Summary information generation method and device, medium and product
By preset filtering of group chat unread information and contextual understanding of large language models, combined with non-text information analysis, accurate and comprehensive group chat summary information is generated, solving the problems of information overload and key information omission in the existing technology.
Patent Information
- Application Number
- CN202411140697.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-22
AI Technical Summary
In large-scale group chats, it is difficult for users to effectively filter and obtain key information. The summary information generated by the existing technology based on keyword extraction or simple statement reorganization has problems such as inaccurate or incompleteness, and it is especially impossible to effectively process non-text information such as pictures and sounds. Important information in the video.
Useless information in group chat unread information through preset filtering conditions, directly use the large language model to understand the text information in context, and analyze and convert non-text information such as pictures, audio and video and file information into text information, combined with Text information is contextually understood to generate accurate and comprehensive summary information.
It realizes accurate and comprehensive summary information generation of group chat unread information, avoids the omission of important information, and improves the accuracy and completeness of information screening.
Smart Images

Figure CN119003758A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a summary information generation method, medium and product. Background Art
[0002] With the popularity of social networks and instant messaging applications, group chat has become one of the most commonly used communication methods in people's daily life and work. However, with the increase in the number of group chat participants and the rapid growth of information in group chats, users are facing a serious information overload problem. In large-scale group chats, users often find it difficult to effectively filter and obtain key information.
[0003] In related schemes, in order to filter out key information from massive amounts of unread group chat information, users are usually required to provide keywords, and based on the keywords provided by the user, group chat information that can match the keywords is filtered out from the group chat messages, and then a summary of the group chat information that can match the keywords is provided to the user; or based on the keywords provided by the user, a simple sentence reorganization process is performed on the group chat information that can match the keywords to generate a summary of the unread group chat information and provide it to the user.
[0004] However, the above solutions rely on keyword extraction technology or simple sentence reorganization technology, and the summary information generated based on unread group chat information is inaccurate or incomplete. Summary of the invention
[0005] Embodiments of the present invention provide a summary information generation method, device, medium and product to generate accurate and comprehensive summary information based on unread group chat information, and can also provide automatic question and answer information related to the summary information to meet the user's personalized needs for unread group chat information.
[0006] In a first aspect, an embodiment of the present invention provides a summary generation method, the method comprising:
[0007] Obtain unread information of a target group chat, where the unread information of the target group chat includes multiple first unread messages;
[0008] Filtering the first unread messages that meet the preset filtering conditions in the target group chat unread messages to obtain a plurality of second unread messages for information analysis;
[0009] Analyze the non-text information in the plurality of second unread messages to obtain text information corresponding to each non-text information; wherein the non-text information includes any one or more of picture information, audio and video information, and file information;
[0010] The text information corresponding to the non-text information and the text information in the plurality of second unread messages are contextually understood through a large language model to obtain summary information of the unread messages of the target group chat.
[0011] In a second aspect, an embodiment of the present invention provides a summary information generation method, comprising:
[0012] Receive a first input from a user on a group chat page;
[0013] In response to the first input, display summary information of unread group chat messages in the group chat corresponding to the group chat page;
[0014] Among them, the summary information of the group chat unread information is obtained by filtering the first unread information that meets the preset conditions in the group chat unread information to obtain multiple second unread information for information analysis, analyzing the non-text information in the multiple second unread information to obtain text information corresponding to each non-text information, and using a large language model to understand the context of the text information corresponding to the non-text information and the text information in the multiple second unread information.
[0015] In a third aspect, an embodiment of the present invention provides a summary generating device, the device comprising:
[0016] An acquisition module, used for acquiring unread information of a target group chat, wherein the unread information of the target group chat includes a plurality of first unread information;
[0017] A filtering module, used to filter the first unread messages that meet the preset filtering conditions in the target group chat unread messages, and obtain a plurality of second unread messages for information analysis;
[0018] An analysis module, configured to analyze the non-text information in the plurality of second unread messages to obtain text information corresponding to each non-text information; wherein the non-text information includes any one or more of picture information, audio and video information, and file information;
[0019] The understanding module is used to perform context understanding on the text information corresponding to the non-text information and the text information in the plurality of second unread messages through a large language model to obtain summary information of the unread messages of the target group chat.
[0020] In a fourth aspect, an embodiment of the present invention provides a summary generating device, the device comprising:
[0021] A receiving module, used to receive a first input from a user on a group chat page;
[0022] A response module, configured to display summary information of unread group chat messages in a group chat corresponding to the group chat page in response to the first input;
[0023] Among them, the summary information of the group chat unread information is obtained by filtering the first unread information that meets the preset conditions in the group chat unread information to obtain multiple second unread information for information analysis, analyzing the non-text information in the multiple second unread information to obtain text information corresponding to each non-text information, and using a large language model to understand the context of the text information corresponding to the non-text information and the text information in the multiple second unread information.
[0024] In a fifth aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, a processor, and a communication interface; wherein executable code is stored on the memory, and when the executable code is executed by the processor, the processor can at least implement the summary generation method of the above aspects.
[0025] In a sixth aspect, an embodiment of the present invention provides a non-temporary machine-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor can at least implement the summary generation method of the above aspects.
[0026] In a seventh aspect, an embodiment of the present invention provides a computer program product, which includes a computer program. When the computer program product is executed by a processor of an electronic device, the processor can at least implement the summary generation method of the above aspects.
