Method, device, and storage medium for initiating an active conversation

Abstract

An embodiment of the present invention provides a method, a device, and a storage medium for initiating an active conversation. The method for initiating an active conversation includes: obtaining a conversation task to be implemented; determining a target user who needs to have an active conversation based on the conversation task; generating an initiating user for having an active conversation with the target user, and determining conversation information corresponding to the target user; initiating a conversation with the target user based on the initiating user and the conversation information, and obtaining a conversation record corresponding to the target user, so as to complete the conversation task based on the conversation record. The technical solution provided in this embodiment effectively realizes the operation of initiating an active conversation with the target user. The right to speak in the conversation is not in the hands of the target user, but in the hands of the initiating user. In this way, the target user can be guided to have a conversation, thereby effectively improving the application scope and usage efficiency of this method. After the conversation with the target user is completed, the conversation task can be completed based on the conversation record with the target user.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular, to a method, device, and storage medium for initiating an active conversation. Background Art

[0002] With the continuous development of artificial intelligence technology, the human-computer interaction system based on dialogue technology has become increasingly mature, and dialogue robots have been applied to all aspects of daily life. At present, the interaction operation between the human-computer dialogue system and the user is mainly in the form of passive response, that is, the human-computer dialogue system mainly responds to the inquiries or demands initiated by the user. At this time, the initiative in the dialogue process is in the hands of the user; and the user needs to actively wake up the human-computer dialogue system. As can be seen from the above, the above-mentioned human-computer dialogue operation mainly relies on the subjective initiative of the user, thereby restricting the application scope and usage efficiency of the dialogue technology. Summary of the Invention

[0003] Embodiments of the present invention provide a method, device, and storage medium for initiating an active conversation, which effectively realizes the ability to actively initiate a conversation operation to a target user, thereby improving the application scope and usage efficiency of the method.

[0004] In a first aspect, an embodiment of the present invention provides a method for initiating an active conversation, including:

[0005] Obtain a dialogue task to be implemented;

[0006] Based on the dialogue task, determine a target user who needs to have an active conversation;

[0007] Generate an initiating user for having an active conversation with the target user, and determine dialogue information corresponding to the target user;

[0008] Initiate a conversation with the target user based on the initiating user and the dialogue information, and obtain a conversation record corresponding to the target user, so as to complete the dialogue task based on the conversation record.

[0009] In a second aspect, an embodiment of the present invention provides a device for initiating an active conversation, including:

[0010] A first obtaining module, configured to obtain a dialogue task to be implemented;

[0011] A first determining module, configured to determine a target user who needs to have an active conversation based on the dialogue task;

[0012] The first determining module is configured to generate an initiating user for having an active conversation with the target user, and determine dialogue information corresponding to the target user;

[0013] A first processing module, configured to initiate a conversation with the target user based on the initiating user and the conversation information, obtain a conversation record corresponding to the target user, and complete the conversation task based on the conversation record.

[0014] In a third aspect, an embodiment of the present invention provides an electronic device, including: a memory and a processor; wherein, the memory is used to store one or more computer instructions, and when the one or more computer instructions are executed by the processor, the method for initiating an active conversation in the first aspect above is implemented.

[0015] In a fourth aspect, an embodiment of the present invention provides a computer storage medium for storing a computer program, and when the computer program is executed by a computer, the method for initiating an active conversation in the first aspect above is implemented.

[0016] In a fifth aspect, an embodiment of the present invention provides a computer program product, including: a computer program, when the computer program is executed by a processor of an electronic device, the processor is caused to execute the steps in the method for initiating an active conversation shown in the first aspect above.

[0017] In a sixth aspect, an embodiment of the present invention provides a method for initiating an active conversation, which is applied to a wearable device, and the method includes:

[0018] Obtain the image information displayed by the wearable device;

[0019] When the image information includes a preset object or a preset area, generate an initiating user for initiating an active conversation with the user, and determine conversation information corresponding to the user, where the conversation information corresponds to the preset object or the preset area;

[0020] Initiate a conversation based on the initiating user and the conversation information, so as to perform an active conversation operation with the user through the wearable device.

