Furthermore, in order to facilitate online recommendation recall, the embodiment of the present application can further establish a mapping relationship between subscription fields and topics, and establish a mapping relationship between subscription fields and tags. When recalling recommended documents, the subscription field can be used as a key to search for the mapped topic or tag, so that the corresponding candidate recommended document (or candidate recommended document list) can be determined based on the topic or tag mapped to the subscription field. The so-called mapping relationship means that the topic or tag mapped to the subscription field can be found through the subscription field. The process of establishing the mapping relationship between the subscription field and the topic includes: determining the similarity between the subscription field and the topic, and when the similarity between the subscription field and the topic is greater than the preset similarity, establishing the mapping relationship between the subscription field and the topic; in addition, the process of establishing the mapping relationship between the subscription field and the tag: determining the similarity between the subscription field and the tag, and when the similarity between the subscription field and the tag is greater than the preset similarity, establishing the mapping relationship between the subscription field and the tag. Of course, the mapping relationship between the subscription field and the topic, and the mapping relationship between the subscription field and the title can also be directly pre-set. For example, directly set the mapping relationship between the subscription field "model" and the label "ChatGPT model" (Generative Pre-trained Transformer).

[0090] 2) Subscription field recommendation and customization

[0091] This embodiment of the application adds a subscription field recommendation and customization module to the information retrieval interface. The subscription field recommendation and customization module supports configuring subscription fields, editing configured subscription fields (such as deleting, modifying, etc.), providing one or more candidate search fields for objects to select, and storing the subscription fields configured by objects. The specific scheme of the subscription field recommendation and customization module is as follows:

[0092] When an object is configured with a subscription field, embodiments of the present application can provide the following methods for configuring the subscription field: ① Since the object and the corresponding candidate search field are already stored in the database, the information retrieval interface includes a search subscription option, and the object can trigger the search subscription option. When the search subscription option is triggered, one or more candidate search fields corresponding to the object can be obtained from the above database, and one or more candidate search fields can be displayed for the object to select. After the object selects a candidate search field, the candidate search field can be configured as a subscription field. ② The object directly configures a custom subscription field. Specifically, the object can enter a search field in the search subscription input window, thereby configuring the entered search field as a subscription field.

[0093] In addition, subscription fields support editing, which can include modification or deletion. For example, when an object is no longer interested in a subscription field, it can delete the uninterested subscription field. In response to a delete operation on the subscription field, the subscription field can be deleted. Deleting a subscription field means that the corresponding search field is unsubscribed, and the object no longer needs to continue to pay attention to information related to the search field.

[0094] In one implementation, operation records such as selecting a candidate search field as a subscription field, configuring a custom subscription field, and deleting a subscription field can be recorded in a custom subscription field database (such as a custom subscription field KV database) corresponding to the object. Specifically, the object identifier of the object is used as the key, and the operation record corresponding to the object is stored as the value in the custom subscription field database. In another implementation, the object and its configured subscription field are saved in a target subscription field database (such as a KV database). Specifically, the object identifier of the object is used as the key, and the subscription field configured by the object is stored as the value in the target subscription field database. Subsequently, relevant attention information can be output for the object based on the subscription field in the target subscription field database.

[0095] It should be understood that any object can configure one or more subscription fields according to its own interests, so as to continuously obtain the attention information related to these subscription fields.

[0096] 3) Online subscription search process:

[0097] 3.1) Display the subscription field in the terminal used by the object. The object can trigger (such as clicking, double-clicking, etc.) a subscription field to enter the information retrieval process. In response to the triggering operation on the subscription field, the server 103 can call the search background CGI (Common Gateway Interface). Schematically, the search background CGI can be a websearch CGI; the search main scheduling service program mainmixer is called through the search background CGI.

[0098] 3.2) The main scheduling service program mainmixer calls the query understanding service program, the object information service program, and the object personalized data acquisition service program in parallel. Among them, the query understanding service program is used to perform query understanding on the subscription field to realize search recall. The query understanding of the subscription field here includes: performing query segmentation on the subscription field, and building a syntax tree based on the segmentation result, and performing intent recognition on the subscription field to obtain the intent recognition result, thereby completing the query understanding. The object information service program is used to obtain object information, and the object information is used for subsequent interest and demand matching sorting of documents retrieved by information retrieval and for search recall of personalized documents. The object personalized data acquisition service program is used to obtain personalized data, and the personalized data is used for interest and demand matching sorting of documents retrieved by information retrieval and for search recall of personalized documents.

[0099] 3.3) The main scheduling service program, mainmixer, can initiate a recall call for documents based on the subscription field within a first preset time period (e.g., the last month or two weeks) to obtain a business recall result. The so-called recall call refers to information retrieval based on the subscription field to obtain a business recall result. The business recall result may include multiple documents, including one or more of the retrieved document and the recommended document. Accordingly, the embodiment of the present application involves two links in initiating a recall call for these two types of documents: one is a search recall link corresponding to the retrieved document, and the other is a recommendation recall link corresponding to the recommended document.

[0100] 3.4) The service recall result includes: a search recall result, and a method for obtaining the search recall result may include the following steps S11-S15:

[0101] S11. Search and recall: Using the subscription field as a query, through the search and recall link, retrieve M (M is a positive integer) candidate search documents that match the subscription field within a first preset time period from the search document library.

[0102] In one implementation, the search document library includes a full search document library, and M candidate search documents whose publication time is within a first preset time period and matches the subscription field are retrieved from the search document library through a search recall link. In another implementation, the search document library includes a full search document library, a timely search document library, and a latest search document library. According to the first preset time period, the ob to be used (i.e., the full search document library (gob), the timely search document library (fob), or the latest search document library (latest ob)) can be determined, and M candidate search documents matching the subscription field are retrieved from the determined ob. For example, if the first preset time period is the most recent day and the latest search document library is constructed with days as the time period, then the ob to be used can be determined to be the latest search document library (latest ob), and then M candidate search documents matching the subscription field can be retrieved from the latest ob.

[0103] Among them, the specific implementation of retrieving M candidate retrieval documents that match the subscription field within the first preset time period from the search document library can be: performing fingerprint calculation on the subscription field to obtain the fingerprint characteristics of the subscription field; respectively calculating the similarity between the fingerprint characteristics of the subscription field and the document characteristics of each candidate retrieval document in the search document library, and taking the candidate retrieval document corresponding to the similarity greater than the similarity threshold as the candidate retrieval document that matches the subscription field.

[0104] In one implementation, the embodiment of the present application can search and roughly sort the P candidate retrieval documents retrieved from the search document library to obtain M candidate retrieval documents that match the subscription field, where P is greater than or equal to M. Rough sorting refers to the process of preliminarily screening and sorting a large number of documents (such as P candidate retrieval documents) in the search document library to extract a portion of documents (such as M candidate retrieval documents) that have a high relevance to the subscription field. Schematically, retrieving M candidate retrieval documents that match the subscription field within a first preset time period from the search document library through a search recall link includes: retrieving P candidate retrieval documents whose document publication time is within the first preset time period and matches the subscription field from the search document library through a search recall link; sorting the P candidate retrieval documents in descending order according to the degree of matching with the subscription field, and determining the M candidate retrieval documents that match the subscription field based on the sorted P candidate retrieval documents. Specifically, the top M candidate retrieval documents can be selected from the sorted P candidate retrieval documents as the M candidate retrieval documents that match the subscription field.

[0105] S12. Based on the historical browsing history of the object, the M candidate retrieval documents are filtered to obtain L unbrowsed candidate retrieval documents. As an implementation method, a browsed document filter constructed based on the historical browsing history can be called to filter the M candidate retrieval documents to obtain L unbrowsed candidate retrieval documents. As another implementation method, fingerprint calculation is performed on the M candidate retrieval documents to obtain fingerprint features of each candidate retrieval document in the M candidate retrieval documents, and the fingerprint features are compared with the document features of each browsed document in the browsed document database. The candidate retrieval documents in the M candidate retrieval documents that match the fingerprint features of the browsed documents are filtered to obtain L unbrowsed candidate retrieval documents, where L is a positive integer and L is less than or equal to M.

[0106] S13. Sort the L unbrowsed candidate search documents to obtain the sorted L candidate search documents. Sorting here can be understood as performing refined sorting based on object information (i.e., search refinement). Refined sorting (i.e., search refinement) refers to the process of sorting search results based on a certain standard (e.g., object information). In one implementation, the L unbrowsed candidate search documents are sorted based on the object information. For example, if the object information indicates a high level of interest in food (the higher the interest, the more interested in the object), the candidate documents related to food in the L unbrowsed candidate search documents can be sorted higher, and the other candidate search documents in the L unbrowsed candidate search documents can be sorted lower.

[0107] S14. Perform personalized document retrieval and recall based on the personalized data of the object to obtain a personalized retrieval document. Personalized data includes, but is not limited to, documents liked, documents shared, documents forwarded, other objects with which the object has an object relationship, documents following other objects, and the like. Performing personalized document retrieval and recall based on the personalized data of the object to obtain a personalized retrieval document may include obtaining a document according to the personalized data and determining the obtained document as the personalized retrieval document.