[0027] In the solution provided by the embodiment of the present invention, target group chat unread information including multiple first unread information is obtained, and useless information in the target group chat unread information, such as emoticons, gossip, and repeated information, is filtered out by preset filtering conditions to obtain multiple second unread information for information analysis. The text information in the multiple second unread information does not need to be processed in any way, and can be directly sent to the large language model for understanding. The non-text information in the multiple second unread information, such as picture information, audio and video information, and file information, is analyzed to obtain text information corresponding to each non-text information. By analyzing the picture information, audio and video information, and file information in the second unread information, it is possible to avoid missing important information. Afterwards, the text information corresponding to the non-text information and the text information in the multiple second unread information are contextually understood by the large language model to obtain summary information of the target group chat unread information. Thus, accurate and comprehensive summary information corresponding to the target group chat unread information can be obtained by understanding the text information and non-text information in the target group chat unread information. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0029] Figure 1 A flowchart of a summary generation method provided by an embodiment of the present invention;
[0030] Figure 2a to Figure 2b A schematic diagram of a user triggering an operation to obtain target group chat information;
[0031] Figure 3 A flowchart of another summary generation method provided by an embodiment of the present invention;
[0032] Figure 4 A flowchart of another method for generating summary information provided by an embodiment of the present invention;
[0033] Figure 5 A flowchart of another summary generation method provided by an embodiment of the present invention;
[0034] Figure 6 A schematic diagram of the structure of a summary generating device provided by an embodiment of the present invention;
[0035] Figure 7 A schematic diagram of the structure of another summary generating device provided by an embodiment of the present invention;
[0036] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0037] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0038] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.
[0039] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the case where there is no conflict between the embodiments, the following embodiments and the features in the embodiments can be combined with each other. In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.
[0040] First, some terms involved in the embodiments of the present invention are explained.
[0041] Large Language Model (LLM): usually refers to a deep learning model trained based on massive text data. It can handle a variety of natural language tasks, such as filtering useless information in unread messages in group chats and understanding the context of text information.
[0042] Image analysis model: used to analyze key elements in image information to convert the key elements contained in the image information into corresponding text information.
[0043] Speech analysis model: used to analyze the audio content in speech and video to convert the speech content in speech and video into corresponding text information.
[0044] File analysis model: used to analyze the content contained in the file information to convert the content in the file information into corresponding text information.
[0045] With the rapid development of social networks and the widespread use of instant messaging applications, group chat has become one of the indispensable ways of communication in people's daily life and work. Whether it is chatting among friends, arranging gatherings for family members, or collaborating on tasks in work teams, group chat provides a convenient communication platform. However, as the scale of group chat grows, information overload has become a problem that cannot be ignored. When the number of group chat members increases, the amount of information in the chat history accumulates rapidly, making it difficult for users to follow up on every message in a timely manner, let alone filter out valuable content from it.
[0046] The current solution to this problem is usually to help users quickly locate important information through keyword extraction technology and simple sentence reorganization technology. For example, users can set specific keywords, and the system will automatically filter messages containing these keywords and generate a brief summary for users to view. Although this method has improved the efficiency of information screening to a certain extent, it still has some limitations: for example, keyword extraction technology may not accurately capture the topics that users really care about, and sometimes important information is not directly reflected in the keywords, but is implied in the context. Therefore, keyword extraction may cause key information to be omitted and the accuracy is not high. Simple sentence reorganization may lose the integrity and context of the original information, and the generated summary often lacks logical coherence and is difficult to fully express the meaning of the original information.
[0047] In addition, group chat messages usually contain some picture information and audio and video information. Sometimes, important information is contained in the picture or audio and video. The above solution can only filter the text in the group chat messages, and cannot identify the content contained in the picture information and audio and video information. This may cause important information to be omitted in the generated summary information.
[0048] Based on this, the embodiment of the present invention provides a new summary generation method, which can first filter out useless information in the unread information of the target group chat through preset filtering conditions, such as emoticons, gossip and repeated information, and then directly understand the text information in the unread information of the filtered target group chat through a large language model, and analyze and identify the non-text information in the unread information of the filtered target group chat, such as picture information, audio and video information and file information, so as to convert the picture information, audio and video information and file information into corresponding text information. After that, the text information is contextually understood through the large language model, and accurate and comprehensive summary information corresponding to the unread information of the target group chat can be obtained.
[0049] The following is an introduction to the summary generation method and device provided by the embodiments of the present invention.
[0050] Figure 1 A flowchart of a summary generation method provided by an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:
[0051] 101. Obtain unread information of a target group chat, where the unread information of the target group chat includes a plurality of first unread messages.
[0052] 102. Filter the first unread messages that meet the preset filtering condition in the target group chat unread messages to obtain a plurality of second unread messages for information analysis.
[0053] 103. Analyze the non-text information in the plurality of second unread messages to obtain text information corresponding to each piece of non-text information; wherein the non-text information includes any one or more of picture information, audio and video information, and file information.
[0054] 104. Perform context understanding on the text information corresponding to the non-text information and the text information in the plurality of second unread messages through a large language model to obtain summary information of the unread messages in the target group chat.
[0055] The operation of obtaining the unread information of the target group chat may be actively triggered by the user, or may be triggered by the electronic device according to a preset trigger condition.
[0056] Specifically, when the user actively triggers the acquisition of unread information of the target group chat, the user first locates the target group chat that the user wants to view after opening the instant messaging software, and after entering the target group chat, a prompt message indicating whether to display the summary of the unread information is displayed on the information display interface of the group chat. If the user selects the need to display the summary of the unread information in response to the prompt message, the electronic device acquires the unread information of the target group chat including the plurality of first unread messages.
[0057] For example, Figure 2a As shown, after opening the instant messaging software, the user enters the main interface of the instant messaging software. Assuming that there are three group chats A, B and C, which include unread messages, as well as contacts a and b, on the main interface, the user wants to view group chat C, that is, group chat C is the user's target group chat, and the user clicks group chat C to enter the information display interface of group chat C. After entering the information display interface of group chat C, a prompt message will be displayed on the information display interface, and the prompt message is used to ask the user whether to display the summary information corresponding to the unread messages in group chat C, such as Figure 2b If the user selects yes, the electronic device is triggered to obtain the target group chat unread information including the plurality of first unread information.
[0058] When the electronic device triggers the acquisition of the target group chat unread information according to the preset trigger condition, after the electronic device detects that the user opens the instant messaging software and enters the target group chat that the electronic device wants to view, if the electronic device determines that the number of unread messages in the target group chat exceeds the preset value, the process of displaying the summary information of the unread messages of the target group chat is triggered, that is, the electronic device obtains the target group chat unread information including multiple first unread messages. On the contrary, if the electronic device determines that the number of unread messages in the target group chat does not exceed the preset value, the process of displaying the summary information of the unread messages of the target group chat is not triggered.