[0021] In a seventh aspect, an embodiment of the present invention provides a device for initiating an active conversation, including:

[0022] A second acquisition module, configured to acquire the image information displayed by the wearable device;

[0023] A second generation module, configured to generate an initiating user for initiating an active conversation with the user and determine conversation information corresponding to the user when the image information includes a preset object or a preset area, where the conversation information corresponds to the preset object or the preset area;

[0024] A second processing module, configured to initiate a conversation based on the initiating user and the conversation information, so as to perform an active conversation operation with the user through the wearable device.

[0025] In an eighth aspect, an embodiment of the present invention provides an electronic device, including: a memory and a processor; wherein, the memory is used to store one or more computer instructions, and when the one or more computer instructions are executed by the processor, the method for initiating an active conversation in the above sixth aspect is implemented.

[0026] In a ninth aspect, an embodiment of the present invention provides a computer storage medium for storing a computer program, and when the computer program is executed by a computer, the method for initiating an active conversation in the above sixth aspect is implemented.

[0027] In a tenth aspect, an embodiment of the present invention provides a computer program product, including: a computer program, when the computer program is executed by a processor of an electronic device, the processor is caused to execute the steps in the method for initiating an active conversation shown in the above sixth aspect.

[0028] The technical solution provided in this embodiment, by obtaining a conversation task to be implemented, determining a target user who needs to have an active conversation based on the conversation task, then generating an initiating user for having an active conversation with the target user, determining conversation information corresponding to the target user, and initiating a conversation with the target user based on the initiating user and the conversation information, effectively realizes that it is not necessary for the target user to perform a wake-up operation, can actively initiate a conversation operation to the target user, and moreover, the right to speak in the conversation is not mastered by the target user, but by the initiating user, so that the target user can be guided to have a conversation, thereby effectively improving the application scope and usage efficiency of this method. In addition, after the conversation with the target user is completed, the conversation task can be completed based on the conversation record with the target user, further improving the practicality of this method and being conducive to market promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0030] Figure 1 It is a schematic flowchart of a method for initiating an active conversation provided by an embodiment of the present invention;

[0031] Figure 2 It is a schematic diagram of initiating an active conversation provided by an embodiment of the present invention;

[0032] Figure 3 It is a schematic flowchart of another method for initiating an active conversation provided by an embodiment of the present invention;

[0033] Figure 4 Flow diagram of another method for initiating an active conversation provided by an embodiment of the present invention;

[0034] Figure 5 Flow diagram of yet another method for initiating an active conversation provided by an embodiment of the present invention;

[0035] Figure 6 Flow diagram of still another method for initiating an active conversation provided by an embodiment of the present invention;

[0036] Figure 7 Flow diagram of a method for initiating an active conversation provided by an application embodiment of the present invention;

[0037] Figure 8 Structural diagram of a device for initiating an active conversation provided by an embodiment of the present invention;

[0038] Figure 9 For Figure 8 Structural diagram of an electronic device corresponding to the device for initiating an active conversation provided by the illustrated embodiment;

[0039] Figure 10 Flow diagram of another method for initiating an active conversation provided by an embodiment of the present invention;

[0040] Figure 11 Structural diagram of a device for initiating an active conversation provided by an embodiment of the present invention;

[0041] Figure 12 For Figure 11 Structural diagram of an electronic device corresponding to the device for initiating an active conversation provided by the illustrated embodiment. Detailed implementation manners

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention, but this does not exclude the case of including at least one. It should be understood that the term “and / or” used herein...

[0043] The terms used in the embodiments of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The singular forms "a", "the" and "said" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms unless the context clearly dictates otherwise. "Plural" generally includes at least two. "And / or" is a description of the relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0044] Depending on the context, the words "if", "when" as used herein can be interpreted as "when...", "when...", "in response to determining", or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detected (stated condition or event)" can be interpreted as "when determined", "in response to determining", "when detecting (stated condition or event)", or "in response to detecting (stated condition or event)".

[0045] It should also be noted that the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, so that a commodity or system including a series of elements not only includes those elements but also includes other elements not explicitly listed, or also includes elements inherent to such commodity or system. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the commodity or system including said element.