[0108] For example, personalized data may include other objects with object relationships. These object relationships may include, but are not limited to, friend relationships and follow relationships in social applications. Based on the personalized data of an object, performing personalized document retrieval to obtain personalized retrieved documents may include obtaining documents viewed by other objects and using the obtained documents as personalized retrieved documents. For another example, if the personalized data includes followed accounts, performing personalized document retrieval to obtain personalized retrieved documents based on the personalized data of the object may include obtaining documents related to the followed account and using the obtained documents related to the followed account as personalized retrieved documents.

[0109] S15: Perform fusion sorting on the sorted L candidate search documents and the personalized search document to obtain a search recall result, wherein the search recall result may include one or more search documents.

[0110] As an implementation method, the fusion sorting process includes at least one of the following: ① Deduplication processing is performed on the sorted L candidate retrieval documents and personalized retrieval documents to obtain deduplication documents, and the deduplication documents are sorted to obtain search recall results. ② Based on the personalized retrieval documents, the order of the sorted L candidate retrieval documents is adjusted to obtain search recall results. Schematically, the personalized retrieval documents include document 1 and document 2, and the L candidate retrieval documents include document 1 and document 3. The order of document 1 in the sorted L candidate retrieval documents can be adjusted according to document 1 in the personalized document (such as adjusting document 1 to be at the front of the L candidate retrieval documents) to obtain a search recall result.

[0111] 3.5) For the recommended recall link, the service recall result may further include: a recommended recall result; and a method for obtaining the service recall result may include the following steps S21-S24:

[0112] S21. Recommendation recall: Using the subscription field as the index keyword (key), determine the subject or tag mapped to the subscription field in the recall recommendation document library provided by the recommendation recall link, and retrieve candidate recommendation documents matching the subscription field based on the mapped subject or tag in the recall recommendation document library provided by the recommendation recall link, so as to realize the recall of documents from multiple recall directions such as subject and tag. Schematically, the subscription field is "model", and the tag "ChatGPT model" mapped to the subscription field "model" is determined in the recall recommendation document library provided by the recommendation recall link. Based on the tag "ChatGPT model" in the recall recommendation document library, retrieve the candidate recommendation documents belonging to the tag "ChatGPT model", and determine the retrieved candidate recommendation documents belonging to the tag "ChatGPT model" as the candidate recommendation documents matching the subscription field.

[0113] S22. Roughly rank the candidate recommended subscription information corresponding to the topic or tag mapped to the subscription field. As an implementation method, a ranking parameter may be obtained, and the retrieved matching candidate recommended documents may be ranked according to the ranking parameter to obtain ranked candidate recommended documents. In this case, the ranking parameter may include: article quality, degree of match with the subscription field, etc. Taking the degree of match with the subscription field as an example, ranking the retrieved matching candidate recommended documents according to the ranking parameter to obtain ranked candidate recommended documents includes: ranking the retrieved matching candidate recommended documents in descending order of degree of match with the subscription field to obtain ranked candidate recommended documents. For example, assuming that the number of retrieved matching candidate recommended documents is 3, namely, Document 1, Document 2, and Document 3, and the degree of match between Document 1 and the subscription field is greater than the degree of match between Document 2 and the subscription field, and the degree of match between Document 2 and the subscription field is greater than the degree of match between Document 3 and the subscription field, then the ranked candidate recommended documents are Document 1, Document 2, and Document 3, respectively.

[0114] S23: Filter the sorted candidate recommended documents based on historical browsing records to obtain unbrowsed candidate recommended documents. The specific implementation of step S23 is similar to that of step S12 and will not be repeated here.

[0115] S24. Recommend and refine the unbrowsed candidate recommendation documents. As an implementation method, the unbrowsed candidate recommendation documents are re-sorted based on the object information and personalized data to obtain a recommendation recall result. The unbrowsed candidate recommendation documents after sorting can be used as recommended documents. The recommendation recall result can include one or more recommended documents. Specifically, according to the object information and personalized data, the unbrowsed candidate recommendation documents can be weighted to obtain the weights of the unbrowsed recommended documents. The unbrowsed candidate recommendation documents are re-sorted in descending order of the weights of the unbrowsed recommended documents to obtain a recommendation recall result. As an implementation method, according to the object information and personalized data, the unbrowsed candidate recommendation documents can be weighted by determining the unbrowsed candidate recommendation documents that the object is concerned about based on the object information and personalized data, and determining the weight of the unbrowsed candidate recommendation documents that the object is concerned about as a larger value, while the weights of other unbrowsed candidate recommendation documents are set to a smaller value.

[0116] It should be understood that historical browsing records include browsed documents. Documents can be videos, articles, audio, and so on. Therefore, in the embodiment of the present application, taking articles and videos as examples, a search recall link and a recommendation recall link corresponding to articles can be added; a search recall link and a recommendation recall link corresponding to videos can be added, thereby achieving information retrieval from both videos and articles based on the subscription field to obtain service recall results. The service recall results obtained by retrieving information from both videos and articles based on the subscription field can refer to the service recall results obtained by retrieving information based on the subscription field mentioned above, and the embodiment of the present application does not impose any limitations on this.

[0117] In addition, the embodiment of the present application can filter the browsed documents as in the above steps S12 and S23. Of course, the embodiment of the present application can also directly recall the browsed documents instead of filtering them and fold them when outputting. At this time, in the search recall link, after retrieving M candidate retrieval documents that match the subscription field within the first preset time period from the search document library through the search recall link, the M candidate retrieval documents can be directly sorted to obtain the sorted M candidate retrieval documents; based on the personalized data of the object, personalized retrieval document recall processing is performed to obtain personalized retrieval documents; the sorted M candidate retrieval documents and the personalized retrieval documents are fused and sorted to obtain the search recall result. In the recall recommendation link, the sorted candidate recommendation documents are sorted again according to the object information and personalized data to obtain the recommended recall result. In addition, the embodiment of the present application can determine the documents that the object is not interested in based on the behavior log of the object. If the behavior log includes documents corresponding to a certain topic or tag marked as uninteresting (such as documents with attention lower than the attention threshold), then such documents can also be filtered. At this point, you can refer to the processing process of browsed documents to process documents you are not interested in, which will not be elaborated here.

[0118] In steps 3.4) and 3.5), the browsed documents or documents of no interest are filtered and folded, and the latest content related to the subject's information or interests is actively displayed to meet the subject's continuous interest consumption needs under the subscription field, and to avoid the subject from repeatedly seeing browsed documents as much as possible.

[0119] It should be understood that in an embodiment of the present application, the recommendation recall link and the search recall link can be two independent links. Of course, the recommendation recall link and the search recall link are the same search link. At this time, in the search recall link, the subject or tag of the document can be used as the text field of the document, and the text field of the document can be vectorized to obtain the vector corresponding to the text field. Then, based on the vector corresponding to the text field of the document, a search index can be constructed to obtain a recall recommendation document library. When performing information retrieval, documents can be directly recalled from the recall recommendation document library and the search document library through the search recall link. Document recall is completed through a variety of search recall methods.

[0120] 3.6) Subscription recommendation fusion processing: After obtaining the search recall results and the recommendation recall results, the search recall results and the recommendation recall results are fused to obtain the business recall results, and the business recall results are returned to the mainmixer. As an implementation method, fusing the search recall results and the recommendation recall results may include: directly determining the search recall results and the recommendation recall results as business recall results. As another implementation method, fusing the search recall results and the recommendation recall results may include: deduplicating the retrieved documents in the search recall results and the recommended documents in the recommendation recall results to obtain the business recall results. As yet another implementation method, fusing the search recall results and the recommendation recall results may include: fusing and sorting the retrieved documents in the search recall results and the recommended documents in the recommendation recall results according to the fusion strategy to obtain the business recall results. The fusion strategy includes attention fusion display rules, subscription field matching fusion rules, and document diversity fusion display rules.

[0121] When the fusion display rule includes an attention fusion display rule, the retrieved documents in the search recall results and the recommended documents in the recommendation recall results are fused and sorted according to the fusion strategy to obtain a business recall result, including: fusion sorting the retrieved documents in the search recall results and the recommended documents in the recommendation recall results in order of attention from high to low to obtain a business recall result. For example, if the attention of the retrieved documents in the search recall results is greater than the attention of the recommended documents in the recommendation recall results, the retrieved documents are sorted in front of the recommended documents in order of attention from high to low to obtain a business recall result. The attention can be determined by the real-time interest needs of the object. For example, if the real-time interest needs indicate that the object's attention direction is food, then the retrieved document is a document about food, and then the attention of the retrieved document is greater than the attention of the recommended document.

[0122] When the fusion display rule includes a subscription field matching fusion rule, the subscription field matching fusion rule may include a matching degree with the subject of the subscription field or a matching degree with the tag of the subscription field. Taking the subscription field matching fusion rule including a matching degree with the subject of the subscription field as an example, according to the fusion strategy, the retrieved documents in the search recall results and the recommended documents in the recommendation recall results are fused and sorted to obtain a business recall result, including: determining the matching degree of the retrieved documents with the subject of the subscription field, and determining the matching degree of the recommended documents with the subject of the subscription field, and then fusion sorting the retrieved documents and the recommended documents in descending order of matching degree with the subject of the subscription field to obtain the business recall result.