[0059] Afterwards, useless information in the multiple first unread messages in the target group chat unread messages needs to be filtered. Specifically, useless information in the multiple first unread messages can be filtered according to preset filtering conditions, wherein useless information generally includes emoticons, idle chats, repeated information, etc. In this embodiment, a large language model for implementing useless information filtering can be pre-trained to implement the filtering of useless information in the multiple first unread messages based on the contextual understanding of the multiple first unread messages.
[0060] When training the large language model for filtering useless information, first obtain a training data set, which includes training sample group chat information and expected filtering results corresponding to the training sample group chat information. The large language model processes the training sample group chat information to obtain a training filtering result, and the training filtering result is compared with the expected filtering result. The training parameters of the large language model are continuously adjusted according to the comparison result until the training filtering result of the large language model reaches the expected filtering result. More test group chat information can also be used to test the trained large language model to obtain a more ideal filtering result, so as to enhance the generalization ability of the large language model.
[0061] Assume that the target group chat unread messages are: [
[0063] {"id":"1","text":"@Everyone, there are cakes in the xxx meeting room on the 7th floor. If you need them, you can pick them up. First come, first served.
[0064] [Sly smile][Sly smile]","type":"text"},
[0065] {"id":"2","text":"Happy Magic Day to all students","type":"text"},
[0066] {"id":"3","text":"[strong]","type":"text"},
[0067] {"id":"4","text":"Happy Magic Day","type":"text"},
[0068] {"id":"5","text":"Happy Magic Day [Fireworks][Fireworks][Fireworks]","type":"text"},
[0069] {"id":"6","text":"Happy Magic Day [Fireworks][Fireworks][Fireworks]","type":"text"},
[0070] {"id":"7","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0071] {"id":"8","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0072] {"id":"9","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0073] {"id":"10","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0074] {"id":"11","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0075] {"id":"12","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0076] {"id":"13","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0077] {"id":"14","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0078] {"id":"15","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},
[0079] {"id":"16","text":"Happy Magic Day [Fireworks][Firecrackers][Fireworks]","type":"text"},]
[0080] The unread messages of the target group chat are filtered by the large language model for filtering useless messages, and the plurality of second unread messages for information analysis are obtained by removing emoticons and repeated messages: [
[0082] {"id":"1","text":"@Everyone, there are cakes in the xxx meeting room on the 7th floor. If you need them, you can pick them up. First come first served","type":"text"},
[0083] {"id":"2","text":"Happy Magic Day to all students","type":"text"},
[0084] {"id":"4","text":"Happy Magic Day","type":"text"} ]
[0086] It can be seen that by using the large language model for filtering useless information, a relatively ideal filtering effect can be achieved.
[0087] Next, it is necessary to analyze and identify the multimodal information in the plurality of second unread messages, and uniformly process the multimodal information into information in text format, wherein the multimodal information refers to non-text information, such as picture information, audio information, video information, file information, etc.
[0088] Specifically, it is possible to first determine whether the target unread information is text information or non-text information based on the mark of the target unread information. For the target unread information determined to be text information, no processing is required here, and the target unread information can be directly sent to the preset information parsing cache. For the target unread information determined to be non-text information, it is necessary to analyze the target unread information through the analysis model corresponding to the message type of the target unread information, obtain the text information corresponding to the target unread information, and send the text information corresponding to the target unread information to the preset information parsing cache; wherein the target unread information is any one of the multiple second unread information.
[0089] Specifically, for the target unread information whose message type is picture information, the key elements in the target unread information, such as text, objects, emotions, etc. in the picture, are analyzed through the picture analysis model to obtain text information corresponding to the key elements, that is, the elements such as text, objects and emotions in the picture are expressed in the form of text information.
[0090] For target unread information whose message type is audio or video information, the voice in the target unread information is analyzed by a voice analysis model to obtain text information corresponding to the voice, that is, the voice content is converted into the form of text information.
[0091] For target unread information whose message type is file information, the text content in the target unread information is analyzed by the file analysis model to obtain text information corresponding to the text content. Specifically, the text content in the file can be identified by the file analysis model, and then the text content can be converted into the form of text information.
[0092] The above lists several common non-text information message types and methods of obtaining text information corresponding to these message types of non-text information. For other message types of non-text information not listed, they can also be converted into text information by identifying the content of the non-text message.
[0093] In addition, in order to improve the response speed, after determining that the target unread information is non-text information, you can also first query whether there is an analysis result corresponding to the target unread information in the preset information analysis cache. If there is no analysis result corresponding to the target unread information in the preset information analysis cache, the target unread information is analyzed by the analysis model corresponding to the message type of the target unread information. If there is an analysis result corresponding to the target unread information in the preset information analysis cache, the analysis result is determined as the text information corresponding to the target unread information. In this way, while improving the response speed, unnecessary computing overhead can also be reduced.
[0094] After converting the non-text information in the plurality of second unread messages into text information, the large language model is used to perform contextual understanding on the text information corresponding to the non-text information and the text information in the plurality of second unread messages. Specifically, the text information corresponding to the target unread information and the text information in the plurality of second unread messages can be obtained from the preset information parsing cache library, and the large language model is used to perform contextual understanding on the text information to accurately capture the main content and dynamics of the target group chat unread information, so as to generate accurate and complete summary information.
[0095] The large language model for contextual understanding of text information is pre-trained. Specifically, during training, a training data set is first obtained. The training sample data in the training data set is data that has been processed by filtering useless information and converting non-text information into text information, including training sample text information and expected summary information corresponding to the training sample text information. For example, the training sample text information (inputs) and the expected summary information (output) corresponding to the training sample text information can be:
[0096] {
[0097] "inputs":[{"id":"1","text":"@Everyone, there are cakes in the xxx meeting room on the 7th floor. If you need them, you can pick them up.
