Interactive processing method, device, and electronic device for information flow

Abstract

The present application provides a method, device, electronic device, and computer-readable storage medium for interactive processing of information streams. The method includes: displaying at least one piece of information in the information stream and a comment entry corresponding to the at least one piece of information; in response to a triggering operation on the comment entry corresponding to target information, wherein the target information is information to be commented on in the information stream, displaying at least one emoticon material in a comment editing area of ​​the target information, wherein the at least one emoticon material is related to a high-frequency comment word in the information stream, and the comment editing area is used to publish a comment corresponding to the target information based on the at least one emoticon material. Through this application, the efficiency of comment editing can be improved and the expression of comments can be enriched.

Description

Technical Field

[0001] The present application relates to Internet technology, and in particular to a method, device, electronic device, and computer-readable storage medium for interactive processing of information flows. Background Art

[0002] With the development and popularization of terminal applications, terminals are widely used for browsing information, such as news and social media updates. Users can comment on information in terminal applications, and user comments can promote the dissemination and promotion of information. However, related art techniques for posting comments require long editing times, are inefficient, and struggle to vividly express users' subjective feelings about the information. Summary of the Invention

[0003] The embodiments of the present application provide a method, device, electronic device, and computer-readable storage medium for interactive processing of information flow, which can improve the efficiency of comment editing and enrich the expression of comments.

[0004] The technical solution of the embodiment of the present application is implemented as follows:

[0005] The present invention provides an interactive processing method for information flow, including:

[0006] Displaying at least one piece of information in the information stream and a comment entry corresponding to the at least one piece of information;

[0007] In response to a triggering operation on a comment entry corresponding to the target information, displaying at least one emoticon material in a comment editing area of ​​the target information;

[0008] The target information is the information to be commented on in the information flow, the at least one expression material is related to high-frequency comment words in the information flow, and the comment editing area is used to publish comments corresponding to the target information based on the at least one expression material.

[0009] The present invention provides an interactive processing device for information flow, including:

[0010] A display module is used to display at least one piece of information in an information flow and a comment entry corresponding to the at least one piece of information; in response to a trigger operation on a comment entry corresponding to target information, at least one expression material is displayed in a comment editing area of ​​the target information; wherein, the target information is information to be commented on in the information flow, the at least one expression material is related to high-frequency comment words in the information flow, and the comment editing area is used to publish a comment corresponding to the target information based on the at least one expression material.

[0011] In the above solution, the high-frequency comment words in the information flow include: keywords in the hot comments of the target information, and the popularity value of the keywords is greater than the popularity threshold; the display module is further used to:

[0012] In the comment editing area of ​​the target information, a plurality of expression materials related to keywords in the hot comments of the target information are displayed in order;

[0013] The sorting method of the plurality of expression materials includes: the popularity value of the keywords corresponding to the expression materials in the plurality of historical comments of the target information; and the usage frequency of the expression materials in editing comments.

[0014] In the above solution, the high-frequency comment words in the information flow include: keywords in the hot comments of the target information, and the popularity value of the keywords is greater than the popularity threshold; the display module is further used to:

[0015] In response to a comment editing operation, obtaining text entered in the comment editing area;

[0016] When the text includes a keyword in a hot comment of the target information, it is determined that the at least one expression material will be displayed in the comment editing area.

[0017] In the above solution, the high-frequency comment words in the information flow include: keywords in comments posted by interactive accounts, wherein the interactive accounts are accounts that have an interactive relationship with the login account, and the login account is the account used to obtain and display the information flow; the display module is further used to:

[0018] In response to a comment editing operation, obtaining text entered in the comment editing area;

[0019] When it is recognized that the text includes a keyword in the comment posted by the interactive account, it is determined that at least one expression material related to the keyword in the comment posted by the interactive account will be displayed in the comment editing area.

[0020] In the above solution, the display module is further used for:

[0021] Before identifying the keywords included in the text, a plurality of emoticon materials corresponding to the plurality of keywords are sequentially displayed in the comment editing area, wherein the order of the plurality of emoticon materials includes: the popularity value of the keywords corresponding to the emoticon materials in a plurality of historical comments on the target information; the frequency of use of the emoticon materials in editing comments;

[0022] After identifying the keywords included in the text, the expression materials corresponding to the keywords are highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

[0023] In the above solution, the display module is further used for:

[0024] Before identifying the keywords included in the text, a plurality of preset expression materials are displayed in the comment editing area in order, wherein the order of the plurality of preset expression materials includes: the frequency of use of the expression materials in editing comments;

[0025] After identifying the keywords included in the text, the expression materials corresponding to the keywords are inserted and highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

[0026] In the above solution, the display module is further used for:

[0027] The comment editing area of ​​the target information is updated in real time based on at least one of the expression materials corresponding to the keywords in the comments in the display state of the target information and the expression materials corresponding to the keywords in the comments of the logged-in account participating in the interaction; wherein the logged-in account is the account used to obtain and display the information flow.

[0028] In the above solution, the display module is further used for:

[0029] Display at least part of the content of at least one piece of information in the information stream, and in response to an information selection operation, display the details of the selected target information, at least part of the comments on the target information, and a comment entry corresponding to the target information; or

[0030] At least a portion of the content of at least one piece of information in the information stream and a comment entry corresponding to each piece of information are displayed.

[0031] In the above solution, the display module is further used for:

[0032] Determining the activity level of a login account in the information stream, wherein the login account is an account used to obtain and display the information stream;

[0033] Displaying expression materials in a comment editing area of ​​the target information in a number positively correlated with the activity level.

[0034] In the above solution, the display module is further used for:

[0035] In response to a comment posting operation of a login account, wherein the login account is an account used to obtain and display the information flow, the text entered by the login account in the comment editing area and the expression material selected from the at least one expression material are combined into a comment, and

[0036] The information is sent to the server that publishes the information stream based on the identity of the login account, so that the server is synchronized with other accounts that obtain the information stream.

[0037] In the above solution, the display module is further used for:

[0038] In response to a comment posting operation by a logged-in account, wherein the logged-in account is an account used to obtain and display the information stream, the text entered by the logged-in account in the comment editing area is sent as a comment to the server that publishes the information stream based on the identity of the logged-in account, so that the server is synchronized with other accounts that obtain the information stream;

[0039] At least one emoticon material automatically selected from the at least one emoticon material is sent as a comment to the server that publishes the information stream based on the identity of the login account or the identity of the comment robot, so that the server is synchronized with other accounts that obtain the information stream.

[0040] In the above solution, the display module is further used for:

[0041] In response to the comment posting operation of the automatic comment target comment of the login account, wherein the login account is the account used to obtain and display the information flow, the target comment is the comment of the target information in the display state, at least one expression material selected from the at least one expression material related to the target comment is used as a comment, and

[0042] The identity of the login account or the identity of the comment robot is sent to the server that publishes the information stream, so that the server is synchronized with other accounts that obtain the information stream.

[0043] In the above solution, the display module is further used for:

[0044] Acquire at least one expression material related to the comment of the target information in the following manner:

[0045] Obtaining multiple comments on the target information;

[0046] Determine the popularity value corresponding to each word in the comment;

[0047] The words with a popularity value higher than the popularity threshold are regarded as keywords, and the comments containing the keywords are regarded as the first hot comments;

[0048] Generate corresponding emoticon materials based on the first hot comment.

[0049] In the above solution, the display module is further used for:

[0050] Determine a text vector of a comment corresponding to each of the multiple messages in the information stream;

[0051] performing clustering processing on the text vectors corresponding to the comments of the plurality of pieces of information to obtain at least one cluster center;

[0052] Determine the number of text vectors similar to each cluster center, and use the cluster center whose number exceeds the clustering threshold as the hotspot cluster center;

[0053] Determining at least one text vector similar to the hotspot cluster center among the text vectors corresponding to the comments of the plurality of pieces of information, and using the comment of the information corresponding to the at least one text vector as a second hotspot comment;

[0054] Taking the union of the first hot comments and the second hot comments as a hot comments set;

[0055] Generate an expression material corresponding to each comment in the hot comment set.

[0056] In the above solution, the display module is further used for:

[0057] Acquire multiple comments corresponding to the target information published during a sampling period; when the number of the multiple comments is less than a quantity threshold, wherein the quantity threshold is a minimum value for identifying hot comments, acquire multiple comments corresponding to the quantity threshold in the target information in descending order of publication time; or,

[0058] Determine an associated account corresponding to a login account based on at least one dimension, wherein the login account is an account used to obtain and display the information flow, and the dimension includes at least one of the following: a social distance from the login account, a geographical distance from the login account; and obtain a plurality of comments published by the associated account on the target information.

[0059] In the above solution, the display module is further used for:

[0060] Determining text material based on the first hot comment, wherein the text material includes at least one of the following: the first hot comment, keywords in the first hot comment, and other text with the same semantics as the first hot comment;

[0061] Acquire an image material that matches the sentiment feature of the first hot comment from a plurality of candidate material images, or fuse the candidate material image with the sentiment feature of the first hot comment to obtain the image material;

[0062] The text material and the image material are combined into an expression material.

[0063] An embodiment of the present application provides an electronic device, including:

[0064] a memory for storing executable instructions;

[0065] The processor is used to implement the interactive processing method of the information flow provided in the embodiment of the present application when executing the executable instructions stored in the memory.

[0066] An embodiment of the present application provides a computer-readable storage medium storing executable instructions for implementing the interactive processing method of information flow provided in the embodiment of the present application when executed by a processor.