[0123] When the fusion display rule includes a document diversity fusion display rule, such as the document diversity fusion display rule includes document diversity including the distribution of the number of documents of the same account, according to the fusion strategy, the search documents in the search recall results and the recommended documents in the recommendation recall results are fused and sorted, and the business recall result is obtained, including: according to the distribution of the number of documents of the same account, the search documents in the search recall results and the recommended documents in the recommendation recall results are fused to obtain the business recall result. Schematically, the number of search documents in the search recall results is 4, of which 3 search documents belong to the same account, and the number of recommended documents in the recommendation recall results is 2, and these two recommended documents belong to the same account. Assuming that the distribution of the number of documents of the same account is not higher than 2, at this time, according to the distribution of the number of documents of the same account, the 4 search documents and the 2 recommended documents can be fused. The fusion processing here can include: deleting any of the 3 search documents belonging to the same account, and the final business recall result includes 3 search documents and 2 recommended documents.

[0124] 3.7) After mainmixer obtains the business recall result, it can directly output the business recall result. Alternatively, after mainmixer obtains the business recall result, it performs mixed arrangement of multiple business results on the documents in the business recall result to obtain the mixed business recall result, and outputs the mixed business recall result. As an implementation method, the documents in the business recall result can be mixed with multiple business results based on real-time interest requirements and / or document diversity requirements. Schematically, based on the requirements of document diversity, the documents in the business recall result are mixed with multiple business results, such as if the documents in the business recall result include articles and videos; the requirements of document diversity include alternating arrangement of videos and articles, then the articles and videos in the business recall result can be alternately arranged based on the requirements of document diversity to obtain a mixed business recall result.

[0125] 3.8) Optionally, if special style customization is set in the subscription search, mainmixer calls the style customization service program (mergesvr) to customize the style of the mixed business recall results, and feeds the customized business recall results back to the terminal so that the terminal outputs the customized business recall results. Style customization includes but is not limited to: topic aggregation style, tag aggregation style, etc. Schematically, style customization includes topic aggregation style, which can aggregate documents in the business recall results by topic and display documents belonging to the same topic in one area.

[0126] 4) Subscribe to search result feedback

[0127] After the subject obtains the service recall result, it can call the feedback CGI (webquery CGI) to provide feedback on the browsed document while browsing the service recall result. In principle, it can mark documents of interest or disinterest. In response to the marking operation, the subject's feedback is recorded in the database (KV database). In the subscription search process 3.6) of merging and sorting the search recall results and the recommended recall results, the weight of the retrieved documents in the search recall results and the recommended documents in the recommended recall results can be adjusted based on the subject's feedback, or directly used to filter the retrieved documents in the search recall results and the recommended documents in the recommended recall results, etc. When the subject's feedback indicates that the subject is not interested in the document, the weight adjustment here can be: when the retrieved document has a high similarity with the marked document, the weight of the retrieved document or the recommended document can be reduced. When the subject's feedback indicates that the subject is interested in the document, the weight adjustment here can be: when the retrieved document has a high similarity with the marked document, the weight of the retrieved document or the recommended document can be increased.

[0128] 5) Content update notification:

[0129] When the content (document) associated with a subscription field is updated, a content update notification for the subscription field can be generated and output. This content update notification can remind the object of the content update in the subscription field it is interested in, facilitating its continued consumption (e.g., facilitating browsing and querying). Please refer to Figure 2b, which is a flowchart of a subscription update notification provided by an exemplary embodiment of the present application. The entire subscription update notification process is as follows:

[0130] 5.1) Subscription Notification Scheduling Service: The server's subscription notification dispatching (dispatch) service scans the target subscription field database of objects on a minute-by-minute basis, and scatters the dispatches over time based on the object dimension, adding them to the dispatch queue. This means that different objects are evenly scattered over time to avoid system pressure from update calculations. Scattered scheduling refers to a scheduling strategy that disrupts the original scheduling order through randomization or other means to improve system performance or fairness. For example, the subscription fields corresponding to 100 objects can be added to the dispatch queue every 1 or 2 minutes.

[0131] 5.2) Subscription search: The server can obtain the subscription field of the object from the dispatch queue and re-recall the document from the search recall link and the recommendation recall link. In one implementation, one or more local search documents that match the subscription field are retrieved from the local search document library through the search recall link; based on the historical browsing record of the object, the one or more local search documents retrieved in the search recall link are filtered to obtain a first filtering result. Since documents are updated quickly, there may be documents that have been updated and notified last time. In this case, based on the historical browsing record and the notified documents, the one or more local search documents retrieved in the search recall link are filtered, that is, the notified documents are filtered out from the one or more local search documents, thereby obtaining a first filtering result.

[0132] The process of retrieving one or more local search documents matching the subscription field from the local search document library is similar to the process of retrieving M candidate search documents matching the subscription field from the search document library, and will not be repeated here.

[0133] 5.3) Determine the subject or tag mapped to the subscription field from the recall recommendation document library provided by the recommendation recall link, and retrieve one or more candidate recommendation documents matching the subscription field in the recall recommendation document library based on the mapped subject or tag; filter the one or more candidate recommendation documents retrieved by the recommendation recall link based on historical browsing history to obtain a second filtering result. The filtering of the one or more candidate recommendation documents retrieved by the recommendation recall link here is similar to 5.2) and will not be elaborated on.

[0134] 5.4) Generate Content Update Notification: The first and second filtering results are merged to obtain an update recall result. If the update recall result is not an empty set, a content update notification for the subscription field is generated based on the update recall result. As one implementation, due to rapid updates, the previous update recall result may not have generated a content update notification. In this case, the previous update recall result can be merged with the current update recall result to obtain a content update notification. In other words, the content update notification includes the update recall results that were not updated in a timely manner as well as the current update recall result. In another implementation, a content update notification for the subscription field is directly generated based on the update recall result. The content update notification may include one or both of the number of updated documents and a recommendation reason. When the content update notification includes a recommendation reason, the documents in the update recall result may be clustered by content tags to obtain the document type in the update recall result. Based on the document type, a recommendation reason is generated. For example, if clustering the documents in the update recall result reveals that the document type in the update recall result is related to the technical principles of the model, the recommendation reason may be output as "document update related to the technical principles of the model."

[0135] 5.5) The server calls the push service program to output the content update notification of the subscription field. Schematically, the content update notification can be output in a prominent manner (eg, a colored origin).

[0136] By actively outputting content update notifications, the object does not need to actively check whether the configured subscription field has content updates in the information retrieval interface, which simplifies the information retrieval operation, improves the information retrieval efficiency, and can better enhance the object experience.

[0137] Next, the information processing method provided in the embodiments of the present application is described.

[0138] Please refer to Figure 3a, which is a flowchart of an information processing method provided by an exemplary embodiment of the present application. The information processing method can be executed by a computer device in the above-mentioned information processing system, and the computer device can be a terminal in the above-mentioned information processing system. The information processing method can include the following steps S301-S303:

[0139] S301: Display an information retrieval interface, where the information retrieval interface is used to perform information retrieval based on a search field.

[0140] The search field may include keywords, complete sentences, etc. For example, the search field may be Xiao Ming, music, food, etc., and the search field may be "How long is the food festival?" For example, please refer to FIG3b, which is a schematic diagram of outputting attention information related to subscription information provided by an embodiment of the present application. The information retrieval interface is shown in FIG3b. A search field input control 31 is provided in the information retrieval interface 301. The object can enter the search field in the search field input control 31 to implement information search based on the search field.

[0141] S302: Display a subscription field, where the subscription field refers to a configured search field that requires continuous attention.

[0142] As an implementation method, an object can configure a subscription field of interest on an information retrieval interface. Accordingly, a terminal can retrieve the configured subscription field and display it. Displaying the subscription field can include at least one of the following: ① Displaying the subscription field on the information retrieval interface; as shown in FIG3b , displaying a subscription field 32 on the information retrieval interface. The display location of the subscription field 32 can be flexibly set, such as displaying the subscription field 32 in a fixed position on the information retrieval interface, or displaying the subscription field 32 in a floating position on the information retrieval interface, etc., and this embodiment of the present application does not impose any restrictions. ② Displaying the subscription field on a first interface independent of the information retrieval interface. The first interface can be any interface other than the information retrieval interface, for example, the information retrieval interface displayed on the client. The first interface can be a client main interface, a service interface, a session message interface, etc., independent of the information retrieval interface. The display location of the subscription field can be flexibly set, such as displaying the subscription field in a fixed position on the first interface, or displaying the subscription field in a floating position on the first interface, etc., and this embodiment of the present application does not impose any restrictions on the display interface of the subscription field or the display location of the subscription field on the display interface.