[0098] First come first served","type":"text"},
[0099] {"id":"2","text":"Happy Magic Day to all students","type":"text"},
[0100] {"id":"3","text":"58.com's magical day, with 560 million exposures, appeared nationwide, from the top to the bottom, you must see 58.com,
[0101] Let every trip show the beauty of life","type":"text"},
[0102] {"id":"4","text":"Audio and video identification content: 58.com, your partner for a wonderful life. Choose a good partner, live a good life, find a job, find a house, find life services, big or small things in life, there is always TA58 to help you every day",
[0103] "type":"text"},
[0104] {"id":"5","text":"@Everyone, please remember to set your alarm","type":"text"},
[0105] …
[0106] ],
[0107] "output":[{"index":"1","text":"Happy Magic Day to all students. There are cakes in the xxx conference room on the 7th floor. If you need them, you can pick them up. First come, first served"},
[0108] {"index":"2","text":"58.com's magical day, debuting nationwide with 560 million impressions"},
[0109] {"index":"3","text":"Remember to set the alarm"},
[0110] ……]
[0111] }
[0112] The training sample text information is summarized by the large language model to obtain training summary information, and the training summary information is compared with the expected summary information. The training parameters of the large language model are continuously adjusted according to the comparison results until the training summary information output by the large language model reaches the expected summary information. Furthermore, more test text information can be used to test the trained large language model to obtain more ideal training summary information, so as to enhance the generalization ability of the large language model.
[0113] Taking the community owner group as the target group chat as an example, assuming that the multiple second unread messages in the community owner group include:
[0114] [{"id":"1","text":"Neighbors, who bought kitchen gas accessories, gas pipes, etc?",
[0115] "type":"text"},
[0116] {"id":"2","text":"Does anyone in my neighborhood have latex paint? I'd like to see what color looks good?",
[0117] "type":"text"},
[0118] {"id":"7","text":"I painted the colorful white","type":"text"},
[0119] {"id":"5","text":"I also painted a colorful white","type":"text"},
[0120] {"id":"4","text":"I have bought it","type":"text"},
[0121] {"id":"6","text":"I also painted a colorful white","type":"text"},
[0122] {"id":"6","text":"White is the color with the most changes","type":"text"},
[0123] {"id":"7","text":"Absolutely colorful","type":"text"},
[0124] {"id":"7","text":"You should have listened to me and bought two pounds of pearls, ground them into powder and added them in","type":"text"},
[0125] {"id":"2","text":"Is there any colorful white?","type":"text"},
[0126] {"id":"2","text":"Who has done it well? Take a look and learn from it","type":"text"},
[0127] {"id":"8","text":"That's a rainbow","type":"text"},
[0128] {"id":"5","text":"My house is too messy, you can go if you want","type":"text"},
[0129] {"id":"5","text":"Linghuabai","type":"text"},
[0130] {"id":"2","text":"It's all messy now","type":"text"},
[0131] {"id":"7","text":"The color of my house is whatever color you see, I don't know what color it is","type":"text"},
[0132] {"id":"7","text":"Does it matter what color you paint? I can't tell","type":"text"},
[0133] {"id":"7","text":"They all look the same after painting, unless you jump the colors like XX, then it will be effective","type":"text"}
[0134] {"id":"8","text":"It's not too challenging","type":"text"},
[0135] {"id":"9","text":"I authorized the designer to help me change the color, but I don't know what the effect will be","type":"text"},
[0136] {"id":"7","text":"Others are collectively referred to as big white walls","type":"text"},
[0137] {"id":"7","text":"I just saw Ningxia selenium sand melon, 10 yuan per catty","type":"text"},
[0138] {"id":"9","text":"Maybe the workers went back to harvest the wheat? They didn't start brushing last week","type":"text"},
[0139] {"id":"8","text":"The wheat harvest has begun","type":"text"},
[0140] {"id":"8","text":"Everything I do can be slow these days","type":"text"},
[0141] {"id":"9","text":"No wonder","type":"text"},
[0142] {"id":"7","text":"Painting takes a long time. My painting cycle is 22 days, and I work overtime every day","type":"text"},
[0143] {"id":"8","text":"You even brushed the ground","type":"text"},
[0144] {"id":"4","text":"How to change the color? Are there any pictures?""type":"text"},
[0145] {"id":"8","text":"I'll send it to you","type":"text"},
[0146] {"id":"7","text":"It doesn't matter. Anyway, I calculated the cost of painting, labor and materials, and it cost 9400, including 200 plaster lines","type":"text"},
[0147] {"id":"7""text":"I think it's a good deal","type":"text"},
[0148] {"id":"8","text":"Waist thickness","type":"text"},
[0149] {"id":"7","text":"It's very cheap now. Paints cost 2,000 to 3,000 yuan","type":"text"},
[0150] {"id":"10","text":"It seems I have to go visit and learn a little bit","type":"text"},
[0151] {"id":"11","text":"The drain pipe on our balcony is clogged with cement and sand. It is best for everyone to check the drainage to avoid delaying the renovation.","type":"text"},
[0152] {"id":"12","text":"My house is also blocked, what should I do? Who can I contact to clear the blockage?","type":"text"},
[0153] {"id":"13","text":"My house is also blocked","type":"text"},
[0154] {"id":"5","text":"You have to go downstairs","type":"text"},
[0155] {"id":"5","text":"Open the inspection port","type":"text"},
[0156] {"id":"5","text":"Then poke","type":"text"},
[0157] {"id":"11","text":"The property I'm looking for","type":"text"},
[0158] {id":"11","text":"Let's check","type":"text"},
[0159] {"id":"5","text":"It's OK if the downstairs is not wrapped, but it's very difficult to wrap the pipes","type":"text”},
[0160] {"id":"11","text":"Fortunately, the renovation of our downstairs hasn't started yet","type":"text"},
[0161] {id":"11","text":"Otherwise it is true","type":"text"},
[0162] {"id":"13","text":"I found it when I was inspecting the house, and they later repaired it""type":"text"},
[0163] {id":"13","text":"Property to be found","type":"text"},
[0164] {id":"5","text":"I opened it myself","type":"text”},
[0165] {id":"13","text":"It was so crowded [speechless]","type":"text”},
[0166] {"id":"5","text":"Open the inspection port downstairs and do it yourself","type":"text"},
[0167] {"id":"5","text":"All cement sand","type":"text"},
[0168] {id":"13","text":"Try pouring some water into the other drains as well","type":"text”},
[0169] {id":"11","text":"The master said he couldn't knock it down","type":"text"},
[0170] {id":"11","text":"How did you open it?","type":"text"},
[0171] {"id":"14", "text": "I found that the inspection opening on my balcony has been sealed inside the ceiling. If there is a blockage upstairs in the future, it will be a problem to open this inspection opening.", "type": "text"},
[0172] {"id":"5", "text": "Poke it bit by bit.", "type": "text"},
[0173] {"id":"5", "text": "Or use something else.", "type": "text"},
[0174] {"id":"14", "text": "Have you sealed the inspection openings on your balconies?", "type": "text"},
[0175] {"id":"13", "text": "No.", "type": "text"}]
[0176] By using a trained large language model that can understand the context of text information to understand the context of multiple second unread messages in the above community owners' group, the summary information obtained can be:
[0177] The following is a brief summary of the unread messages:
[0178] · Neighbors discussed topics such as purchasing kitchen gas fittings and choosing the color of latex paint.