[0067] An embodiment of the present application provides a computer program product or a computer program, wherein the computer program product or the computer program includes computer instructions, which are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the interactive processing method of information flow provided in the embodiment of the present application.

[0068] The embodiments of the present application have the following beneficial effects:

[0069] When a trigger operation is received for the comment entry corresponding to the target information, at least one emoticon material related to the high-frequency comment words in the information flow is displayed in the comment editing area of ​​the target information, so that the user can publish the emoticon material as a comment without editing the text, thereby improving the efficiency of comment editing and publishing. In addition, the emoticon material can vividly convey the user's subjective feelings towards the target information, enriching the expression form of the comment. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1A is a schematic diagram of the architecture of the interactive processing system 10 provided in an embodiment of the present application;

[0071] Figure 1B is a schematic diagram of the architecture of the interactive processing system 10 provided in an embodiment of the present application;

[0072] Figure 2A It is a flowchart of the interactive processing method of information flow provided by the embodiment of the present application;

[0073] Figure 2BIt is a flowchart of the interactive processing method of information flow provided by the embodiment of the present application;

[0074] Figure 3A This is a schematic diagram of an information flow page provided by an embodiment of the present application;

[0075] Figure 3B is a schematic diagram of a details page of target information provided in an embodiment of the present application;

[0076] Figure 3C This is a schematic diagram of an information flow page provided by an embodiment of the present application;

[0077] Figure 3D Schematic diagram of the sorting of expression materials provided in the embodiment of the present application;

[0078] Figure 3E Schematic diagram of the sorting of expression materials provided in the embodiment of the present application;

[0079] Figure 3F Schematic diagram of updating expression materials provided in an embodiment of the present application;

[0080] Figure 3G This is a schematic diagram of the automatic review provided by an embodiment of the present application;

[0081] Figure 3H This is a schematic diagram of triggering automatic review provided by an embodiment of the present application;

[0082] Figure 4 Schematic diagram of the interactive processing system provided in an embodiment of the present application;

[0083] Figure 5 This is a schematic diagram of hot comment mining provided by an embodiment of the present application;

[0084] Figure 6 This is a schematic diagram of the generation of expression materials provided in the embodiment of the present application;

[0085] Figure 7 This is a flowchart of determining hot comments provided by an embodiment of the present application;

[0086] Figure 8 Schematic diagram of the training BERT model provided in the embodiment of the present application;

[0087] Figure 9 is a schematic diagram of vectorization provided by an embodiment of the present application;

[0088] Figure 10 Schematic diagram of the distance between the text vector of the comments and the cluster center provided in the embodiment of the present application;

[0089] Figure 11 is a schematic diagram of the expression material provided in the embodiment of the present application;

[0090] Figure 12 A schematic diagram of generating expression materials provided in the application embodiment;

[0091] Figure 13 This is a schematic diagram of posting comments provided by an embodiment of the present application;

[0092] Figure 14 It is a structural diagram of the terminal 400-1 provided in an embodiment of the present application. DETAILED DESCRIPTION

[0093] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0094] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0095] In the following description, the terms "first / second / third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understandable that "first / second / third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0096] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0097] The relevant data collection and processing in the embodiments of this application should be strictly in accordance with the requirements of relevant laws and regulations when applied in examples, and the informed consent or separate consent of the personal information subject should be obtained. Subsequent data use and processing should be carried out within the scope of authorization of laws and regulations and the personal information subject.

[0098] Before further describing the embodiments of the present application in detail, the nouns and terms involved in the embodiments of the present application are explained. The nouns and terms involved in the embodiments of the present application are subject to the following interpretations.

[0099] 1) Terminal: An electronic device used by a user to receive messages, which has an intelligent operating system installed in it.

[0100] 2) User: A person who uses an electronic device.

[0101] 3) Terminal programs: Products that run on terminals and have various functions, such as instant messaging programs, shopping programs, browser programs, etc.

[0102] 4) Information flow products: a product form of terminal program, on which various video and article information can be obtained.

[0103] 5) Comments: Users’ responses to videos or articles in information flow products.

[0104] 6) Emoticon material: also known as emoticon package, which is a digital media block featuring images. It is a series of images formed after the rise of social software or social networking sites. It usually uses popular celebrities, quotations, cartoons, film and television screenshots, etc. as materials, and is accompanied by a series of matching texts to express specific emotions.

[0105] 7) Hotspot: An event that attracts the attention of a large number of users within a certain period of time.

[0106] 8) Bidirectional Encoder Representations from Transformers (BERT): Utilizes the left and right context of text to pre-train deep bidirectional neural network representations. During the fine-tuning phase, only a small amount of labeled data is needed to complete training for specific tasks.

[0107] 9) K-means clustering algorithm: Also known as the "K-means algorithm," this algorithm originated as a vector quantization method in signal processing and is now more popular in data mining as a cluster analysis method. Its goal is to partition N points (which can be a single observation or instance of a sample) into k clusters, ensuring that each point belongs to the cluster with the nearest mean (i.e., the cluster center), using this as the clustering criterion.

[0108] 10) N-gram: This refers to n consecutive words in a text. The n-gram model is a probabilistic language model based on an (n-1)-order Markov chain, which infers the structure of a sentence by the probability of n words appearing.

[0109] In the related art, comment editing takes a long time, is inefficient, and fails to vividly express the user's subjective feelings about the information. To address these issues, embodiments of the present application provide a method, device, electronic device, and computer-readable storage medium for interactive processing of information streams, which can improve the efficiency of comment editing and enrich the expression of comments.

[0110] The interactive processing method of information flow provided in the embodiment of the present application can be implemented by various electronic devices. For example, it can be implemented by a terminal alone, or it can be implemented by a server and a terminal in collaboration. For example, the terminal alone executes the interactive processing method of information flow described below, or the terminal and the server collaborate to execute the interactive processing method of information flow described below. For example, in response to a trigger operation on a comment entry corresponding to target information in an information flow, the terminal displays at least one emoticon material in the comment editing area of ​​the target information, and in response to a comment publishing operation of a logged-in account on the content (such as text and emoticon material) entered in the comment editing area, sends the content entered in the comment editing area to the server that publishes the information flow, and the server synchronizes the content to the terminals corresponding to other accounts that obtain the information flow.

[0111] In the embodiments of the present application, the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the test device can be a smartphone, tablet computer, laptop computer, desktop computer, smart TV, etc., but is not limited to these. The test device and the server can be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiments of the present application.

[0112] Taking servers as an example, it can be a server cluster deployed in the cloud, opening artificial intelligence cloud services (AIaaS, AI as a Service) to users. The AIaaS platform will split several common AI services and provide independent or packaged services in the cloud. This service model is similar to an AI theme mall. All users can access and use one or more artificial intelligence services provided by the AIaaS platform through the application programming interface.

[0113] For example, one AI cloud service may be an interactive processing service, where a cloud server encapsulates an interactive processing program provided by an embodiment of the present application. In response to a trigger operation for a comment entry corresponding to target information in an information stream, the terminal sends an expression material acquisition instruction to the cloud server. Upon receiving the expression material acquisition instruction, the cloud server invokes the encapsulated interactive processing program, analyzes the comments on the information in the information stream, generates at least one expression material related to the comment on the target information, and sends the at least one expression material to the terminal, which then displays the at least one expression material in the comment editing area of ​​the target information on the terminal.

[0114] See also Figure 1A , Figure 1A: This is a schematic diagram of the architecture of the interactive processing system 10 provided in an embodiment of the present application. The terminal 400-1 corresponding to the login account is connected to the server 200 via the network 300. The network 300 can be a wide area network or a local area network, or a combination of the two. In response to a trigger operation on the comment entry corresponding to the target information in the information stream, the terminal 400-1 displays at least one emoticon material in the comment editing area of ​​the target information. In response to the login account's comment publishing operation on the comment (such as text and emoticon material) entered in the comment editing area, the terminal 400-1 sends the comment in the comment editing area to the server 200 that publishes the information stream. The server 200 synchronizes the comment to the terminal 400-2 corresponding to the other account that obtains the information stream.

[0115] In some embodiments, see Figure 1B , Figure 1B : This is a schematic diagram of the architecture of the interactive processing system 10 provided in an embodiment of the present application. Both the server and the terminal can join the blockchain network 500 and become one of the nodes (i.e., a node in the blockchain network 500 is running). The type of blockchain network 500 is flexible and diverse. For example, it can be any one of a public chain, a private chain, or a consortium chain. Taking the public chain as an example, any electronic device of a business entity, such as a server, can access the blockchain network 500 without authorization to serve as a consensus node of the blockchain network 500. For example, the server 200 is mapped to the consensus node 500-2 in the blockchain network 500, the terminal 400-2 is mapped to the consensus node 500-0 in the blockchain network 500, and the terminal 400-3 is mapped to the consensus node 500-1 in the blockchain network 500.