[0143] In an embodiment of the present application, the number of subscription fields may be N, where N is a positive integer. Displaying subscription fields may include at least one of the following: ① Displaying N subscription fields in a tiled manner. Schematically, FIG4a shows a schematic diagram of displaying N subscription fields in accordance with an exemplary embodiment of the present application. In FIG4a, taking the display of subscription fields on an information retrieval interface as an example, subscription fields 1, 2, and 3 are tiled in area 33 of information retrieval interface 301. ② Arranging and displaying the N subscription fields in the order in which they were configured. Schematically, N = 3, where subscription field A was configured before subscription field B, and subscription field B was configured before subscription field C. Taking the display of subscription fields on an information retrieval interface as an example, subscription fields A, B, and C may be arranged and displayed in the information retrieval interface in the order in which these three subscription fields were configured. ③ Arranging and displaying the N subscription fields in descending order of interest. The higher the interest level of a subscription field, the more interested the subject is in that subscription field. Schematically, the attention level of subscription field A is higher than that of subscription field B, and the attention level of subscription field B is higher than that of subscription field C, as shown in FIG4b , which provides a schematic diagram of displaying N subscription fields for another exemplary embodiment of the present application. Taking the subscription field displayed on the information retrieval interface as an example, as shown in FIG4b , N=3, subscription field A, subscription field B, and subscription field C are arranged and displayed in the information retrieval interface in descending order of attention level. ④ Randomly display N subscription fields. ⑤ Classify and display the N subscription fields according to the topics to which the N subscription fields respectively belong. As an implementation method, taking the subscription field displayed on the information retrieval interface as an example, classifying and displaying the N subscription fields according to the topics to which the N subscription fields respectively belong may include: displaying subscription fields belonging to different topics in different areas of the information retrieval interface, and displaying subscription fields belonging to the same topic in the same area. Schematically, taking the subscription field displayed on the information retrieval interface as an example, as shown in FIG4c , a schematic diagram of displaying N subscription fields is provided for another exemplary embodiment of the present application. In Figure 4c, subscription field a and subscription field b belong to the same topic, and subscription field c and subscription field d belong to the same topic. In the information retrieval interface 302, subscription field a and subscription field b are displayed in area 34, and subscription field c and subscription field d are displayed in area 35. ⑥ Fold and display N subscription fields. Taking the display of subscription fields in the information retrieval interface as an example, as shown in Figure 4d, a schematic diagram of folding and displaying N subscription fields is provided as an exemplary embodiment of the present application. In Figure 4d, 2 (N=2) subscription fields are folded and displayed in the information retrieval interface, and the folded subscription fields can support being triggered to expand and display.The information retrieval interface in FIG4d includes a subscription field expansion option 36. In response to a trigger operation for the subscription field expansion option 36, two folded subscription fields (i.e., subscription field 1 and subscription field 2) are displayed. ⑦ Arrange and display the N subscription fields in descending order of their field lengths. For example, the field length of the subscription field "XX City" is 4, and the field length of the subscription field "XXXX Cuisine" is 6. Taking the subscription fields displayed on the information retrieval interface as an example, the subscription fields "XXXX Cuisine" and "XX City" are arranged and displayed in descending order of their field lengths on the information retrieval interface.

[0144] S303: Output the attention information related to the subscription field.

[0145] In one implementation, the number of subscription fields is N, and outputting the attention information related to the subscription fields includes: in response to a selection operation on the subscription field, outputting the attention information related to the subscription field selected by the selection operation. The attention information includes, but is not limited to: business recall results obtained by information retrieval based on the subscription field, content update notifications, and the like. The business recall results may include multiple documents matching the subscription field, and the multiple documents may include recommended documents and / or retrieved documents. For example, in FIG3b , documents 321, 322, 323, and 324 related to the subscription field are output. The content update notification can be used to prompt that there is an update to the document related to the subscription field. The content update notification may include one or both of the number of updated documents and the reason for the recommendation.

[0146] In an embodiment of the present application, outputting information of interest related to a subscription field may include at least one of the following: ① Outputting the information of interest in a fixed area within the information retrieval interface. The fixed area may be any area within the information retrieval interface, such as the top area, the middle area, the right area, etc., and is not limited in this embodiment of the present application. For example, the information of interest includes a service recall result, and the service recall result includes Document 1 and Document 2 matching the subscription field. FIG5a is a schematic diagram of outputting multiple documents matching the subscription field in a fixed area, provided in accordance with an exemplary embodiment of the present application. In FIG5a, Document 1 and Document 2 matching the subscription field are displayed in fixed area 51 of information retrieval interface 501. ② Outputting the information of interest in a floating area within the information retrieval interface. For example, the information of interest includes a content update notification. In FIG5b, a schematic diagram of outputting a content update notification in a floating area, provided in accordance with an exemplary embodiment of the present application, is shown. In FIG5b, a schematic diagram of outputting a content update notification in a floating area, provided in accordance with an exemplary embodiment of the present application, is shown. In FIG5b, a content update notification 1 related to the subscription field is displayed in floating area 52 of information retrieval interface 501. ③ Outputting the information of interest in a second interface independent of the information retrieval interface. For example, the information of interest includes a business recall result, and the business recall result includes documents 1 and 2 that match the subscription field. As shown in Figure 3b, documents 321, 322, 323, and 324 that match the subscription field are displayed in a second interface 302 that is independent of the information retrieval interface 301. The second interface can be any interface other than the information retrieval interface. For example, the information retrieval interface is displayed in the client. The second interface can be a client main interface, a service interface, a session message interface, etc. that is independent of the information retrieval interface. This embodiment of the application does not impose any restrictions on the second interface.

[0147] In an embodiment of the present application, an information retrieval interface is displayed, which is used to retrieve information based on a search field; a subscription field is displayed, which refers to a configured search field; and information of interest related to the subscription field is output. Through the above method, not only can information retrieval be performed based on the search field in the information retrieval interface, but the subject can also configure the search field as a subscription field based on their own continuous attention needs. By displaying the subscription field, the configured search field that needs to be continuously followed can be presented, and the search field that the subject is continuously interested in can be learned. Information can also be continuously retrieved through the subscription field to obtain and output information of continuous interest to the subject. While enriching the information retrieval method, the flexibility of information retrieval is improved, thereby meeting the subject's need for continuous attention to information.

[0148] Next, the specific implementation process of obtaining the subscription field provided in the embodiment of the present application is described.

[0149] In one embodiment, obtaining the subscription field may include but is not limited to the following methods:

[0150] ① A search subscription option is provided in the information retrieval interface; a subject can trigger the search subscription option to perform a configuration operation and enter a subscription field in the search subscription input window. In this implementation, obtaining the subscription field includes: in response to the search subscription option being triggered, displaying the search subscription input window, then receiving a first search field entered in the search subscription input window, and configuring the first search field as a subscription field. The search subscription option can be triggered by a click, a double-click, a voice trigger, a swipe, or the like. For illustration, see Figure 6a, which illustrates a schematic diagram of configuring a subscription field in accordance with an exemplary embodiment of the present application. In Figure 6a, a search subscription option 61 is provided in the information retrieval interface. A subject clicks on the search subscription option 61. In response to the triggering of the search subscription option 61, a search subscription input window 601 is displayed. The subject can enter the first search field "piano music" in the search subscription input window 601. Accordingly, the first search field "piano music" entered in the search subscription input window 601 is received and configured as a subscription field.

[0151] ② A search subscription option is provided in the information retrieval interface; obtaining a subscription field includes: displaying one or more candidate search fields in response to the search subscription option being triggered; receiving a selection operation for the candidate search field, and configuring the selected candidate search field as a subscription field. The one or more candidate search fields may be displayed directly in the information retrieval interface, or the one or more candidate search fields may be displayed in a search subscription input window, or the one or more candidate search fields may be displayed in a search subscription interface independent of the information retrieval interface, and this embodiment of the application does not impose any limitation on this.

[0152] Schematically, please refer to FIG6b, which is a schematic diagram of setting a subscription field provided in another exemplary embodiment of the present application. In FIG6b, a search subscription option 61 is provided in the information retrieval interface. The subject clicks on the search subscription option 61. In response to the triggering of the search subscription option 61, two candidate search fields are displayed: candidate search field 1 and candidate search field 2. The subject can select one or two candidate search fields from these two candidate search fields to configure as subscription fields. Assume that the subject selects candidate search field 1, a selection operation for candidate search field 1 is received, and the selected candidate search field 1 is configured as a subscription field.

[0153] Among them, the one or more candidate search field acquisition formula includes: based on the object identifier of the object, obtaining one or more candidate search fields corresponding to the object identifier from the database, the candidate search field is the behavior log of the collection object within a certain time period (such as day level, hour level, etc.), and using content tag clustering technology to calculate the candidate search subscription field of the behavior log.

[0154] It should be understood that when an object triggers the search subscription option, one or more candidate search subscription fields may be displayed in the search subscription input window. In this case, the object may customize the input of the first search field to be configured as the subscription field according to interest requirements in the search subscription input window, or may directly select the candidate search field to be configured as the subscription field from one or more candidate search fields.

[0155] ③ The information retrieval interface includes a search field input control; in this implementation, obtaining the subscription field includes: receiving a second search field input in the search field input control, and in response to a configuration operation for the second search field, configuring the second search field as a subscription field.