[0179] · Someone shared their own painting experience, including using pearl powder to create special effects.
[0180] · Discussed possible problems during decoration, such as blockage of the drain pipe under the balcony floor drain, and corresponding solutions.
[0181] · Someone mentioned the cost of decoration materials and how to enhance the decoration effect through color-blocking design.
[0182] · Neighbors provided each other with decoration suggestions, including how to dredge pipes and visit to learn decoration skills.
[0183] It can be seen that the above summary information is generated based on the context understanding of multiple second unread messages, rather than simple keyword extraction and sentence recombination. The above summary information covers all the key points discussed in multiple second unread messages and describes each key point in a concise and clear sentence for display to the user.
[0184] Based on the solution provided by the above embodiment of the present invention, useless information in the unread information of the target group chat is filtered out by preset filtering conditions, such as emoticons, gossip, and repeated information, and multiple second unread messages for information analysis are obtained. The text information in the multiple second unread messages does not need to be processed in any way, and can be directly sent to the large language model for understanding. The non-text information in the multiple second unread messages, such as picture information, audio and video information, and file information, is analyzed to obtain the text information corresponding to each non-text information. By analyzing the picture information, audio and video information, and file information in the second unread information, it is possible to avoid missing important information. Afterwards, the text information corresponding to the non-text information and the text information in the multiple second unread messages are contextually understood by the large language model to obtain the summary information of the unread information of the target group chat. Thus, accurate and comprehensive summary information corresponding to the unread information of the target group chat can be obtained by understanding the text information and non-text information in the unread information of the target group chat.
[0185] Figure 3 Flow chart of another summary generation method provided by an embodiment of the present invention. Figure 3 As shown, based on the above embodiment, the method may further include the following steps:
[0186] 301. Generate multiple automatic question and answer information related to the summary information and display them to the user.
[0187] 302. In response to a user's selection input for a plurality of automatic question and answer information, display a question and answer result corresponding to the automatic question and answer information selected by the user.
[0188] In this embodiment, after generating summary information of the target group chat unread messages, the topics involved in the target group chat unread messages are analyzed to evaluate the popularity of each topic, for example, the popularity of each topic is evaluated based on indicators such as discussion frequency and the number of times related keywords appear. Then, all topics discussed in the content of the target group chat unread messages are arranged according to the popularity of the topics, and several topics with higher popularity are determined as content that the user may be interested in, thereby determining hot topics that the user may be interested in.
[0189] Afterwards, multiple automatic question-and-answer information is generated based on the listed hot topics, which may include detailed explanations, relevant data, or answers to common questions, etc. These automatic question-and-answer information are displayed to users, allowing them to quickly obtain key information related to the hot topics without having to read a lot of information.
[0190] By displaying multiple automatic question and answer information to the user, the user can view and select any one or more of these question and answer information at will to find the topic that he is interested in. When the user selects any one of the multiple automatic question and answer information, in response to the user's selection input, the question and answer result corresponding to the selected automatic question and answer information is obtained and displayed to the user. In this way, the user can obtain the question and answer result corresponding to the topic that he is interested in without manually filtering and analyzing a large amount of information.
[0191] In actual applications, a conversation window can also be provided to the user on the display interface. If the user does not find the content of interest in the multiple displayed word-based Q&A information, the user can enter the topic of interest through the conversation window. For example, the user can ask a specific question or discuss a topic through the window, and then generate and display the relevant Q&A results based on the user's input. This way of interacting through the conversation window enhances interactivity and ensures that the user can obtain information that matches his or her needs, thereby further meeting the user's personalized needs.
[0192] For example, assuming that the summary information is generated based on the plurality of second unread messages in the above-mentioned community owner group, the plurality of automatically generated automatic question and answer messages are as follows:
[0193] “1) What was discussed about purchasing kitchen gas accessories?
[0194] 2) What was discussed about the choice of latex paint?
[0195] 3) What is your experience in painting?
[0196] 4) What are the problems you may encounter during decoration?
[0197] 5) What are the techniques for unclogging pipes?
[0198] If the topic that the user is interested in is the technique of unclogging pipes, then "5) What are the techniques of unclogging pipes?" is selected from the above five automatic question and answer information, and the selection method may be, for example, clicking on the automatic question and answer information on the display screen. In response to the user's selection operation, the electronic device obtains the question and answer result corresponding to the automatic question and answer information and displays it to the user. Specifically, the question and answer result may be as follows:
[0199] "Based on the content you provided, the techniques for unclogging pipes mainly include the following points:
[0200] 1. Go downstairs and open the inspection port: If the pipe is blocked, you can go downstairs to open the inspection port and directly deal with the blocked part.