[0116] Taking blockchain network 500 as an example of a consortium chain, the server can access blockchain network 500 after obtaining authorization. After receiving the comment on the target information sent by terminal 400-1, server 200 first determines the first hot comment corresponding to the target information before synchronizing the comment to the terminals corresponding to other accounts that obtain the information flow, and sends the first hot comment to other terminals (such as terminal 400-2 and terminal 400-3). Other terminals can confirm the first hot comment by executing a smart contract to verify whether the first hot comment determined by terminal 400-1 is a hot comment. When more than a threshold number of nodes confirm that the verification is passed, a digital signature (i.e., endorsement) will be signed for it. When the first hot comment has sufficient endorsements, server 200 generates corresponding emoticon materials based on the first hot comment, and obtains emoticon materials corresponding to the keywords in the comment sent by terminal 400-1 from the generated emoticon materials, and sends the emoticon materials as follow-up comments and comments to other terminals. At the same time, blockchain network 500 will store the first hot comment on the chain. When other accounts post comments, the stored first hot comment can be used to determine whether the comments posted by other accounts contain keywords.

[0117] It can be seen that in the embodiment of the present application, the accuracy of the first hot comment in the hot comment set can be guaranteed by performing consensus verification.

[0118] The interactive processing method of information flow provided in the embodiment of the present application will be described below with reference to the accompanying drawings. The execution subject of the following interactive processing method of information flow can be a terminal, which can be implemented by the terminal running the various computer programs mentioned above, or it can be implemented by a client, an operating system, a software module, a script or a small program executing a computer program; of course, based on the understanding of the following, it is not difficult to see that the interactive processing method of information flow provided in the embodiment of the present application can also be implemented by the terminal and the server in collaboration.

[0119] See also Figure 2A , Figure 2A This is a flow chart of the interactive processing method of information flow provided by the embodiment of the present application, which will be combined with Figure 2A The steps shown are explained.

[0120] In step 101, an information flow is obtained.

[0121] In some embodiments, a terminal can actively obtain information streams from a server, or the server can push information streams to the terminal. The obtained information streams can be presented in various information stream products on the terminal, such as instant messaging applications, browser applications, shopping applications, live streaming applications, etc. The information streams include multiple pieces of information, and the types of information can be social, technology, sports, entertainment, etc. The information can include text, images, video, audio, etc.

[0122] In step 102, at least one piece of information in the information stream and a comment entry corresponding to the at least one piece of information are displayed.

[0123] In some embodiments, displaying at least one piece of information in an information stream and a comment entry corresponding to the at least one piece of information can be implemented in the following manner: displaying at least a portion of the content of at least one piece of information in the information stream, and in response to an information selection operation, displaying details of the selected target information, at least a portion of the comments on the target information, and a comment entry corresponding to the target information.

[0124] After the information stream is acquired, it will be displayed on the information stream page. When there is too much information in the information stream and it is impossible to fully display all the information on the information stream page, partial information (at least one piece of information) can be displayed on the information stream page, and all the information can be browsed by dragging the scroll bar (horizontal scroll bar or vertical scroll bar) or sliding the information stream page. When displaying partial information, the complete content of the information can be displayed, or at least part of the content of the information (such as full text, summary, illustrations, etc.) can be displayed. For example, for news, the summary and illustrations of the information can be displayed on the information stream page. In response to the information selection operation for the target information in the information stream page, the detailed content of the selected target information, at least part of the comments on the target information, and the comment entry corresponding to the target information are displayed on the details page of the target information. Among them, when the target information has fewer comments (not exceeding the preset number of comments, or all comments can be fully presented in the details page of the target information), all comments are displayed; when the target information has more comments (exceeding the preset number of comments, or all comments cannot be fully presented in the details page of the target information), the comments can be displayed selectively, for example, the most popular comments, the latest comments, or comments posted by interactive accounts that have an interactive relationship with the login account that obtains the information flow, etc., and other comments are folded and displayed through the comment expansion entrance. When the comment expansion entrance is triggered, all comments are displayed.

[0125] like Figure 3A As shown, Figure 3A The information flow page is a schematic diagram of the information flow page provided by the embodiment of the present application. The information flow page displays part of the content (abstract and drawings) of some information. In response to the information selection operation for the target information 301, the information flow page is displayed. Figure 3B In the target information details page shown, because the target information has only one comment, all comments can be displayed. At the same time, the comment entry 302 corresponding to the target information is displayed, allowing users to comment on the target information through the comment entry 302. In this way, users can understand the specific content of the information on the details page and then make objective comments on it.

[0126] In other embodiments, displaying at least one piece of information in an information stream and a comment entry corresponding to the at least one piece of information may also be achieved by displaying at least a portion of the content of the at least one piece of information in the information stream and a comment entry corresponding to each piece of information. In some possible examples, while displaying multiple pieces of information in an information stream, multiple existing comments under each piece of information may also be displayed.

[0127] like Figure 3C As shown, Figure 3C This is a schematic diagram of the information flow page provided in an embodiment of the present application. Figure 3C In the example, each displayed information has a corresponding comment entry 303 . For example, the first information has a corresponding comment entry 303 , and the user can comment on the first information through the comment entry 303 .

[0128] In this way, the comment entry corresponding to each information is displayed in the information flow page. Users can comment on the information without entering the details page of the information, which can increase user participation in the information. It can also help users understand the general content of the information through the existing comments on the information displayed in the information flow page, determine whether the information is worth browsing, help users make browsing decisions, save selection time, and save the complicated operations of entering and exiting the details page.

[0129] In step 103 , in response to a triggering operation on a comment entry corresponding to the target information, at least one expression material is displayed in a comment editing area of ​​the target information.

[0130] In some embodiments, the target information may be information to be commented on in the information flow, that is, the logged-in account can comment on the information, and at the same time, there may be multiple other accounts that have posted comments on the target information. Expression materials are images used to express user emotions, including static images (such as emoji emotional symbols) and animated images. Expression materials are displayed in the comment editing area by default. The comment editing area is used to publish comments corresponding to the target information based on at least one expression material. High-frequency comment words include at least one of the following: keywords in the hot comments of the target information, keywords in the hot comments of all information in the information flow. Keywords are words with a heat value greater than a heat threshold. At least one expression material related to the high-frequency comment words of the information flow includes at least one of the following: at least one expression material related to the keywords in the hot comments of the target information, and at least one expression material related to the keywords in the hot comments of all information in the information flow.

[0131] In some embodiments, before displaying at least one expression material in the comment editing area of ​​the target information, it is necessary to Figure 2B Steps 1031 to 1034 in the embodiment of the present invention obtain at least one expression material related to the comment of the target information.

[0132] In step 1031 , multiple comments on the target information are obtained.

[0133] In some embodiments, obtaining multiple comments on target information can be implemented in the following manner: obtaining multiple comments corresponding to the target information published during a sampling period; and when the number of the multiple comments is less than a quantity threshold, obtaining multiple comments in the target information corresponding to the quantity threshold in descending order of publication time.

[0134] Among them, the sampling period is the granularity of the interval time for obtaining comments, and the sampling period can be 12 hours, one day, one week, etc. The quantity threshold is the minimum value used to identify hot comments. After obtaining multiple comments on the corresponding target information published in the sampling period, the number of multiple comments is determined. If the number is greater than or equal to the quantity threshold, step 1032 is executed. If the number is less than the quantity threshold, the hot comments cannot be accurately identified due to the small number. At this time, the sampling period is no longer used as a restriction, and multiple comments with the same number as the quantity threshold are obtained in the order of publication time. For example, the sampling period is set to 1 day, and the quantity threshold is 100. If the target information has 50 comments in 1 day, which is less than the quantity threshold, the first 100 comments are obtained according to the publication time of the comments. In this way, the number of comments obtained can be ensured, which facilitates the subsequent accurate determination of hot comments.

[0135] In other embodiments, obtaining multiple comments on the target information may also be achieved by: determining an associated account corresponding to the login account based on at least one dimension; and obtaining multiple comments posted by the associated account on the target information.

[0136] Among them, the login account is an account used to obtain and display the information flow, and the dimensions include at least one of the following: social distance from the login account, geographical distance from the login account. Social distance indicates the number of intermediate accounts that need to be passed from the login account to the associated account. Based on at least one dimension, the associated account corresponding to the login account is determined, which can be an account whose social distance from the login account is within the social distance threshold as the associated account, or an account whose geographical distance from the login account is within the geographical distance threshold as the associated account. If there is at least one account in the associated account that has commented on the target information, the comments posted by the at least one account are obtained. Compared with general comments, the user corresponding to the login account is more likely to pay attention to the comments posted by the associated account with which it is associated. Therefore, obtaining the comments posted by the associated account can facilitate the subsequent determination of hot comments that the user is concerned about.

[0137] It should be noted that when at least one expression material related to high-frequency comment words in an information flow includes at least one expression material related to keywords in hot comments of all information in the information flow, the above-mentioned acquisition of multiple comments on the target information is actually the acquisition of multiple comments on all information in the information flow. At this time, the method of acquiring multiple comments on all information in the information flow is the same as the method of acquiring multiple comments on the target information, and steps 1032 to 1034 are also applicable.

[0138] In step 1032, the popularity value corresponding to each word in the comment is determined.

[0139] In some embodiments, the comments are first segmented to obtain corresponding segmentation results. For example, if the comment is "I gave my deskmate a bouquet of flowers today", after segmentation, the segmentation results (i.e., the words obtained) are: I, today, gave, classmate, a bouquet, and flowers.