[0156] Among them, the configuration operation for the second search field may include any of the following: A. A subscription setting option is displayed in the information retrieval interface, and the configuration operation for the second search field may include a triggering operation for the subscription setting option. Among them, the subscription setting option can be displayed in the information retrieval interface after receiving the second search field entered in the search field input control. Alternatively, the subscription setting option is always displayed in the information retrieval interface to configure the second search field entered in the search field input control as a subscription field. Schematically, please refer to Figure 6c, which is a schematic diagram of setting a subscription field provided in another exemplary embodiment of the present application. In Figure 6c, the information retrieval interface includes a search field input control 62. The object can enter the second search field "artificial intelligence model" in the search field input control 62. Accordingly, the second search field "artificial intelligence model" entered in the search field input control is received, and the subscription setting option 63 is displayed in the information retrieval interface. In response to the triggering operation of displaying the subscription setting option 63, the second search field "artificial intelligence model" is configured as a subscription field. B. Click, double-click, and other operations in the preset area of ​​the information retrieval interface. The preset area may be the middle area, the left area, the right area, and the like in the information retrieval interface. The embodiments of the present application do not impose any restrictions on this. C. Double-click, move, click, and other operations on the second search field. Schematically, please refer to Figure 6d, which is a schematic diagram of setting a subscription field provided by another exemplary embodiment of the present application. In Figure 6d, the information retrieval interface includes a search field input control 62. The object can enter the second search field "artificial intelligence model" in the search field input control 62. Accordingly, the object receives the second search field "artificial intelligence model" entered in the search field input control. When the object clicks on the second search field "artificial intelligence model", the second search field "artificial intelligence model" is configured as a subscription field in response to the click operation on the second search field. D. Input operation of specific gestures in the information retrieval interface. Specific gestures may include but are not limited to: M gesture, OK gesture, and the like.

[0157] ④ After the object retrieves information based on the search field, it will record the historical search field. Therefore, in an embodiment of the present application, obtaining the subscription field may include: displaying information of at least one historical search field, which refers to the search field used in the historical information retrieval process; receiving a selection operation for the historical search field, and configuring the selected historical search field as a subscription field. Schematically, please refer to Figure 6e, which is a schematic diagram of setting a subscription field provided by an exemplary embodiment of the present application. In Figure 6e, two historical search fields are displayed in the information retrieval interface, namely the historical search field "model" and the historical search field "calligraphy and painting". The object can select one or both of the historical search field "model" and the historical search field "calligraphy and painting" as subscription fields. Assuming that the object selects the historical search field "model", accordingly, it receives a selection operation for the historical search field "model" and uses the historical search field "model" as a subscription field.

[0158] In an embodiment of the present application, a subscription field can support editing, where editing can include modification or deletion. When editing includes modification, in one implementation, a modification option for the subscription field is displayed, which can be triggered by a single click or double click. In response to the triggering operation on the modification option, the subscription field can be set to an editable state. When a modification operation is received on the subscription field, the subscription field is modified.

[0159] When editing includes deletion, in one implementation, a delete option for the subscription field may be displayed, and the object may trigger the delete option by single-clicking, double-clicking, etc., and in response to the triggering operation for the delete option, the subscription field is deleted. Schematically, as shown in FIG7a, a schematic diagram of deleting a subscription field word is provided for an exemplary embodiment of the present application. In FIG7a, subscription field 1 and subscription field 2 are displayed in the information retrieval interface, and a delete option 71 corresponding to subscription field 1 and a delete option 72 corresponding to subscription field 2 are also displayed. When the object clicks on the delete option 72 corresponding to subscription field 2, in response to the click operation on the delete option 72 corresponding to subscription field 2, the subscription field 2 is deleted in the information retrieval interface. In another implementation, the object may move the subscription field to be deleted to a fixed area, and in response to the move operation on the subscription field, the moved subscription field may be deleted from the information retrieval interface.

[0160] In some feasible embodiments, when a subscription field changes due to editing, the changed subscription field needs to be refreshed and displayed. Schematically, in Figure 7a, after deleting subscription field 2, the changed subscription field is refreshed and displayed in the information retrieval interface, that is, only field 1 is displayed in the information retrieval interface. The number of changed subscription fields is M, M is less than or equal to N, and refreshing the display of the changed subscription field may include at least one of the following: ① displaying the M changed subscription fields in a tiled manner. ② displaying the M changed subscription fields in the order of the time when the M changed subscription fields were configured; ③ displaying the M changed subscription fields in the order of attention from high to low; ④ displaying the M changed subscription fields randomly; ⑤ displaying the M changed subscription fields in the order of the field length of the M changed subscription fields from large to small.

[0161] The above description uses the display of the subscription field in the information retrieval interface as an example. It should be understood that when the subscription field changes due to editing, the changed subscription field can also be refreshed and displayed in a second interface independent of the information retrieval interface.

[0162] As described above, the embodiments of the present application provide a flexible way to configure subscription fields. For example, when configuring subscription fields, users can customize the subscription fields and select candidate search fields to be configured as subscription fields. For another example, when configuring subscription fields, users can configure the first search field entered in the search subscription input window as the subscription field. This flexible configuration of subscription fields can promote content consumption in scenarios where users have ongoing interest.

[0163] Next, the output and subscription field-related information involved in the embodiments of the present application will be described.

[0164] As previously mentioned, the attention information includes service recall results, content update notifications, and the like obtained by searching for information based on the subscription field. In one possible embodiment, the attention information includes service recall results obtained by searching for information based on the subscription field. The service recall results may include multiple documents matching the subscription field, including articles, web pages, videos, images, or audio. In this embodiment, the attention information output related to the subscription field includes at least one of the following:

[0165] (1) Display multiple documents in random order.

[0166] (2) Arrange and display multiple documents in descending order of the degree of match between each document and the subscription field. As an implementation method, the degree of match between each document and the subscription field can be determined, and then arrange and display multiple documents in descending order of the degree of match between each document and the subscription field. Schematically, as shown in FIG7b , a schematic diagram of displaying multiple documents matching a subscription is provided for an exemplary embodiment of the present application. In FIG7b , the business recall result includes 3 documents, the degree of match between document 1 and the subscription field is 98%, the degree of match between document 2 and the subscription field is 88%, and the degree of match between document 3 and the subscription field is 80%. Document 1, document 2, and document 3 are arranged and displayed in descending order of the degree of match between each document and the subscription field.

[0167] (3) Arrange and display multiple documents in descending order of attention. Attention can be understood as the degree of interest of the object. The higher the interest of the object, the higher the attention. Schematically, the subscription field is "ChatGPT model" (Generative Pre-trained Transformer model), and the multiple documents matching the subscription field include: Document 1 "Technical Implementation of ChatGPT Model" and Document 2 "Overview of AI Model". The object is interested in the ChatGPT model, and the attention of Document 1 "Technical Implementation of ChatGPT Model" is higher than that of Document 2 "Overview of AI Model". Document 1 "Technical Implementation of ChatGPT Model" and Document 2 "Overview of AI Model" are arranged and displayed in descending order of attention.

[0168] The document's attention level can be determined based on the subject's real-time interest needs. The real-time interest needs can be obtained by clustering the content labels of the subject's behavior logs and subject information within the time period closest to the system time (such as the last three days or the last week). For example, if the real-time interest needs indicate that the subject is interested in the ChatGPT model, a higher attention level can be set for documents related to the ChatGPT model.

[0169] (4) Arrange and display multiple documents according to the requirements of document diversity. The document diversity may include but is not limited to: the diversity of document carriers, the diversity of document sources, etc. The document carriers may include videos, articles, web pages, audio, etc. The document sources may include: the account to which the document belongs, the document publishing platform, etc. In an illustrative manner, the requirements for document diversity include: the requirements for document carrier diversity, that is, the document carriers are required to include articles and videos arranged alternately, as shown in FIG7c, which is a schematic diagram of an exemplary embodiment of the present application for arranging and displaying multiple documents. In FIG7c, the multiple documents include document A, document B and document C. The document carrier of document A is a video, the document carrier of document B is a video, and the document carrier of document C is an article. According to the requirements for document carrier diversity, document A, document C and document B are arranged and displayed. In an illustrative manner, the requirements for document diversity include: the requirements for the diversity of the account to which the document belongs, that is, the requirement that no more than two documents belonging to the same account be displayed at a time. The multiple documents include document A, document B, document C and document D. Document A, document B and document C all belong to account A, and document D belongs to account B. According to the requirements for the diversity of the account to which the document belongs, document A, document B, document D and document C are arranged and displayed.

[0170] (5) Clustering and displaying multiple documents according to the topics or tags to which the documents belong. As an implementation method, clustering and displaying multiple documents according to the topics to which the documents belong includes: classifying and displaying documents belonging to different topics in different areas, and clustering and displaying documents belonging to the same topic in the same area. Schematically, document 1 and document 2 belong to topic 1, and document 3 and document 4 belong to topic 2. As shown in Figure 7d, a schematic diagram of clustering and displaying multiple documents is provided as an exemplary embodiment of the present application. In Figure 7d, according to the topics to which the documents belong, document 1 and document 2 are clustered and displayed in area 73, and document 3 and document 4 are clustered and displayed in area 74. As another implementation method, clustering and displaying multiple documents according to the tags to which the documents belong includes: classifying and displaying documents belonging to different tags in different areas, and clustering and displaying documents belonging to the same tag in the same area. Schematically, document 1 and document 2 belong to the tag "beautiful scenery", and document 3 and document 4 belong to the tag "tree planting". Document 1 and document 2 are clustered and displayed in the first area, and document 3 and document 4 are clustered and displayed in the second area.