[0201] 2. Use tools to clear the blockage: You can use appropriate tools (such as pipe clearers) to poke or poke the blockage to remove cement, sand and other debris in the pipe.
[0202] 3. Property assistance: If you cannot solve the problem yourself, you can seek help from the property and let professionals check and clear the problem.
[0203] 4. Check other drains: When dealing with a blockage point, you should also check whether other drains are also blocked to ensure that the entire drainage system is unobstructed.
[0204] 5. Preventive measures: Pay attention to protecting the pipes during renovation to prevent cement, sand and other construction materials from entering the pipes and causing blockages.
[0205] 6. Regular inspection: Even if there is no obvious blockage, the condition of the pipeline should be checked regularly to ensure that there is no potential risk of blockage.
[0206] 7. Use chemical unclogging agents: In some cases, you may consider using chemical unclogging agents to help clear the blockage, but pay attention to the product instructions and safe use.
[0207] 8. Keep pipes clear: After the renovation is completed, ensure that all pipes are clear and there is no leftover construction waste.
[0208] These tips can help neighbors take effective measures when they encounter pipe blockage problems."
[0209] Based on the solution provided by the above embodiment of the present invention, by generating question and answer information related to the summary information and displaying the question and answer results corresponding to the question and answer information after the user selects the question and answer information, the user can quickly obtain the hot topics in the unread information of the target group chat, and the interactivity can be enhanced by providing a conversation window to meet the user's personalized needs.
[0210] Figure 4 A flowchart of another method for generating summary information provided by an embodiment of the present invention. Figure 4 As shown, the method may include the following steps:
[0211] 401. Receive a first input from a user on a group chat page.
[0212] 402. In response to the first input, display summary information of unread group chat messages in the group chat corresponding to the group chat page.
[0213] When the user opens the instant messaging software, if there are multiple group chats in the main interface of the instant messaging software, the user can select the group chat that he wants to view as the target group chat and enter the group chat page of the group chat. At this time, a prompt message "Do you need to display the summary information of unread messages" can be popped up in the group chat page.
[0214] The user sends a first input for the prompt message on the group chat page. The first input may include multiple forms, such as click input, voice input, gesture input or touch input. The click input may be a single-click input, a double-click input or a triple-click input. The voice input may be a voice command, and the electronic device may recognize the content of the voice command based on voice recognition technology. Gesture input may be an operation in which a user inputs a command using a pre-set gesture to trigger the display of summary information of unread messages in the group chat. Touch input may be an operation in which a user touches an icon or area on the screen to trigger the display of summary information of unread messages in the group chat. The embodiments of the present invention do not limit the specific form of the first input.
[0215] For example, when the first input is a single-click input, the user can click "yes" displayed under the prompt message to trigger the operation of displaying the summary information of the unread information in the group chat. When the first input is a voice input, the user can choose whether to display the summary information of the unread information by speaking a voice command. Assuming that the user says "need to display the summary information of the unread information", the electronic device triggers the operation of displaying the summary information of the unread information in the group chat when it recognizes the content of the voice command based on the voice recognition technology. When the first input is a gesture input, the electronic device needs to capture the user's hand movement through the camera, and compare the captured hand movement with the pre-set gesture indicating affirmation. If the similarity between the captured hand movement and the pre-set gesture indicating affirmation meets the threshold, it indicates that the user needs to display the summary information of the unread information, and the operation of displaying the summary information of the unread information in the group chat is triggered. When the first input is a touch input, the electronic device detects that the user touches the icon component corresponding to the "yes" icon on the screen that expresses the need to display the summary information of the unread information, and then triggers the operation of displaying the summary information of the unread information in the group chat.
[0216] After receiving the first input, the electronic device displays summary information of the unread group chat information in the group chat corresponding to the group chat page in response to the first input. Specifically, the summary information of the unread group chat information is obtained by filtering the first unread information that meets the preset conditions in the unread group chat information to obtain multiple second unread information for information analysis, analyzing the non-text information in the multiple second unread information to obtain text information corresponding to each non-text information, and performing contextual understanding of the text information corresponding to the non-text information and the text information in the multiple second unread information through the large language model.
[0217] Among them, the electronic device obtains multiple second unread messages for information analysis by filtering the first unread messages that meet preset conditions in the group chat unread messages, analyzes the non-text information in the multiple second unread messages to obtain text information corresponding to each non-text message, and uses a large language model to perform contextual understanding on the text information corresponding to the non-text information and the text information in the multiple second unread messages to obtain summary information of the group chat unread messages. The process can refer to the relevant description in the above embodiments and will not be repeated here.
[0218] Based on the solution provided by the above embodiment of the present invention, accurate and comprehensive summary information can be generated for unread group chat information according to user needs.
[0219] Figure 5 A flowchart of another summary generation method provided by an embodiment of the present invention. Figure 5 As shown, the method may include the following steps:
[0220] 501. Receive a first input from a user on a group chat page.
[0221] 502. In response to the first input, display a plurality of automatic question and answer information related to the summary information; wherein the plurality of automatic question and answer information is generated based on the summary information.
[0222] 503. Receive a second input from the user regarding any automatic question and answer information.
[0223] 504. In response to the second input, display the question and answer result corresponding to the automatic question and answer information selected by the user.
[0224] For the description of step 501 , reference may be made to the relevant description of step 401 , which will not be repeated here.
[0225] After receiving the first input, in response to the first input, in addition to displaying the summary information of the unread group chat information in the group chat corresponding to the group chat page, multiple automatic question and answer information related to the summary information can also be displayed. Specifically, the multiple automatic question and answer information is to analyze the topics involved in the unread information of the target group chat after the summary information is generated to evaluate the popularity of each topic, for example, the popularity of each topic is evaluated according to indicators such as the discussion frequency and the number of times related keywords appear. Then, all the topics discussed in the content of the unread information of the target group chat are arranged according to the popularity of the topic, and several topics with higher popularity are determined as content that the user may be interested in, thereby determining the hot topics that the user may be interested in. Afterwards, it is generated based on the listed high-heat topics. This information may include detailed explanations, relevant data, or answers to common questions.