[0140] N-gram can be used to determine the probability of a specific word appearing after one or more words in a sentence. For example, for "The weather is very good today", n-gram can determine the probability of "weather" appearing after "today", that is, the probability of the word "today's weather" appearing. In an embodiment of the present application, n-gram can be used to determine all possible words in the comment based on the word segmentation processing result. When n=1, a single word is obtained, which is consistent with the word segmentation processing result. When n=2, the word is formed by combining two adjacent words in the word segmentation processing result in sequence, which can be: I today, I sent today, I sent a classmate, a bouquet of flowers to my classmate, a bouquet of flowers. When n=3, the word is formed by combining three adjacent words in the word segmentation processing result in sequence, which can be: I sent today, I sent a classmate today, I sent a bouquet of flowers to my classmate, a bouquet of flowers to my classmate. The embodiment of the present application obtains words when n is different values, and all possible words corresponding to the comment can be obtained. For example, if n takes the values ​​of 1, 2, and 3 respectively, then for the comment "I gave my deskmate a bouquet of flowers today", all possible corresponding words can be obtained: I, today, gave, classmate, a bouquet of, flowers, I today, gave today, gave my classmate, my classmate a bouquet, a bouquet of flowers, I gave today, gave my classmate today, gave my classmate a bouquet, my classmate a bouquet of flowers.

[0141] Related technologies only determine the popularity value of a single word, but in real life, keywords or hot comments usually do not appear in the form of a single word, but are composed of multiple words. The embodiment of the present application uses n-grams to determine all possible words. These words can be single words or composed of multiple words, thereby greatly increasing the number of candidate keywords and improving the accuracy of the determined keywords and hot comments.

[0142] Then, the frequency of each word appearing in all the comments is counted as the corresponding word frequency. For example, if the sampling period is one week and the word "sent today" appears a total of 3 times in all the comments in a week, then its word frequency is 3.

[0143] Afterwards, the corresponding popularity value can be calculated based on the word frequency of each word. The popularity value is a numerical value used to measure the popularity of a word. In some possible examples, for the i-th word in the comment, its popularity value can be calculated by the following formula (1):

[0144]

[0145] Among them, count(w i ) represents the frequency of each word on the day when the frequency is counted; t represents time, and its value range is d~T, T=6, d is the time difference between the day when the frequency is counted and the previous day (the difference is a day within a week), and its value range is 0~6.

[0146] In this way, the popularity value of each word can be accurately determined, and whether the corresponding word is a keyword can be determined.

[0147] In step 1033, words with a popularity value higher than a popularity threshold are used as keywords, and comments including the keywords are used as first hot comments.

[0148] In some embodiments, words with a popularity value higher than a popularity threshold can be used as keywords, or the words can be sorted in descending order by popularity value, with the top words being used as keywords. Accordingly, the comments containing the keywords are used as the first hot comments. For example, if the keyword is "I today", the first hot comment can be "I cleaned the house today".

[0149] In step 1034, corresponding expression material is generated based on the first hot comment.

[0150] In some embodiments, generating corresponding emoticon materials based on the first hot comment can be implemented in the following manner. First, determine the text vector of the comment corresponding to each of the multiple messages included in the information stream. The text vector corresponding to the comment can be generated by a model such as a BERT model, a text convolutional neural network (Text CNN), a long short-term memory network (LSTM), or a tiny BERT model.

[0151] Then, the text vectors corresponding to the comments of the multiple messages are clustered to obtain at least one cluster center. In some possible examples, the cluster center can be determined by methods such as K-means algorithm, mean shift clustering, density-based clustering, and agglomerative hierarchical clustering.

[0152] The text vectors corresponding to multiple comments can be randomly divided into multiple categories, and the central vectors of each category (the text vector at the center) can be initialized; the distance from each text vector to each central vector can be calculated, and the text vectors with a distance less than a distance threshold can be divided into the category where the corresponding central vector is located; the text vectors in each category that have the same distance to each text vector can be re-used as the central vector of this category, and the central vector can be used as the cluster center. The above steps of calculating the distance, re-dividing the text vectors, and re-determining the central vector can be iterated multiple times until the central vector of each category remains basically unchanged after the iteration, then the iteration can be stopped, and the final central vector can be used as the cluster center. In this way, the cluster center of multiple text vectors can be accurately determined, and the cluster center is the text vector that is most likely to correspond to the hot comment.

[0153] Afterwards, the number of text vectors similar to each cluster center is determined, and the cluster centers whose corresponding number exceeds the text vector number threshold are regarded as hot cluster centers.

[0154] In some possible examples, the Euclidean distance between the text vector in each category and the cluster center can be calculated, and the text vectors with a Euclidean distance less than a preset distance are regarded as text vectors similar to the cluster center. The number of text vectors similar to the cluster center in each category is counted, and the cluster centers whose corresponding number exceeds the text vector number threshold are regarded as hotspot cluster centers. Among them, the Euclidean distance can be replaced by cosine distance, Manhattan distance, Pearson correlation coefficient, etc. It can be seen that the hotspot cluster center is a text vector that is semantically close to most text vectors (small Euclidean distance), and the hotspot cluster center can be quickly determined by calculating the distance between vectors.

[0155] Afterwards, at least one text vector similar to the hotspot cluster center is determined among the text vectors corresponding to the comments of the multiple messages, and the comment of the message corresponding to the at least one text vector is taken as the second hotspot comment, and the union of the first hotspot comment and the second hotspot comment is taken as the hotspot comment set. The comment corresponding to the hotspot cluster center is a hotspot comment, and accordingly, the comment corresponding to the text vector similar to the hotspot cluster center is also a hotspot comment. The hotspot comments in the hotspot comment set can be determined by calculating the heat value, or by the cluster center corresponding to the text vector of the comment. In this way, the number of hotspot comments can be increased, thereby providing users with more corresponding expression materials.

[0156] Finally, an expression material corresponding to each comment in the hot comment set is generated.

[0157] In some embodiments, generating corresponding expression materials based on the first hot comment can be achieved in the following manner: determining text materials based on the first hot comment, where the text materials include at least one of the following: the first hot comment, keywords in the first hot comment, and other text with the same semantics as the first hot comment; obtaining image materials that match the emotional characteristics of the first hot comment from multiple candidate material images, or fusing the candidate material images with the emotional characteristics of the first hot comment to obtain image materials; and combining the text materials and the image materials into expression materials.

[0158] Among them, the candidate material image can be an online image or a user's face image. Taking the candidate material image as a face image as an example, the emotional features of the first hot comment are extracted, and the emotional features of the candidate material image are extracted, and the similarity of the emotional features of the two is calculated. When the similarity is greater than the similarity threshold, it is determined that the candidate material image matches the first hot comment. At this time, the facial expression (such as a smiling expression) in the face image corresponds to the emotion expressed by the first hot comment (happy). In some possible examples, when the candidate material image has a corresponding label (such as "smiling face", "crying", "sad", etc.), the emotional features of the label can also be extracted, and the emotional features of the first hot comment are matched with the emotional features of the label of the candidate material image, and the candidate material image corresponding to the matched label is used as the image material. For example, the emotional features of the first hot comment represent a joyful mood, and the candidate material image with the label "smiling face" is the image material that matches it. It can be seen that the embodiment of the present application can find the expression material that accurately conveys the emotion of the hot comment by obtaining the image material that matches the emotional features of the first hot comment. There is no need to further process the expression material, and it can be directly used, which is more convenient.

[0159] In some possible examples, the candidate material image and the emotional features of the first hot comment are fused to obtain image material. This process can be performed by encoding the first hot comment and the candidate material image through a neural network model such as an autoencoder network model to obtain the corresponding emotional features, and then decoding the obtained emotional features to generate image material expressing the corresponding emotions. For example, the user's face image and the emotional features of the first hot comment (happy) can be fused to generate a smiling face image. It can be seen that the embodiment of the present application can obtain the corresponding image material by fusing the emotional features of the candidate material image and the first hot comment. While retaining the original content of the candidate material image, the emotion in the hot comment can be conveyed through the expression material, and the material image as a whole is more natural and vivid.

[0160] In some possible examples, the text material may be fused with an existing background template to generate a corresponding expression material; or the image material may be fused with an existing background template to generate a corresponding expression material.

[0161] In some embodiments, the high-frequency comment words in the information flow include at least one of the following: keywords in the hot comments of the target information, keywords in the hot comments of all the information in the information flow, and the heat value of the keywords is greater than a heat threshold. Before displaying at least one emoticon material in the comment editing area of ​​the target information, in response to the comment editing operation, the text entered in the comment editing area is obtained; when the text includes the keywords in the hot comments of the target information, or when the text includes the keywords in the hot comments of all the information in the information flow, it is determined that at least one emoticon material will be displayed in the comment editing area.

[0162] The emoticons in the comment editing area are displayed in a default order. When it is determined that the text contains keywords in a hot comment, the position of at least one emoticon corresponding to the keyword will be adjusted to the front of the comment editing area, and the sort order of other emoticons will be moved to the back. Therefore, at least one emoticon related to the keyword will be displayed in the comment editing area. This makes it easier for users to choose emoticons corresponding to keywords in the text to post comments, meet user needs, quickly find emoticons that meet their needs, increase the usage rate of emoticons corresponding to keywords, increase the number of comments posted by users, and improve user experience.

[0163] In some embodiments, the at least one emoticon material associated with a comment on the target information includes: at least one emoticon material associated with a keyword in a comment (e.g., a hot comment) posted by an interactive account, wherein the interactive account is an account with which a login account has an interactive relationship, and the login account is an account used to obtain and display the information flow. Before displaying the at least one emoticon material in the comment editing area of ​​the target information, in response to a comment editing operation, text entered in the comment editing area is obtained; when it is recognized that the text includes a keyword in a comment posted by the interactive account, it is determined that at least one emoticon material associated with the keyword in the comment posted by the interactive account will be displayed in the comment editing area.