[0171] In some feasible embodiments, the multiple documents may include viewed documents and unviewed documents, and outputting the attention information related to the subscription field includes at least one of the following:

[0172] ① Filter the browsed documents and display only the unbrowsed documents. As an implementation method, after obtaining the business recall result, the business recall result can be filtered for browsed documents based on the historical browsing records to obtain unbrowsed documents, and only the unbrowsed documents can be displayed in the interface. Specifically, the historical browsing records contain the fingerprint features of the browsed documents. According to the historical browsing records, the business recall result is filtered for browsed documents to obtain the unbrowsed documents, which includes: determining the fingerprint features of the documents in the business recall result, and comparing the fingerprint features of the documents in the business recall result with the fingerprint features of the browsed documents in the historical browsing records, and filtering the documents with the same fingerprint features as the browsed documents to obtain the unbrowsed documents.

[0173] It should be noted that in the embodiment of the present application, the multiple documents may only include unviewed documents. That is, when performing information retrieval based on the subscription field, the retrieved documents will be filtered for viewed documents, so the service recall results will no longer include viewed documents. In this case, the unviewed documents will be displayed directly.

[0174] ② Displaying browsed documents and unbrowsed documents in different areas of the same interface. Schematically, Figure 7e shows a schematic diagram of displaying browsed documents and unbrowsed documents in different areas of an information retrieval interface, provided by an exemplary embodiment of the present application. In Figure 7e, browsed documents 711 and 712 are displayed in area 75 of the information retrieval interface, and unbrowsed documents 713 and 714 are displayed in area 76 of the information retrieval interface.

[0175] ③ Display browsed documents and unbrowsed documents in different interfaces.

[0176] ④ Display unbrowsed documents, and fold the browsed documents in the display interface of unbrowsed documents. The folded browsed documents support being triggered to expand and display. In one implementation, the folded browsed documents correspond to a display option. When the display option is triggered, the browsed documents can be expanded and displayed in the display interface or a new interface independent of the display interface. Schematically, as shown in Figure 7f, a schematic diagram of displaying browsed documents and unbrowsed documents is provided for an exemplary embodiment of the present application. In Figure 7f, the unbrowsed documents include document 715, and document 715 is displayed in the display interface 702, and the browsed documents are folded in the display interface 702. The folded browsed documents correspond to a display option 77. When the display option 77 is triggered, the browsed documents are expanded and displayed in a new interface 703 independent of the display interface.

[0177] ⑤ Display multiple documents and highlight unbrowsed documents among the multiple documents; highlighting unbrowsed documents among the multiple documents may include at least one of the following: bolding the unbrowsed documents, tilting the unbrowsed documents, etc.

[0178] In other feasible embodiments, multiple documents may include retrieved documents and recommended documents, and the output of attention information related to the subscription field includes at least one of the following AEs: A. Hide the retrieved documents and only display the recommended documents. B. Hide the recommended documents and only display the retrieved documents. It should be understood that the hidden documents support being triggered to expand and display. C. Categorize and display the retrieved documents and recommended documents. Schematically, the retrieved documents can be displayed in the first display area of ​​the same interface, and the recommended documents can be displayed in the second area. D. Fusion display of the retrieved documents and recommended documents. As an implementation method, the fusion display of the retrieved documents and recommended documents may include: deduplication processing of the retrieved documents and recommended documents to obtain deduplication retrieved documents and recommended documents, and display the deduplication retrieved documents and recommended documents. For example, the retrieved documents include document 1 and document 2, and the recommended documents include document 1 and document 3. The fusion display of the retrieved documents and recommended documents may be deduplication processing of document 1, document 2, document 1, and document 3 to obtain document 1, document 2, and document 3, and display document 1, document 2, and document 3.

[0179] As another implementation method, the retrieved documents and the recommended documents are fused and displayed according to the fusion display rules. The fusion display rules include one or more of the following: attention fusion display rules, subscription field matching fusion rules, and document diversity fusion display rules. When the fusion display rules include attention fusion display rules, according to the fusion display rules, the fusion display of the retrieved documents and the recommended documents includes: fusion displaying the retrieved documents and the recommended documents in order of attention from high to low. For example, if the attention of the retrieved document is greater than that of the recommended document, the retrieved document and the recommended document are fused and displayed in order of attention from high to low, that is, the retrieved document and the recommended document are arranged and displayed in order of attention from high to low.

[0180] When the fusion display rule includes a subscription field matching fusion rule, the subscription field matching fusion rule may include the matching degree of the subject corresponding to the subscription field or the matching degree of the tag corresponding to the subscription field. If the subscription field matching fusion rule includes the matching degree of the subject corresponding to the subscription field, and according to the fusion display rule, fusion display of the retrieved document and the recommended document includes: determining the matching degree of the retrieved document and the subject corresponding to the subscription field, and determining the matching degree of the recommended document and the subject corresponding to the subscription field, and fusion displaying the retrieved document and the recommended document in descending order of the matching degree of the subject corresponding to the subscription field.

[0181] When the fusion display rule includes a document diversity fusion display rule, such as document diversity including the distribution of the number of documents of the same account, the searched documents and the recommended documents are fused according to the distribution of the number of documents of the same account to obtain the searched documents and recommended documents after fusion processing, and the searched documents and recommended documents after fusion processing are displayed. Schematically, the number of searched documents is 4, of which 3 searched documents belong to the same account, the number of recommended documents is 2, and these two recommended documents belong to the same account. Assume that the distribution of the number of documents of the same account is no more than 2. At this time, according to the distribution of the number of documents of the same account, the 4 searched documents and the 2 recommended documents can be fused. The fusion processing here can include: deleting any of the 3 searched documents belonging to the same account, and finally obtaining 3 searched documents and 2 recommended documents, and displaying the 3 searched documents and 2 recommended documents.

[0182] E. Filter the retrieved documents and recommended documents in descending order of attention, and display the filtered documents. As an implementation method, filter the retrieved documents and recommended documents in descending order of attention to obtain Y documents, and display the Y documents. For example, Y = 2, and the multiple documents include retrieved document 1, retrieved document 2, and recommended document 1. Retrieved document 1 has a higher level of attention than retrieved document 2, and retrieved document 2 has a higher level of attention than recommended document 1. Filtering the retrieved documents and recommended documents in descending order of attention yields two documents: retrieved document 1 and retrieved document 2.

[0183] In some feasible embodiments, there may be updates to documents related to the subscription field. In this case, the attention information includes a content update notification, which is used to prompt that there are updates to documents related to the subscription field. In this embodiment, outputting the attention information related to the subscription field may include at least one of the following: (1) outputting a content update notification at the display position of the subscription field. Schematically, as shown in FIG8a, a content update notification schematic diagram is provided for an exemplary embodiment of the present application. The content update notification includes the number of updated documents related to the subscription field. In FIG8a, at the display position 81 of the subscription field "piano music", the number of updated documents related to the subscription field "piano music" is 20, and at the display position 82 of the subscription field "food", the number of updated documents related to the subscription field "piano music" is 10. (2) outputting a content update notification around the display position of the subscription field. The surrounding of the display position can be understood as: any position whose distance from the display position of the subscription field is less than or equal to a preset distance (such as 1 mm, 0.1 mm, etc.). Schematically, as shown in FIG8b, a content update notification schematic diagram is provided for another exemplary embodiment of the present application. The content update notification includes a recommendation reason. In FIG8b , the recommendation reason "There is a document update about the model technology" is displayed around the display position 83 of the subscription field. (3) A third interface independent of the information retrieval interface is used to output the content update notification. The third interface can be any interface in the terminal except the information retrieval interface. For example, the third interface can be the main interface of the terminal, the standby interface, the message prompt interface, the information interface where the client is located in the terminal, the service interface, etc., which is not limited in the embodiment of the present application.

[0184] The output form of the content update notification may include at least one of the following: 1. Outputting the content update notification in text form. For example, in Figures 8a and 8b, the content update notification is output in text form. 2. Outputting the content update notification in a highlighted manner. The highlighted manner herein may be, for example, a colored graphic manner (such as a colored dot manner), a bold manner, a tilted manner, an enlarged manner, etc. As shown in Figure 8a, the content update notification is output in bold. 3. Outputting the content update notification in voice form. When a content update notification exists, the content update notification is directly played by voice.

[0185] As described above, by actively outputting content update notifications, the object does not need to actively conduct an information retrieval interface to check whether there is content update in the subscribed subscription field, which simplifies the information retrieval operation, improves the information retrieval efficiency, and can better enhance the object experience.

[0186] Next, the information processing device provided in the embodiments of the present application will be described.

[0187] Please refer to Figure 9, which is a schematic diagram of the structure of an information processing device provided in an embodiment of the present application. The information processing device can be a computer program (including program code) in a computer device, for example, the information processing device can be an application software in the computer device; the information processing device can be used to perform some or all of the steps in the method embodiment shown in Figure 3a. Referring to Figure 9, the information processing device includes the following units:

[0188] Display unit 901, used to display an information retrieval interface, which is used to perform information retrieval based on a search field;

[0189] The display unit 901 is further used to display a subscription field, which refers to a configured search field;

[0190] The processing unit 902 is configured to output attention information related to the subscription field.