[0226] If the user is interested in any one or more of the multiple automatic question and answer information related to the summary information displayed, the user can trigger the operation of displaying the question and answer result corresponding to the selected automatic question and answer information by issuing a second input.
[0227] Specifically, the second input may include multiple forms, such as click input, voice input, gesture input or touch input. The click input may be a single-click input, a double-click input or a triple-click input. The voice input may be a voice command, and the electronic device may recognize the content of the voice command based on voice recognition technology. Gesture input may be an operation in which a user inputs a command using a pre-set gesture to trigger the display of summary information of unread group chat messages. Touch input may be an operation in which a user touches an icon or area on the screen to trigger the display of summary information of unread group chat messages. The embodiments of the present invention do not limit the specific form of the first input.
[0228] For example, when the second input is a single-click input, the user can trigger the operation of displaying the question and answer result corresponding to the selected automatic question and answer information by clicking on the automatic question and answer information that the user wants to know more about. When the second input is a voice input, the user can select the automatic question and answer information that he wants to know more about by speaking a voice command. For example, the user can directly read out the automatic question and answer information that he wants to know more about. When the electronic device recognizes the content of the voice command based on the voice recognition technology, it triggers the operation of displaying the question and answer result corresponding to the selected automatic question and answer information. When the second input is a gesture input, the electronic device needs to capture the user's hand movement through the camera, and compare the captured hand movement with the pre-set gesture for indicating the selected area. If the similarity between the captured hand movement and the pre-set gesture for indicating the selected area meets the threshold, it indicates that the user wants to further understand the automatic question and answer information displayed in the selected area, and the operation of displaying the question and answer result corresponding to the selected automatic question and answer information is triggered. When the second input is a touch input, the electronic device detects that the user touches a certain automatic question and answer information displayed on the screen, and then triggers the operation of displaying the question and answer result corresponding to the touched automatic question and answer information.
[0229] After receiving the second input, in response to the second input, the question and answer result corresponding to the selected automatic question and answer information is obtained and displayed to the user. In this way, the user can obtain the question and answer result corresponding to the topic of interest to him / her without manually screening and analyzing a large amount of information.
[0230] Based on the solution provided by the above embodiment of the present invention, by responding to the user's first input, while displaying accurate and comprehensive summary information of the group chat unread information, automatic question and answer information related to the summary information is displayed, and the corresponding question and answer results can be displayed based on the automatic question and answer information selected by the user, so that the user can quickly obtain hot topics in the unread information of the target group chat without reading a large amount of information. While enhancing interactivity, it can also meet the personalized needs of users.
[0231] The summary generation device of one or more embodiments of the present invention will be described in detail below. Those skilled in the art will appreciate that these devices can be configured using commercially available hardware components through the steps taught in this solution.
[0232] Figure 6 A schematic diagram of a summary generating device provided by an embodiment of the present invention is shown in FIG. Figure 6 As shown, the summary generating device 60 may include:
[0233] The acquisition module 61 is used to acquire unread information of a target group chat, where the unread information of the target group chat includes a plurality of first unread information.
[0234] The filtering module 62 is used to filter the first unread messages that meet the preset filtering conditions in the target group chat unread messages to obtain multiple second unread messages for information analysis.
[0235] The analysis module 63 is used to analyze the non-text information in the plurality of second unread messages to obtain text information corresponding to each non-text information; wherein the non-text information includes any one or more of picture information, audio and video information and file information.
[0236] The understanding module 64 is used to perform context understanding on the text information corresponding to the non-text information and the text information in the plurality of second unread messages through a large language model to obtain summary information of the unread messages of the target group chat.
[0237] In one embodiment, the analysis module 63 can be specifically used to analyze the target unread information through the analysis model corresponding to the message type of the target unread information when it is determined that the target unread information is non-text information according to the mark of the target unread information, obtain the text information corresponding to the target unread information, and send the text information corresponding to the target unread information to the preset information parsing cache library; the target unread information is any one of the multiple second unread information. When the target unread information is determined to be text information according to the mark of the target unread information, the text information is sent to the preset information parsing cache library. Correspondingly, the understanding module 64 is also used to obtain the text information corresponding to the target unread information and the text information in the multiple second unread information from the preset information parsing cache library, and perform contextual understanding of the text information through the large language model.
[0238] In one embodiment, the analysis module 63 can be specifically used for analyzing key elements in the target unread information through the image analysis model when the message type of the target unread information is picture information, and obtaining text information corresponding to the key elements. When the message type of the target unread information is audio and video information, the voice in the target unread information is analyzed through the voice analysis model to obtain text information corresponding to the voice. Or, when the message type of the target unread information is file information, the text content in the target unread information is analyzed through the file analysis model to obtain text information corresponding to the text content.
[0239] In one embodiment, the summary generating device 60 may further include a query module for querying whether there is an analysis result corresponding to the target unread information in the preset information analysis cache. If not, the triggering analysis module 63 analyzes the target unread information through the analysis model corresponding to the message type of the target unread information. If it exists, the analysis result is determined as the text information corresponding to the target unread information.
[0240] In one embodiment, the summary generating device 60 may further include a generating module and a responding module, wherein the generating module is used to generate a plurality of automatic question and answer information related to the summary information and display them to the user. The responding module is used to respond to the user's selection input for the plurality of automatic question and answer information and display the question and answer result corresponding to the automatic question and answer information selected by the user.
[0241] Figure 6 The summary generating device shown can execute the steps in the aforementioned embodiments. The detailed execution process and technical effects can be found in the description of the aforementioned embodiments, which will not be repeated here.
[0242] Figure 7 A schematic diagram of the structure of another summary generating device provided by an embodiment of the present invention. Figure 7 As shown, the summary generating device 70 may include:
[0243] The receiving module 71 is used to receive a first input from a user on a group chat page.
[0244] The response module 72 is used to display summary information of unread group chat messages in the group chat corresponding to the group chat page in response to the first input.