[0164] Interactive relationships include comment relationships, like relationships, and forwarding relationships. When a text is identified as containing a keyword from a comment posted by an interactive account, at least one emoticon material related to the keyword will be displayed in the comment editing area. This makes it easier for users to choose these emoticons to post comments. Because these emoticons are related to the keywords in the comments posted by the interactive account, that is, these emoticons are also commonly used and paid attention to by the interactive account, it can encourage the interactive account to follow up on the user's comments, thereby promoting comment interaction among users.

[0165] In some embodiments, displaying at least one emoticon material in the comment editing area of ​​the target information can be implemented in the following manner: in the comment editing area of ​​the target information, multiple emoticons related to keywords in the hot comments of the target information are displayed in order; wherein the order of the multiple emoticons includes: the popularity value of the keywords corresponding to the emoticon material in multiple historical comments of the target information; the popularity value of the keywords corresponding to the emoticon material in historical comments of all information in the information stream; and the frequency of use of the emoticon material in editing comments.

[0166] The frequency of using emoticons in editing comments can be the frequency of using them in editing comments in the target information or the frequency of using them in editing comments in all information in the information flow. When the comments corresponding to the emoticons are determined by calculating the popularity value, the emoticons are sorted in descending order according to the popularity value, and the emoticons sorted in descending order are displayed in the comment editing area of ​​the target information. For example, Figure 3D In the comment editing area 304, multiple emoticon materials are displayed, and these multiple emoticon materials are displayed in descending order according to their popularity. In this way, the most popular emoticon materials can be placed at the front of the comment editing area, making it easier for users to use popular emoticon materials to comment, thereby increasing the usage rate of popular emoticon materials.

[0167] When the comments corresponding to the emoticon materials are determined using a K-means algorithm, the frequency of use of at least one emoticon material in the comment editing area is counted, and the emoticon materials are sorted in descending order based on frequency of use. The emoticon materials in descending order are then displayed in the comment editing area of ​​the target message. This allows commonly used emoticons to be placed at the top of the comment editing area, making it easier for users to select commonly used emoticons for comments and reducing the time it takes to search for emoticons.

[0168] In some embodiments, displaying at least one emoticon material in the comment editing area of ​​the target information can be implemented in the following manner. Before identifying the keywords included in the text, multiple emoticon materials corresponding to multiple keywords are displayed in the comment editing area in order, wherein the sorting method of the multiple emoticon materials includes: the popularity value of the keywords corresponding to the emoticon materials in multiple historical comments of the target information; the frequency of use of the emoticon materials in editing comments; after identifying the keywords included in the text, the emoticon materials corresponding to the keywords are highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, magnifying, text prompt, voice prompt, and sorting in front. Among them, after identifying the keywords included in the text, if the emoticon material corresponding to the keyword is sorted at the back, it will be sorted in the front, and a text prompt or voice prompt can be given to it; if the emoticon material corresponding to the keyword is in the front, it can be highlighted, magnified, text prompted, or voice prompted.

[0169] like Figure 3E As shown, if the logged-in account enters the text "I want to eat too!" in the comment editing area, it is determined that the corresponding keyword is "eat", and the order of the expression material 305 corresponding to the keyword "eat" is adjusted to the first place in the comment editing area. In this way, by adjusting the order and prompting, the user can be prompted to select the corresponding expression material to post a comment, thereby improving the utilization rate and relevance of the expression material, and also improving the relevance between the expression material and the text in the comment, so that the comment can convey more rich and detailed emotions.

[0170] In some embodiments, displaying at least one emoticon material in the comment editing area of ​​the target information can be implemented in the following manner: before identifying keywords included in the text, displaying multiple preset emoticon materials in order in the comment editing area, wherein the order of the multiple preset emoticon materials includes: the frequency of use of the emoticon materials in editing comments; after identifying keywords included in the text, inserting and highlighting the emoticon material corresponding to the keyword in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, magnifying, text prompt, voice prompt, and sorting first.

[0171] The preset emoticon materials can be ordinary emoticon materials (such as emoji emotional symbols), that is, emoticon materials corresponding to keywords are not displayed by default. After the keywords included in the text are identified, the emoticon materials corresponding to the keywords are obtained and inserted into the comment editing area (such as the front) for highlighting. In this way, it is convenient for users to choose emoticon materials corresponding to keywords to post comments, optimize the user experience, and avoid displaying a large amount of useless emoticon materials, which squeezes the display space of ordinary emoticon materials.

[0172] In some embodiments, displaying at least one emoticon material in a comment editing area of ​​target information can be implemented in the following manner: Based on at least one of an emoticon material corresponding to a keyword in a displayed comment on the target information and an emoticon material corresponding to a keyword in a comment interactively participated in by a logged-in account, the comment editing area displaying the target information is updated in real time to display the emoticon material in the comment editing area; wherein the logged-in account is the account used to obtain and display the information stream.

[0173] like Figure 3F As shown, the expression materials in the comment editing area 306 are updated (the newly added expression materials are placed in the first place, and the positions of the other expression materials are moved back accordingly), and a new comment editing area 307 is obtained.

[0174] Because emoticons can convey users' emotions more vividly, when multiple users use the same emoticon in their comments on the same information, it indicates that the users have similar emotions toward the information. Therefore, by updating the emoticons in the comment editing area of ​​the target information using the emoticons corresponding to the keywords in the displayed comments of the target information, users can quickly select emoticons to express their emotions when commenting. By updating the emoticons in the comment editing area of ​​the target information using the emoticons corresponding to the keywords in the comments of those who have logged in to participate in the interaction, the emoticons in the comment editing area can be updated to the emoticons that users are accustomed to, making it easier for users to choose the emoticons that they are most familiar with and most used to express their emotions.

[0175] In some embodiments, displaying at least one emoticon material in the comment editing area of ​​a target message can be implemented in the following manner: determining the activity level of a logged-in account in the information stream, where the logged-in account is the account used to obtain and display the information stream; and displaying, in the comment editing area of ​​the target message, emoticons whose number is positively correlated with the activity level.

[0176] In some possible examples, the use rights of all emoticon materials may also be opened to all users.

[0177] It can be seen that in the embodiment of the present application, the number of emoticon materials available to users is positively correlated with their level of activity, that is, the higher the level of activity, the more emoticon materials available. In this way, users can be encouraged to participate more deeply in the information, such as longer in-depth reading, more comments, reposts, likes, etc., thereby improving user stickiness.

[0178] In some embodiments, in response to a comment posting operation of a logged-in account, the text entered by the logged-in account in the comment editing area and the emoticon material selected from at least one emoticon material are combined into a comment through a comment editing operation, and are sent to the server that publishes the information stream based on the identity of the logged-in account, so that the server is synchronized to other accounts that obtain the information stream.

[0179] In some possible examples, the text and the expression materials may be published separately as different comments.

[0180] In some embodiments, in response to a comment posting operation by a logged-in account, the text entered by the logged-in account in the comment editing area is sent as a comment to the server that publishes the information stream based on the identity of the logged-in account, so that the server is synchronized with other accounts that obtain the information stream; at least one emoticon material automatically selected from at least one emoticon material is sent as a comment to the server that publishes the information stream based on the identity of the logged-in account or the identity of a comment robot, so that the server is synchronized with other accounts that obtain the information stream. The selected emoticon material is related to the text entered in the comment editing area.

[0181] See also Figure 3G , Figure 3G This is a schematic diagram of the automatic review provided by the embodiment of this application. Figure 3G In the example, the text posted by the login account nicknamed "Lucky" is "Go Duck", and the comment posted by the server based on the identity of the login account is the "Go Duck" expression material 308 below. In this way, after the user posts a comment, the expression material related to the comment can be automatically obtained, and the comment can be automatically followed by the expression material.

[0182] In some embodiments, in response to a comment posting operation for an automatically followed target comment by a logged-in account, where the target comment is a comment on the target information that is currently displayed, at least one emoticon selected from at least one emoticon associated with the target comment is sent as a comment to the server that publishes the information stream based on the identity of the logged-in account or the commenting bot, so that the server synchronizes the comment with other accounts that access the information stream. The selected emoticon is associated with the text entered in the comment editing area.

[0183] See also Figure 3H , Figure 3H This is a schematic diagram of triggering automatic review provided by the embodiment of this application. Figure 3H In the example, there is an automatic comment button under each comment of the target information. In response to the triggering operation of the automatic comment button 309 in the target comment "Go!" (i.e., comment publishing operation), a comment will be automatically posted under the target comment with the corresponding "Go!" expression material, such as Figure 3GAs shown in the "Go Duck" expression material 308 in FIG. In this way, after the user triggers the automatic comment and posts a comment, the automatic comment can be realized through the expression material.

[0184] It can be seen that in an embodiment of the present application, when a trigger operation is received for a comment entry corresponding to the target information, at least one expression material related to the high-frequency comment words in the information flow is displayed in the comment editing area of ​​the target information, so that the user can publish the expression material as a comment without editing the text, thereby improving the efficiency of comment editing and publishing, and the expression material can vividly convey the user's subjective emotions towards the target information, enriching the expression form of the comment.

[0185] The following describes an exemplary application of the interactive processing method of information flow provided in an embodiment of the present application in a practical application scenario.