[0191] The processing unit 902 is further configured to:

[0192] Get the subscription field. The subscription field supports editing, which includes modification or deletion.

[0193] When the subscription fields are changed due to editing, the changed subscription fields are refreshed and displayed.

[0194] The information retrieval interface includes a search and subscription option. The processing unit 902 is specifically configured to:

[0195] In response to the search subscription option being triggered, displaying a search subscription input window;

[0196] Receiving a first search field input in the search subscription input window;

[0197] Configure the first search field as a subscription field.

[0198] The information retrieval interface includes a search field input control; the processing unit 902 is specifically configured to:

[0199] a second search field receiving input in a search field input control;

[0200] In response to the configuration operation on the second search field, the second search field is configured as a subscription field.

[0201] The processing unit 902 is specifically configured to:

[0202] Display at least one historical search field, where the historical search field refers to a search field used in a historical information retrieval process;

[0203] Receive a selection operation on a historical search field and configure the selected historical search field as a subscription field.

[0204] The information retrieval interface includes a search and subscription option. The processing unit 902 is specifically configured to:

[0205] In response to the search subscription option being triggered, displaying one or more candidate search fields;

[0206] A selection operation for a candidate search field is received, and the selected candidate search field is configured as a subscription field.

[0207] The number of subscription fields is N; the display unit 901 is specifically used to:

[0208] Display subscription fields in the information retrieval interface;

[0209] Displaying a subscription field in a first interface independent of the information retrieval interface;

[0210] Display N subscription fields in a tiled manner;

[0211] Arrange and display N subscription fields in the order of their configuration time.

[0212] Arrange and display N subscription fields in descending order of attention;

[0213] Randomly display N subscription fields;

[0214] Display N subscription fields according to their respective topics;

[0215] Fold and display N subscription fields;

[0216] The N subscription fields are displayed in descending order of their field lengths.

[0217] The processing unit 902 is specifically configured to:

[0218] Outputting the information of interest in a fixed area of ​​the information retrieval interface;

[0219] Output the focused information in the floating area of ​​the information retrieval interface;

[0220] The information of interest is output in a second interface independent of the information retrieval interface.

[0221] The information of interest includes a business recall result obtained by performing information retrieval based on the subscription field. The business recall result includes multiple documents matching the subscription field. The documents include articles, web pages, videos, images, or audio. The processing unit 902 is specifically configured to:

[0222] Display multiple documents in random order;

[0223] Arrange and display multiple documents in descending order of matching degree with the subscription field;

[0224] Arrange and display multiple documents in descending order of attention;

[0225] Arrange and display multiple documents according to the needs of document diversity;

[0226] Clustering displays multiple documents according to the topics or tags to which the documents belong.

[0227] The attention information includes a business recall result obtained by performing information retrieval based on the subscription field, and the business recall result includes multiple documents matching the subscription field, and the multiple documents include browsed documents and unbrowsed documents; the processing unit 902 is specifically configured to:

[0228] Filter the browsed documents to display only the unbrowsed documents;

[0229] Display browsed documents and unbrowsed documents in different areas of the same interface;

[0230] Display browsed documents and unbrowsed documents in different interfaces;

[0231] Display unbrowsed documents and collapse browsed documents in the display interface of unbrowsed documents. The collapsed browsed documents can be triggered to expand and display;

[0232] Display multiple documents and highlight unbrowsed documents among the multiple documents.

[0233] The attention information includes a business recall result obtained by performing information retrieval based on the subscription field. The business recall result includes multiple documents matching the subscription field, and the multiple documents include retrieved documents and recommended documents. The processing unit 902 is specifically configured to:

[0234] Hide searched documents and only show recommended documents;

[0235] Hide recommended documents and only display searched documents;

[0236] Categorize and display retrieved documents and recommended documents;

[0237] Fusion display of retrieved documents and recommended documents;

[0238] The retrieved documents and recommended documents are filtered in descending order of attention, and the filtered documents are displayed.

[0239] The attention information includes a content update notification of the subscription field; the processing unit 902 is specifically configured to:

[0240] Output content update notification at the display position of the subscription field;

[0241] Output content update notification around the display position of the subscription field;

[0242] Use a third interface independent of the information retrieval interface to output content update notifications;

[0243] The content update notification includes one or both of the number of updated documents and the reason for recommendation; the output form of the content update notification includes at least one of the following: outputting the content update notification in text form; outputting the content update notification in a prominent manner; and outputting the content update notification in voice form.

[0244] The attention information includes a service recall result obtained by performing information retrieval based on the subscription field, and the service recall result includes a search recall result. The processing unit 902 is further configured to:

[0245] By searching the recall link, M candidate search documents matching the subscription field within a first preset time period are retrieved from the search document library; M is a positive integer;

[0246] Based on the historical browsing records of the object, M candidate search documents are filtered to obtain L unbrowsed candidate search documents; L is less than or equal to M;

[0247] Sorting the L unbrowsed candidate retrieval documents to obtain the sorted L candidate retrieval documents;

[0248] Based on the personalized data of the object, personalized retrieval document recall processing is performed to obtain personalized retrieval documents;

[0249] Perform fusion sorting on the sorted L candidate retrieval documents and personalized retrieval documents to obtain the search recall result;

[0250] Based on the search recall results, the recall results obtained by information retrieval based on the subscription field are obtained.

[0251] The service recall result also includes a recommendation recall result. The processing unit 902 is further configured to:

[0252] Based on the subscription field, a topic or tag mapped to the subscription field is determined from a recall recommendation document library provided by the recommendation recall link; the recall recommendation information library includes at least one of the following: candidate recommendation documents corresponding to different topics, and candidate recommendation documents corresponding to different tags; the candidate recommendation documents in the recall recommendation document library are sorted in descending order of candidate recommendation document weights within a target time period and constructed based on the first Y candidate recommendation documents in the sorted candidate recommendation documents;

[0253] Retrieve candidate recommendation documents that match the subscription fields based on the mapped topics or tags in the recall recommendation document library;

[0254] Sorting the retrieved matching candidate recommendation documents according to the sorting parameters to obtain sorted candidate recommendation documents;

[0255] Based on historical browsing records, the sorted candidate recommendation documents are filtered to obtain unbrowsed candidate recommendation documents;

[0256] According to the object information and personalized data, the candidate recommendation documents that have not been browsed are re-ranked to obtain the recommendation recall results;

[0257] The search recall results and recommendation recall results are integrated to obtain the business recall results.

[0258] The attention information includes a content update notification of a subscription field. The processing unit 902 is further configured to:

[0259] Obtaining the subscription field from the scheduling queue, and retrieving one or more local search documents matching the subscription field from the local search document library by searching the recall link; the publication time of the local search documents in the local search document library is all within the second preset time period;

[0260] Determine the subject or tag mapped to the subscription field from the recall recommendation document library provided by the recommendation recall link, and retrieve one or more candidate recommendation documents matching the subscription field in the recall recommendation document library based on the mapped subject or tag;

[0261] Based on the historical browsing records, one or more local search documents retrieved in the search recall link are filtered to obtain a first filtering result;

[0262] Based on the historical browsing records, one or more candidate recommendation documents retrieved by the recommendation recall link are filtered to obtain a second filtering result;

[0263] Fusing the first filtering result and the second filtering result to obtain an updated recall result;

[0264] If the update recall result is not an empty set, a content update notification of the subscription field is generated based on the update recall result.

[0265] In the embodiment of the present disclosure, an information retrieval interface is displayed, which is used to retrieve information based on a search field; a subscription field is displayed, which refers to a configured search field; and attention information related to the subscription field is output. Through the above method, not only can information retrieval be performed based on the search field in the information retrieval interface, but the object can also configure the search field as a subscription field according to its own continuous attention needs. By displaying the subscription field, the configured search field that needs to be continuously followed can be presented, and the search field that the object is continuously interested in can be learned; information can also be continuously retrieved through the subscription field to obtain and output information that the object is continuously interested in. While enriching the information retrieval method, the flexibility of information retrieval is improved, so that it can better adapt to the scenario where the object continuously pays attention to information.

[0266] Next, the computer device provided in the embodiments of the present application is described.

[0267] Furthermore, an embodiment of the present application also provides a structural schematic diagram of a computer device, and the structural schematic diagram of the computer device can be seen in Figure 10; the computer device can be the above-mentioned server, and the computer device may include: a processor 1001, an input device 1002, an output device 1003 and a memory 1004. The above-mentioned processor 1001, input device 1002, output device 1003 and memory 1004 are connected via a bus. The memory 1004 is used to store a computer program, and the computer program includes program instructions. The processor 1001 is used to execute the program instructions stored in the memory 1004. Among them, the processor 1001 executes the description of the information processing method in the embodiment corresponding to Figure 3a above by running the program instructions in the memory 1004, and can also execute the description of the information processing device in the embodiment corresponding to Figure 9 above, which will not be repeated here. In addition, the description of the beneficial effects of adopting the same method will not be repeated.