[0245] Among them, the summary information of the group chat unread information is obtained by filtering the first unread information that meets the preset conditions in the group chat unread information to obtain multiple second unread information for information analysis, analyzing the non-text information in the multiple second unread information to obtain text information corresponding to each non-text information, and using a large language model to understand the context of the text information corresponding to the non-text information and the text information in the multiple second unread information.
[0246] In one embodiment, the response module 72 may also be used to: in response to the first input, display multiple automatic question and answer information related to the summary information; wherein the multiple automatic question and answer information is generated based on the summary information. The receiving module 71 may also be used to receive a second input from the user for any automatic question and answer information. Accordingly, the response module 72 may also be used to display a question and answer result corresponding to the automatic question and answer information selected by the user in response to the second input.
[0247] Figure 7 The summary generating device shown can execute the steps in the aforementioned embodiments. The detailed execution process and technical effects can be found in the description of the aforementioned embodiments, which will not be repeated here.
[0248] The embodiment of the present invention also provides an electronic device, such as Figure 8 As shown, the electronic device may include: a processor 81, a memory 82, and a communication interface 83. The memory 82 stores executable codes, and when the executable codes are executed by the processor 81, the processor 81 can at least implement the summary generation method provided in the above embodiments.
[0249] In addition, an embodiment of the present invention provides a non-transitory machine-readable storage medium, on which executable code is stored. When the executable code is executed by a processor of an electronic device, the processor can at least implement the summary generation method provided in the aforementioned embodiment.
[0250] In addition, an embodiment of the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor of an electronic device, the processor can at least implement the summary generation method provided in the above embodiment.
[0251] The device embodiments described above are merely illustrative, and the network elements described as separate components may or may not be physically separated. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art may understand and implement the solution without creative effort.
[0252] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by adding a necessary general hardware platform, and of course can also be implemented by combining hardware and software. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a computer product, and the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
[0253] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for generating summary information, characterized in that: The method comprises: Acquire unread information of a target group chat, where the unread information of the target group chat includes a plurality of first unread messages; Filtering the first unread messages that meet the preset filtering condition in the target group chat unread messages to obtain a plurality of second unread messages for information analysis; Analyze the non-text information in the plurality of second unread messages to obtain text information corresponding to each of the non-text information; wherein the non-text information includes any one or more of picture information, audio and video information, and file information; The text information corresponding to the non-text information and the text information in the plurality of second unread messages are contextually understood by a large language model to obtain summary information of the target group chat unread messages.
2. The method according to claim 1, characterized in that The step of analyzing the non-text information in the plurality of second unread information to obtain text information corresponding to each of the non-text information includes: When it is determined that the target unread information is non-text information based on the mark of the target unread information, the target unread information is analyzed by using an analysis model corresponding to the message type of the target unread information to obtain text information corresponding to the target unread information, and the text information corresponding to the target unread information is sent to a preset information parsing cache library; the target unread information is any one of the multiple second unread information.
3. The method according to claim 2, characterized in that The method further comprises: When it is determined that the target unread information is text information according to the mark of the target unread information, the target unread information is sent to the preset information parsing cache library.
4. The method according to claim 2 or 3, characterized in that: The performing context understanding on the text information corresponding to the non-text information and the text information in the plurality of second unread information by using the large language model includes: The text information corresponding to the target unread information and the text information in the plurality of second unread information are obtained from the preset information parsing cache, and the context of the text information is understood by the large language model.
5. The method according to claim 2, characterized in that: The step of analyzing the target unread information by using an analysis model corresponding to the message type of the target unread information to obtain text information corresponding to the target unread information includes: When the message type of the target unread information is picture information, analyzing key elements in the target unread information by using a picture analysis model to obtain text information corresponding to the key elements; When the message type of the target unread information is audio or video information, the voice in the target unread information is analyzed by a voice analysis model to obtain text information corresponding to the voice; or When the message type of the target unread information is file information, the text content in the target unread information is analyzed by a file analysis model to obtain text information corresponding to the text content.
6. The method according to claim 2, characterized in that The method further comprises: querying whether there is an analysis result corresponding to the target unread information in the preset information analysis cache library; If not, analyzing the target unread information by using an analysis model corresponding to the message type of the target unread information; If so, the analysis result is determined as the text information corresponding to the target unread information.
7. The method according to claim 1, characterized in that The method further comprises: Generate multiple automatic question and answer information related to the summary information and display them to the user; In response to the user's selection input for the plurality of automatic question and answer information, a question and answer result corresponding to the automatic question and answer information selected by the user is displayed.
8. A method for generating summary information, characterized in that: The method comprises: Receive a first input from a user on a group chat page; In response to the first input, display summary information of unread group chat messages in the group chat corresponding to the group chat page; The summary information of the group chat unread information is obtained by filtering the first unread information that meets preset conditions in the group chat unread information to obtain multiple second unread information for information analysis, analyzing the non-text information in the multiple second unread information to obtain text information corresponding to each of the non-text information, and performing contextual understanding of the text information corresponding to the non-text information and the text information in the multiple second unread information through a large language model.
9. The method according to claim 8, characterized in that The method further comprises: In response to the first input, display a plurality of automatic question and answer information related to the summary information; wherein the plurality of automatic question and answer information is generated based on the summary information; Receiving a second input from a user regarding any of the automatic question and answer information; In response to the second input, a question-and-answer result corresponding to the automatic question-and-answer information selected by the user is displayed.
10. An electronic device, characterized in that: include: A memory, a processor, and a communication interface; wherein the memory stores executable code, and when the executable code is executed by the processor, the processor executes the method as described in any one of claims 1 to 7, or the processor executes the method as described in claim 8 or 9.
11. A non-transitory machine-readable storage medium, characterized in that: The non-transitory machine-readable storage medium stores executable code, and when the executable code is executed by a processor of an electronic device, the processor executes the method as described in any one of claims 1 to 7, or the processor executes the method as described in claim 8 or 9.
12. A computer program product, characterized in that include: A computer program, when executed by a processor of an electronic device, causes the processor to execute the method as claimed in any one of claims 1 to 7, or causes the processor to execute the method as claimed in claim 8 or 9.