[0186] In information flow products, as more and more content is generated, users are engaging in more interactive behaviors within the information flow, including likes, reposts, and comments. Comments are the most revealing of users' thoughts and actions. When commenting, users can express their opinions through text or emoticons. In information flow products, nearly tens of millions of comments are generated daily, and many users express the same or similar opinions. To improve the efficiency of user expression and enhance the overall atmosphere of the comment editing area, it is necessary to guide user comments. Use legitimate and positive emoticons to encourage users to post positive comments, and use interesting emoticons to encourage users to comment actively. Therefore, it is necessary to identify hot topics in user comments, i.e., hot comments, and generate corresponding emoticons based on the specific content of the hot topics. These emoticons are posted in the comment editing area, guiding users to use these emoticons in their comments.

[0187] See also Figure 4 , Figure 4 It is a schematic diagram of the composition of the interactive processing system provided by the embodiment of the present application. The interactive processing system includes a comment hotspot mining module and a comment expression material generation module. Among them, the comment hotspot mining module and the comment expression material generation module are functional modules of the background server. The comment hotspot mining module mines all the comments posted on the day to find the central idea and corresponding comment content of the hot comments added on the day. The comment hotspot mining module consists of two submodules, namely the word frequency recognition algorithm submodule and the BERT vector clustering algorithm submodule. Combined with Figure 5The processing of the above modules is described. The word frequency recognition algorithm submodule segments the comments and uses n-gram to obtain all possible words corresponding to each comment. It calculates the word frequency of the words at a daily granularity, that is, it determines the word frequency of each word at a daily granularity, and determines the corresponding heat value based on the word frequency of each word. It obtains the TopN (N=10) hot words (i.e., keywords) ranked by heat value and the hot comments containing these hot words (i.e., the first hot comment). The BERT vector clustering algorithm submodule uses the pre-trained BERT model to vectorize all comments to obtain text vectors, and clusters the text vectors corresponding to all comments through the K-means algorithm, determines the cluster center, and counts the number of comments similar to each cluster center (i.e., Figure 5 ), determine the cluster centers whose number exceeds the upper limit of the number of comments (generally 2000), and use these cluster centers and the comments corresponding to similar text vectors as hot comments (i.e., the second hot comments). The hot comments determined by the word frequency recognition algorithm submodule and the BERT vector clustering algorithm submodule are taken as the union to obtain a hot comment set. The comment expression material generation module is used to process the hot comments in the hot comment set. It includes two submodules: the expression material generation submodule and the expression material publication submodule. The expression material generation submodule uses candidate expression materials to generate corresponding hot expression materials based on hot comments, and the expression material publication submodule adds the hot expression materials to the hot expression library in the comment editing area. When the comment that the user needs to publish belongs to a hot comment, the corresponding hot expression material can be selected for publication. In this way, the number of user comments can be greatly increased, and users are encouraged to use expression materials to publish relevant comments, thereby increasing the user's usage time and interaction rate of the terminal application.

[0188] like Figure 6 As shown, Figure 6 This is a schematic diagram of the expression material generation provided by the embodiment of this application. The distributed content (such as articles) will be presented in the information flow product, and the comments posted by users will be obtained, and the hot comments will be mined. The corresponding expression materials will be generated based on the hot comments. After that, the content under which users posted hot comments will be identified, and the expression materials corresponding to the hot comments will be published in the corresponding content comment editing area to promote more users to use and spread.

[0189] The modules mentioned above are described in detail below.

[0190] 1. Comment hotspot mining module

[0191] 1.1. Word frequency recognition algorithm submodule

[0192] The following will be passed Figure 7Steps 201 to 205 in the figure illustrate the process of determining hot comments by the word frequency recognition algorithm submodule.

[0193] Step 201 : Collect comments in a week at a daily granularity.

[0194] Step 202: perform word segmentation processing on the comments of each day of the week to obtain a word segmentation processing result corresponding to each comment.

[0195] For example, "I gave my deskmate a bouquet of flowers today", the result of word segmentation processing is: "I, gave, my classmate, a bouquet of flowers today".

[0196] Step 203: Determine the words corresponding to each comment in an n-gram manner.

[0197] Among them, n-gram combines the words in the word segmentation processing result to obtain multiple words, which can increase the number of candidate words. In the embodiment of the present application, the value of n is 1, 2, or 3. For example, if the word segmentation processing result is "I, today, gave, a classmate, a bouquet of, flowers", then the multiple words after n-gram combination can be "I, today, gave, a classmate, a bouquet of, flowers, I gave, gave a classmate, a classmate a bouquet of, a bouquet of flowers today, I gave, gave a classmate, a classmate a bouquet of, a classmate a bouquet of flowers today."

[0198] Step 204: Count the word frequencies of all the words corresponding to the comments within a week.

[0199] Among them, word frequency refers to the number of times each word appears in all comments.

[0200] Step 205: Calculate the popularity value of each word using a word frequency difference algorithm, sort the words in descending order based on the popularity value, take the top words as hot words, and take the comments including the hot words as hot comments.

[0201] The formula corresponding to the word frequency difference algorithm is shown in the formula (1) above. After sorting the words in descending order based on the popularity value, the top 10 words can be regarded as hot words.

[0202] 1.2. BERT Vector Clustering Algorithm Submodule

[0203] The BERT vector clustering algorithm submodule first obtains the BERT model through training. Figure 8 As shown, the BERT model is trained by unlabeled comment samples, where E1-E N Indicates input, T1-T N Indicates output, T rmDenote the intermediate vector. "×a" in the amplification structure indicates the existence of a repeated structures (multi-head attention layer, addition normalization layer, feed-forward layer, and addition normalization layer).

[0204] BERT pre-training is similar to word embedding. The network structure of the BERT model is predefined, and the model is trained using unlabeled corpora for specific tasks. To improve the performance of the BERT model in the comment task, the BERT model is pre-trained using 150 million comment corpora, and a special dictionary for comments is established. After tokenizing the comments, the whole-word masking method is used to incrementally train the BERT model to generate a BERT model suitable for the comment task.

[0205] Then, the comments are vectorized through the BERT model.

[0206] As Figure 9 shown, the BERT model consists of 12 layers of Transformer (Transformer0 - Transformer11). The input of the BERT model is a comment (text) of length L (L refers to the number of all characters in a comment). Among them, the output of each layer includes the vector of each word in the comment and the cls vector (W cls ) and the sep vector (W sep ). Therefore, the output of each layer has a total of (L + 2) vectors. The cls vector of Transformer0 in the 0th layer of the BERT model is concatenated with the cls vector of Transformer11 in the 11th layer, and the obtained text vector is used as the output of the BERT model.

[0207] After that, the text vectors corresponding to the comments are clustered through the K-means algorithm.

[0208] After clustering the text vectors through the K-means algorithm, multiple cluster centers are obtained. Calculate the distance between the text vector corresponding to each comment and the cluster center, and regard the comments corresponding to the text vectors with a distance within 20 from the cluster center as similar comments to the comments corresponding to the cluster center. Count the similar comments corresponding to each cluster center. When the number of similar comments exceeds the cluster threshold (such as 200), it is considered that the comments corresponding to the cluster center and the similar comments belong to the comment hot category, that is, the comments in the comment hot category are regarded as hot comments.

[0209] Among them, the method for calculating the distance between the text vector and the cluster center is the Euclidean distance. The distance between the obtained text vector of the comment and the corresponding cluster center is as Figure 10 shown. For example, the Euclidean distance between the text vector corresponding to the comment "Go for it!" and the cluster center is 36.36.

[0210] Finally, the hot comments determined by the word frequency recognition algorithm submodule and the hot comments determined by the BERT vector clustering algorithm submodule are taken as the union to obtain the total hot comments.

[0211] 2. Comment expression material generation module

[0212] After determining the hot comments, generate the corresponding hot expression materials based on the hot comments. For example Figure 11 The "Go Duck" expression material shown in .

[0213] like Figure 12 As shown, when a new comment is segmented and identified as a hot word, or when the text vector obtained by vectorizing the new comment is similar to each cluster center and the new comment is determined to belong to the hot comment category, the new comment is considered a hot comment. If the user's comment hits a hot topic, that is, hits a hot word, or the text vector corresponding to the comment is similar to a cluster center, the user can select the corresponding hot expression material from the comment editing area to post. At the same time, the posted hot expression material will be included in the next round of hot comment calculation as a new comment.

[0214] like Figure 13 As shown, "Go Duck" is a hot word. When a user wants to post a comment with the content "Go Duck", because the comment is a hot comment, the (hot) expression library in the comment editing area has a corresponding "Go Duck" expression material. At this time, the user can select the "Go Duck" expression material 1301 in the comment editing area to post as a comment.

[0215] Currently, by using the interactive information flow processing method provided in the embodiment of the present application, more than 10 to 20 hot comments can be discovered every week. It is estimated that 2 to 5 emoticon materials can be produced every week after going online, and the overall usage of emoticon materials is no less than 100,000 times / week, which can increase the average consumption time per person in the comment editing area by 5% and the user interaction rate by 9%.

[0216] The following describes an exemplary structure of the terminal described above. Figure 14 , Figure 14 is a schematic diagram of the structure of the terminal 400-1 provided in an embodiment of the present application. Figure 14 The terminal 400-1 shown includes: at least one processor 410, a memory 450, at least one network interface 420, and a user interface 430. The various components in the terminal 400-1 are coupled together via a bus system 440. It is understood that the bus system 440 is used to achieve connection and communication between these components. In addition to including a data bus, the bus system 440 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the bus system 440 is not described in detail. Figure 14 Various buses are labeled as bus system 440 .