[0268] In addition, it should be noted here that: the embodiment of the present application also provides a computer-readable storage medium, and a computer program is stored in the computer-readable storage medium, and the computer program includes program instructions. When the processor executes the above program instructions, it can execute the method in the embodiment corresponding to Figure 3a above, so it will not be described in detail here. For technical details not disclosed in the computer-readable storage medium embodiment involved in this application, please refer to the description of the method embodiment of this application. As an example, the program instructions can be deployed on a computer device, or executed on multiple computer devices located at one location, or, executed on multiple computer devices distributed at multiple locations and interconnected by a communication network.

[0269] According to one aspect of the present application, a computer program product is provided, comprising a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium and executes the computer program, thereby enabling the computer device to perform the method described in the embodiment corresponding to FIG. 3 a above. Therefore, detailed description thereof will not be given here.

[0270] The above disclosure is only a preferred embodiment of the present application, and certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. An information processing method, characterized in that: include: Displaying an information retrieval interface, wherein the information retrieval interface is used to perform information retrieval based on a search field; Displaying a subscription field, wherein the subscription field refers to a configured search field; Output the attention information related to the subscription field.

2. The method according to claim 1, characterized in that The method further comprises: Get a subscription field, where the subscription field supports editing, including modification or deletion; When the subscription field is changed due to editing, the changed subscription field is refreshed and displayed.

3. The method according to claim 1 or 2, characterized in that The information retrieval interface is provided with a search subscription option; the acquisition subscription field includes: In response to the search subscription option being triggered, displaying a search subscription input window; Receiving a first search field input in the search subscription input window; The first search field is configured as a subscription field.

4. The method according to any one of claims 1 to 3, characterized in that: The information retrieval interface includes a search field input control; the acquisition subscription field includes: receiving a second search field input in the search field input control; In response to a configuration operation on the second search field, the second search field is configured as a subscription field.

5. The method according to any one of claims 1 to 4, characterized in that: The acquisition of subscription fields includes: Display at least one historical search field, wherein the historical search field refers to a search field used in a historical information retrieval process; A selection operation for the history search field is received, and the selected history search field is configured as a subscription field.

6. The method according to any one of claims 1 to 5, characterized in that: The information retrieval interface is provided with a search subscription option; the acquisition subscription field includes: In response to the search subscription option being triggered, displaying one or more candidate search fields; A selection operation for a candidate search field is received, and the selected candidate search field is configured as a subscription field.

7. The method according to any one of claims 1 to 6, characterized in that: The number of the subscription fields is N; the displaying of the subscription fields in the information retrieval interface includes at least one of the following: Displaying a subscription field in the information retrieval interface; Displaying a subscription field in a first interface independent of the information retrieval interface; Display N subscription fields in a tiled manner; Arrange and display the N subscription fields in the order of the time when the N subscription fields were configured; Arrange and display the N subscription fields in descending order of attention; Randomly display the N subscription fields; Classify and display the N subscription fields according to the topics to which the N subscription fields respectively belong; Fold and display the N subscription fields; The N subscription fields are arranged and displayed in descending order of their field lengths.

8. The method according to any one of claims 1 to 7, characterized in that: The outputting of the attention information related to the subscription field includes at least one of the following: Outputting the information of interest in a fixed area in the information retrieval interface; Outputting the information of interest in a floating area in the information retrieval interface; The information of interest is output in a second interface independent of the information retrieval interface.

9. The method according to any one of claims 1 to 8, characterized in that The attention information includes a service recall result obtained by performing information retrieval based on the subscription field, and the service recall result includes a plurality of documents matching the subscription field, and the documents include articles, web pages, videos, images or audios; and the outputting of the attention information related to the subscription field includes at least one of the following: Arrange and display the plurality of documents in a random order; Arrange and display the multiple documents in descending order of the degree of matching with the subscription field; Arrange and display the multiple documents in descending order of attention; Arrange and display the multiple documents according to the needs of document diversity; The plurality of documents are displayed in clusters according to the topics or tags to which the documents belong.

10. The method according to any one of claims 1 to 9, characterized in that: The attention information includes a service recall result obtained by performing information retrieval based on the subscription field, and the service recall result includes a plurality of documents matching the subscription field, wherein the plurality of documents include browsed documents and unbrowsed documents; and the outputting of the attention information related to the subscription field includes at least one of the following: Filtering the browsed documents to display only the unbrowsed documents; Displaying the browsed documents and the unbrowsed documents in different areas of the same interface respectively; Displaying the browsed documents and the unbrowsed documents in different interfaces respectively; Display the unbrowsed document, and fold the browsed document in the display interface of the unbrowsed document, so that the folded browsed document supports being triggered to be expanded and displayed; The multiple documents are displayed, and unbrowsed documents among the multiple documents are highlighted.

11. The method according to any one of claims 1 to 10, characterized in that: The attention information includes a service recall result obtained by performing information retrieval based on the subscription field, and the service recall result includes multiple documents matching the subscription field, and the multiple documents include a search document and a recommended document; and the outputting of the attention information related to the subscription field includes at least one of the following: Hide the searched documents and only display the recommended documents; Hide the recommended documents and only display the searched documents; Displaying the retrieved documents and the recommended documents by categories; Fusion display of the retrieved document and the recommended document; The retrieved documents and the recommended documents are screened in descending order of attention, and the screened documents are displayed.

12. The method according to any one of claims 1 to 11, characterized in that: The attention information includes a content update notification of the subscription field; and the outputting of the attention information related to the subscription field includes at least one of the following: Outputting the content update notification at the display position of the subscription field; outputting the content update notification around the display position of the subscription field; Outputting the content update notification using a third interface independent of the information retrieval interface; Among them, the content update notification includes one or two of the number of updated documents and recommendation reasons; the output form of the content update notification includes at least one of the following: outputting the content update notification in text form; outputting the content update notification in a prominent manner; outputting the content update notification in voice form.

13. The method according to any one of claims 1 to 12, characterized in that: The concerned information includes a service recall result obtained by performing information retrieval based on the subscription field, and the service recall result includes a search recall result; and the method further includes: By searching the recall link, M candidate search documents matching the subscription field within a first preset time period are retrieved from the search document library; M is a positive integer; Based on the historical browsing records of the object, the M candidate search documents are filtered to obtain L unbrowsed candidate search documents; L is less than or equal to M; Sorting the L unbrowsed candidate retrieval documents to obtain sorted L candidate retrieval documents; Perform personalized retrieval document recall processing based on the personalized data of the object to obtain personalized retrieval documents; The sorted L candidate retrieval documents and the personalized retrieval document are subjected to fusion sorting processing to obtain the search recall result.

14. The method according to any one of claims 1 to 13, characterized in that: The service recall result also includes a recommended recall result; and the method further includes: According to the subscription field, a subject or label mapped to the subscription field is determined from a recall recommendation document library provided by a recommendation recall link; the recall recommendation information library includes at least one of the following: candidate recommendation documents corresponding to different subjects, and candidate recommendation documents corresponding to different labels; the candidate recommendation documents in the recall recommendation document library are constructed based on the first Y candidate recommendation documents in the candidate recommendation documents after ranking the candidate recommendation documents within the target time period in descending order of the weight of the candidate recommendation documents; Retrieving candidate recommended documents matching the subscription field in the recalled recommended document library based on the mapped topics or tags; Sorting the retrieved matching candidate recommendation documents according to the sorting parameters to obtain sorted candidate recommendation documents; Based on historical browsing records, the sorted candidate recommendation documents are filtered to obtain unbrowsed candidate recommendation documents; Re-ranking the unbrowsed candidate recommendation documents according to the object information and the personalized data to obtain the recommendation recall result; The search recall result and the recommendation recall result are fused to obtain the business recall result.

15. The method according to any one of claims 1 to 14, characterized in that: The attention information includes a content update notification of the subscription field; the method further includes: The subscription field is obtained from the scheduling queue, and one or more local search documents matching the subscription field are retrieved from the local search document library by searching the recall link; the document publishing time of the local search documents in the local search document library is all within the second preset time period; Determine a subject or tag mapped to the subscription field from a recall recommendation document library provided by the recommendation recall link, and retrieve one or more candidate recommendation documents matching the subscription field in the recall recommendation document library based on the mapped subject or tag; Based on the historical browsing record of the object, one or more local search documents retrieved in the search recall link are filtered to obtain a first filtering result; Based on the historical browsing record, one or more candidate recommended documents retrieved by the recommendation recall link are filtered to obtain a second filtering result; Performing a fusion process on the first filtering result and the second filtering result to obtain an updated recall result; If the update recall result is not an empty set, a content update notification of the subscription field is generated based on the update recall result.

16. An information processing device, characterized in that: include: A display unit, used to display an information retrieval interface, wherein the information retrieval interface is used to perform information retrieval based on a search field; The display unit is further used to display a subscription field, where the subscription field refers to a configured search field; The processing unit is used to output the attention information related to the subscription field.

17. A computer device, characterized in that: include: a processor suitable for executing a computer program; A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, the information processing method according to any one of claims 1 to 15 is executed.

18. A computer-readable storage medium, characterized in that: The computer storage medium stores a computer program, and when the computer program is executed by a processor, the information processing method according to any one of claims 1 to 15 is executed.

19. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the information processing method according to any one of claims 1 to 15 is implemented.

Images