[0217] The processor 410 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., where the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0218] The user interface 430 includes one or more output devices 431 that enable presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0219] The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard drives, optical drives, etc. The memory 450 may optionally include one or more storage devices that are physically remote from the processor 410.

[0220] The memory 450 includes volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 450 described in the embodiments of the present application is intended to include any suitable type of memory.

[0221] In some embodiments, the memory 450 can store data to support various operations, examples of which include programs, modules, and data structures, or a subset or superset thereof, as exemplified below.

[0222] The operating system 451 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic businesses and process hardware-based tasks.

[0223] The network communication module 452 is used to reach other computing devices via one or more (wired or wireless) network interfaces 420. Exemplary network interfaces 420 include Bluetooth, Wireless LAN (WiFi), and Universal Serial Bus (USB).

[0224] The presentation module 453 is configured to enable presentation of information (eg, a user interface for operating peripheral devices and displaying content and information) via one or more output devices 431 (eg, a display screen, a speaker, etc.) associated with the user interface 430 .

[0225] The input processing module 454 is configured to detect one or more user inputs or interactions from one of the one or more input devices 432 and to translate the detected inputs or interactions.

[0226] In some embodiments, the interactive processing device for information flow provided in the embodiments of the present application can be implemented in software. Figure 14 An interactive processing device 455 of the information flow stored in the memory 450 is shown, which can be software in the form of a program and a plug-in, etc., including an acquisition module 4551 and a display module 4552.

[0227] The functions of each module will be described below.

[0228] An acquisition module 4551 is used to acquire an information flow, wherein the information flow includes multiple pieces of information; a display module 4552 is used to display at least one piece of information in the information flow, and a comment entry corresponding to the at least one piece of information; in response to a trigger operation on the comment entry corresponding to the target information, at least one expression material is displayed in the comment editing area of ​​the target information; wherein the target information is information to be commented on in the information flow, the at least one expression material is related to high-frequency comment words in the information flow, and the comment editing area is used to publish comments corresponding to the target information based on the at least one expression material.

[0229] In some embodiments, high-frequency comment words in the information flow include: keywords in the hot comments of the target information, and the heat value of the keywords is greater than the heat threshold; the display module 4552 is also used to display multiple expression materials related to the keywords in the hot comments of the target information in order in the comment editing area of ​​the target information; wherein, the sorting method of the multiple expression materials includes: the heat value of the keywords corresponding to the expression materials in multiple historical comments of the target information; the frequency of use of the expression materials in editing comments.

[0230] In some embodiments, high-frequency comment words in the information flow include: keywords in hot comments related to the target information, and the heat value of the keywords is greater than the heat threshold; the display module 4552 is also used to respond to the comment editing operation and obtain the text entered in the comment editing area; when the text includes keywords in the hot comments, it is determined that at least one expression material related to the keywords in the comments posted by the interactive account will be displayed in the comment editing area.

[0231] In some embodiments, at least one expression material related to a comment on the target information includes: at least one expression material related to a keyword in a comment posted by an interactive account, wherein the interactive account is an account that has an interactive relationship with the login account, and the login account is an account used to obtain and display information flow; the display module 4552 is also used to obtain text entered in the comment editing area in response to a comment editing operation; when it is recognized that the text includes a keyword in a comment posted by an interactive account, it is determined that at least one expression material related to the keyword will be displayed in the comment editing area.

[0232] In some embodiments, the display module 4552 is also used to display multiple expression materials corresponding to multiple keywords in the comment editing area in order before identifying the keywords included in the text, wherein the sorting method of the multiple expression materials includes: the popularity value of the keywords corresponding to the expression materials in multiple historical comments of the target information; the frequency of use of the expression materials in editing comments; after identifying the keywords included in the text, highlighting the expression materials corresponding to the keywords in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

[0233] In some embodiments, the display module 4552 is also used to display a plurality of preset expression materials in order in the comment editing area before identifying the keywords included in the text, wherein the sorting method of the plurality of preset expression materials includes: the frequency of use of the expression materials in editing comments; after identifying the keywords included in the text, inserting and highlighting the expression materials corresponding to the keywords in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

[0234] In some embodiments, the display module 4552 is also used to update the comment editing area displaying the target information in real time based on at least one of the expression materials corresponding to the keywords in the comments of the target information in the display state and the expression materials corresponding to the keywords in the comments of the logged-in account participating in the interaction; wherein the logged-in account is an account used to obtain and display the information flow.

[0235] In some embodiments, the display module 4552 is also used to display at least part of the content of at least one piece of information in the information flow, and in response to an information selection operation, display the detailed content of the selected target information, at least part of the comments on the target information, and the comment entry corresponding to the target information; or display at least part of the content of at least one piece of information in the information flow and the comment entry corresponding to each piece of information.

[0236] In some embodiments, the display module 4552 is also used to determine the activity level of the logged-in account in the information flow, where the logged-in account is the account used to obtain and display the information flow; and to display emoticon materials whose number is positively correlated with the activity level in the comment editing area of ​​the target information.

[0237] In some embodiments, the display module 4552 is also used to respond to a comment posting operation of a logged-in account, wherein the logged-in account is an account used to obtain and display the information stream, and the text entered by the logged-in account in the comment editing area and the emoticon material selected in at least one emoticon material are combined into a comment, and sent to the server that publishes the information stream based on the identity of the logged-in account, so that the server is synchronized to other accounts that obtain the information stream.

[0238] In some embodiments, the display module 4552 is also used to respond to a comment posting operation of a logged-in account, wherein the logged-in account is an account used to obtain and display the information flow, and the text entered by the logged-in account in the comment editing area is sent as a comment to the server that publishes the information flow based on the identity of the logged-in account, so that the server is synchronized to other accounts that obtain the information flow; at least one emoticon material automatically selected from at least one emoticon material is sent as a comment to the server that publishes the information flow based on the identity of the logged-in account or the identity of the comment robot, so that the server is synchronized to other accounts that obtain the information flow.

[0239] In some embodiments, the display module 4552 is also used to respond to the comment posting operation of the automatic follow-up target comment of the login account, wherein the login account is the account used to obtain and display the information flow, the target comment is the comment of the target information in the display state, and at least one emoticon material selected from at least one emoticon material related to the target comment is used as a comment, and is sent to the server that publishes the information flow based on the identity of the login account or the identity of the comment robot, so that the server is synchronized to other accounts that obtain the information flow.

[0240] In some embodiments, the display module 4552 is also used to obtain at least one expression material related to the comments of the target information in the following manner: obtaining multiple comments on the target information; determining the popularity value corresponding to the words in each comment; using words with popularity values ​​higher than the popularity threshold as keywords, and using the comments including the keywords as the first hot comment; generating corresponding expression materials based on the first hot comment.

[0241] In some embodiments, the display module 4552 is also used to determine the text vector of the comment corresponding to each of the multiple messages in the information flow; cluster the text vectors corresponding to the comments of the multiple messages to obtain at least one cluster center; determine the number of text vectors similar to each cluster center, and use the cluster centers whose number exceeds the clustering threshold as hot cluster centers; determine at least one text vector similar to the hot cluster center among the text vectors corresponding to the comments of the multiple messages, and use the comment of the information corresponding to the at least one text vector as the second hot comment; use the union of the first hot comment and the second hot comment as the hot comment set; and generate expression materials corresponding to each comment in the hot comment set.

[0242] In some embodiments, the display module 4552 is also used to obtain multiple comments corresponding to the target information published during the sampling period; when the number of multiple comments is less than a quantity threshold, where the quantity threshold is the minimum value for identifying hot comments, multiple comments corresponding to the quantity threshold in the target information are obtained in order from new to old in terms of publication time; or, based on at least one dimension, an associated account of the corresponding login account is determined, where the login account is an account used to obtain and display information flow, and the dimension includes at least one of the following: social distance from the login account, geographical distance from the login account; and multiple comments published by the associated account on the target information are obtained.

[0243] In some embodiments, the display module 4552 is also used to determine text materials based on the first hot comment, wherein the text materials include at least one of the following: the first hot comment, keywords in the first hot comment, and other texts with the same semantics as the first hot comment; obtaining image materials that match the emotional characteristics of the first hot comment from multiple candidate material images, or fusing the candidate material images with the emotional characteristics of the first hot comment to obtain image materials; and combining the text materials and the image materials into expression materials.

[0244] The present invention provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the interactive processing method of information flows described in the present invention.

[0245] The embodiment of the present application provides a computer-readable storage medium storing executable instructions, wherein the executable instructions are stored. When the executable instructions are executed by a processor, the processor will execute the interactive processing method of the information flow provided by the embodiment of the present application, for example, Figure 2A An interactive processing method for information flow is shown.

[0246] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface storage, optical disk, or CD-ROM; or various devices including one or any combination of the above memories.

[0247] In some embodiments, executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0248] As an example, executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file that stores other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinating files (e.g., files storing one or more modules, subroutines, or code portions).

[0249] By way of example, executable instructions may be deployed to be executed on one computing device, or on multiple computing devices at one site, or on multiple computing devices distributed across multiple sites and interconnected by a communication network.

[0250] To sum up, in an embodiment of the present application, when a trigger operation is received for a comment entry corresponding to the target information, at least one emoticon material related to the high-frequency comment words in the information flow is displayed in the comment editing area of ​​the target information, so that the user can publish the emoticon material as a comment without editing the text, thereby improving the efficiency of comment editing and publishing, and the emoticon material can vividly convey the user's subjective emotions towards the target information, enriching the expression form of the comment.

[0251] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. Any modifications, equivalent replacements, and improvements made within the spirit and scope of the present application are included in the scope of protection of the present application.

Claims

1. A method for interactive processing of information flow, characterized in that: The method comprises: Displaying at least one piece of information in the information stream and a comment entry corresponding to the at least one piece of information; In response to a triggering operation on a comment entry corresponding to the target information, a plurality of expression materials related to keywords in the hot comments of the target information are displayed in sequence in a comment editing area of ​​the target information; Among them, the target information is the information to be commented on in the information flow, the heat value of the keyword in the hot comments of the target information is greater than the heat threshold, and the comment editing area is used to publish comments corresponding to the target information based on the at least one expression material; the sorting method of the multiple expression materials includes: the heat value of the keyword corresponding to the expression material in multiple historical comments on the target information; the frequency of use of the expression material in editing comments.

2. The method according to claim 1, characterized in that Before sequentially displaying a plurality of expression materials related to keywords in the hot comments of the target information in the comment editing area of ​​the target information, the method further includes: In response to a comment editing operation, obtaining text entered in the comment editing area; When the text includes a keyword in a hot comment of the target information, it is determined that the plurality of expression materials will be displayed in the comment editing area.

3. The method according to claim 2, characterized in that The method further comprises: Before identifying the keywords included in the text, a plurality of emoticon materials corresponding to the plurality of keywords are sequentially displayed in the comment editing area, wherein the order of the plurality of emoticon materials includes: the popularity value of the keywords corresponding to the emoticon materials in a plurality of historical comments on the target information; the frequency of use of the emoticon materials in editing comments; After identifying the keywords included in the text, the expression materials corresponding to the keywords are highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

4. The method according to claim 2, characterized in that The method further comprises: Before identifying the keywords included in the text, a plurality of preset expression materials are displayed in the comment editing area in order, wherein the order of the plurality of preset expression materials includes: the frequency of use of the expression materials in editing comments; After identifying the keywords included in the text, the expression materials corresponding to the keywords are inserted and highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

5. The method according to claim 1, wherein The method further comprises: The comment editing area of ​​the target information is updated in real time based on at least one of the expression materials corresponding to the keywords in the comments in the display state of the target information and the expression materials corresponding to the keywords in the comments of the logged-in account participating in the interaction; wherein the logged-in account is the account used to obtain and display the information flow.

6. The method according to any one of claims 1 to 2 and 5, characterized in that The display of at least one piece of information in the information flow and a comment entry corresponding to the at least one piece of information includes: Display at least part of the content of at least one piece of information in the information stream, and in response to an information selection operation, display the details of the selected target information, at least part of the comments on the target information, and a comment entry corresponding to the target information; or At least a portion of the content of at least one piece of information in the information stream and a comment entry corresponding to each piece of information are displayed.

7. The method according to any one of claims 1 to 2 and 5, characterized in that The method further comprises: Determining the activity level of a login account in the information stream, wherein the login account is an account used to obtain and display the information stream; Displaying expression materials in a comment editing area of ​​the target information in a number positively correlated with the activity level.

8. The method according to any one of claims 1 to 2 and 5, characterized in that The method further comprises: In response to a comment posting operation of a login account, wherein the login account is an account used to obtain and display the information flow, the text entered by the login account in the comment editing area and the expression material selected from the at least one expression material are combined into a comment, and The information is sent to the server that publishes the information stream based on the identity of the login account, so that the server is synchronized with other accounts that obtain the information stream.

9. The method according to any one of claims 1 to 2 and 5, characterized in that The method further comprises: In response to a comment posting operation by a logged-in account, wherein the logged-in account is an account used to obtain and display the information stream, the text entered by the logged-in account in the comment editing area is sent as a comment to the server that publishes the information stream based on the identity of the logged-in account, so that the server is synchronized with other accounts that obtain the information stream; At least one emoticon material automatically selected from the at least one emoticon material is sent as a comment to the server that publishes the information stream based on the identity of the login account or the identity of the comment robot, so that the server is synchronized with other accounts that obtain the information stream.

10. The method according to any one of claims 1 to 2 and 5, characterized in that: The method further comprises: In response to the comment posting operation of the automatic comment target comment of the login account, wherein the login account is the account used to obtain and display the information flow, the target comment is the comment of the target information in the display state, at least one expression material selected from the at least one expression material related to the target comment is used as a comment, and The identity of the login account or the identity of the comment robot is sent to the server that publishes the information stream, so that the server is synchronized with other accounts that obtain the information stream.

11. The method according to any one of claims 1 to 2 and 5, characterized in that: Before sequentially displaying a plurality of expression materials related to keywords in the hot comments of the target information in the comment editing area of ​​the target information, the method further includes: Acquire at least one expression material related to the comment of the target information in the following manner: Obtaining multiple comments on the target information; Determine the popularity value corresponding to each word in the comment; The words with a popularity value higher than the popularity threshold are regarded as keywords, and the comments containing the keywords are regarded as the first hot comments; Generate corresponding emoticon materials based on the first hot comment.

12. The method according to claim 11, characterized in that Generating corresponding expression materials based on the first hot comment includes: Determine a text vector of a comment corresponding to each of the multiple messages in the information stream; performing clustering processing on the text vectors corresponding to the comments of the plurality of pieces of information to obtain at least one cluster center; Determine the number of text vectors similar to each cluster center, and use the cluster center whose number exceeds the clustering threshold as the hotspot cluster center; Determining at least one text vector similar to the hotspot cluster center among the text vectors corresponding to the comments of the plurality of pieces of information, and using the comment of the information corresponding to the at least one text vector as a second hotspot comment; Taking the union of the first hot comments and the second hot comments as a hot comments set; Generate an expression material corresponding to each comment in the hot comment set.

13. The method according to claim 11, characterized in that The obtaining of multiple comments on the target information includes: Acquire multiple comments corresponding to the target information published during a sampling period; when the number of the multiple comments is less than a quantity threshold, wherein the quantity threshold is a minimum value for identifying hot comments, acquire multiple comments corresponding to the quantity threshold in the target information in descending order of publication time; or, Determine an associated account corresponding to a login account based on at least one dimension, wherein the login account is an account used to obtain and display the information flow, and the dimension includes at least one of the following: a social distance from the login account, a geographical distance from the login account; and obtain a plurality of comments published by the associated account on the target information.

14. The method according to claim 11, characterized in that Generating corresponding expression materials based on the first hot comment includes: Determining text material based on the first hot comment, wherein the text material includes at least one of the following: the first hot comment, keywords in the first hot comment, and other text with the same semantics as the first hot comment; Acquire an image material that matches the sentiment feature of the first hot comment from a plurality of candidate material images, or fuse the candidate material image with the sentiment feature of the first hot comment to obtain the image material; The text material and the image material are combined into an expression material.

15. An interactive processing device for information flow, characterized in that: include: A display module, configured to display at least one piece of information in an information stream and a comment entry corresponding to the at least one piece of information; In response to a triggering operation on a comment entry corresponding to target information, a plurality of expression materials related to keywords in hot comments of the target information are displayed in sequence in the comment editing area of ​​the target information; wherein, the target information is information to be commented on in the information flow, and the heat value of the keyword in the hot comments of the target information is greater than a heat threshold, and the comment editing area is used to publish a comment corresponding to the target information based on the at least one expression material; the sorting method of the plurality of expression materials includes: the heat value of the keyword corresponding to the expression material in multiple historical comments of the target information; the frequency of use of the expression material in editing comments.

16. The device according to claim 15, characterized in that The display module is further used for: In response to a comment editing operation, obtaining text entered in the comment editing area; When the text includes a keyword in a hot comment of the target information, it is determined that the plurality of expression materials will be displayed in the comment editing area.

17. The device according to claim 16, characterized in that The display module is further used for: Before identifying the keywords included in the text, a plurality of emoticon materials corresponding to the plurality of keywords are sequentially displayed in the comment editing area, wherein the order of the plurality of emoticon materials includes: the popularity value of the keywords corresponding to the emoticon materials in a plurality of historical comments on the target information; the frequency of use of the emoticon materials in editing comments; After identifying the keywords included in the text, the expression materials corresponding to the keywords are highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

18. The device according to claim 16, characterized in that The display module is further used for: Before identifying the keywords included in the text, a plurality of preset expression materials are displayed in the comment editing area in order, wherein the order of the plurality of preset expression materials includes: the frequency of use of the expression materials in editing comments; After identifying the keywords included in the text, the expression materials corresponding to the keywords are inserted and highlighted in the comment editing area, wherein the highlighting method includes at least one of the following: highlighting, enlarged display, text prompt, voice prompt, and sorting in front.

19. An electronic device, characterized in that: include: a memory for storing executable instructions; A processor is configured to implement the information flow interactive processing method according to any one of claims 1 to 14 when executing the executable instructions stored in the memory.

20. A computer-readable storage medium, characterized in that Executable instructions are stored for implementing the interactive processing method of information flow according to any one of claims 1 to 14 when executed by a processor.

21. A computer program product, characterized in that The computer program product includes computer instructions for being executed by a processor to implement the interactive processing method of information flow according to any one of claims 1 to 14.

Images