[0046] To facilitate understanding of the specific implementation process and implementation effects of the technical solutions in this application, the related technologies are briefly described below:

[0047] Currently, general-purpose dialogue robots can be roughly divided into two categories, namely task-based robots and chat-based robots. For task-based robots, their main goal is to complete specific tasks in as few short conversations as possible. At this time, the initiative of the conversation is in the hands of the user. In specific implementation, task-based robots include but are not limited to map navigation robots, meal ordering and ticket booking robots, and smart device robots. For chat-based robots, their main goal is to chat with users in as long conversations as possible. At this time, the initiative of the conversation is also in the hands of the user. In addition, since chatting can also be defined as a task, chat-based robots can be embedded in the framework of task-based robots. In specific implementation, chat-based robots include but are not limited to open-ended chatting robots, psychological counseling robots, and question-and-answer query robots.

[0048] It can be understood that the dialogue robot is not limited to the above two types and may also include question-and-answer robots, etc. However, no matter which type of robot or dialogue system it is, currently, the dialogue system not only requires the user to actively wake up the dialogue system. For example, the user wakes up the chat robot to play the idiom solitaire game, etc.; and the interaction with the user is mainly in the form of passive replies, that is, the system's replies mainly target the questions or needs initiated by the user, and the initiative in the entire dialogue process is in the hands of the user. Since the interaction between people in normal scenarios requires both active and passive methods, and the dialogue system under the passive human-computer interaction method depends on the subjective initiative of the user, this greatly limits the application scope and usage efficiency of the dialogue technology.

[0049] In addition, the interfaces of existing dialogue systems are usually open to users to achieve system wake-up operations. In this way, when the dialogue system monitors the user's input and triggers a session, it will run the system to reply. In addition, the passive dialogue interface needs to maintain real-time detection of the input end. For example, the dialogue system of a home audio needs to monitor the sound to determine whether the user wakes up, which poses a risk of violating the user's personal information security.

[0050] To solve the above technical problems, this embodiment provides a method, device, and storage medium for initiating an active dialogue. Among them, the execution subject of the method for initiating an active dialogue is a device for initiating an active dialogue, and this device for initiating an active dialogue can be implemented as a request end. Specifically, when implemented, the request end can be implemented as a physical robot, a virtual robot, etc. Specifically:

[0051] The request end can be any computing device with a certain data transmission ability. Specifically, when implemented, the request end can be a mobile phone, a tablet computer, a set application program, a robot, a cluster server, a conventional server, a cloud server, a cloud host, a virtual center, etc. In addition, the basic structure of the request end can include: at least one processor. The number of processors depends on the configuration and type of the request end. The request end can also include a memory, which can be volatile, such as RAM, or non-volatile, such as read-only memory (ROM), flash memory, etc., or can also include both types at the same time. Usually, an operating system (OS), one or more application programs, and program data, etc. are stored in the memory. In addition to the processing unit and the memory, the request end also includes some basic configurations, such as a network card chip, an IO bus, a display component, and some peripheral devices, etc. Optionally, some peripheral devices can include, for example, a keyboard, a mouse, a stylus, a printer, etc. Other peripheral devices are well known in the art and will not be elaborated here.

[0052] In the above-mentioned embodiment of the present invention, the requesting end can establish a network connection with other terminal devices to be conversed with, and this network connection can be a wireless or wired network connection. If the requesting end is communicatively connected to other terminal devices to be conversed with, the network mode of the mobile network can be any one of 2G (GSM), 2.5G (GPRS), 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G+ (LTE+), WiMax, 5G, etc.

[0053] In the embodiment of the present application, the requesting end can obtain the dialogue task to be realized, determine the target user who needs to initiate an active dialogue based on the dialogue task and the initiating user who initiates the active dialogue with the target user. The number of the target users can be one or more, and the corresponding initiating users for different target users can be different. The initiating user can be a virtual user. Then, determine the dialogue information corresponding to the target user, and initiate a dialogue with the target user based on the initiating user and the dialogue information, thus effectively realizing the initiation operation of the active dialogue. After the initiating user and the target user finish the dialogue, the dialogue task can be completed based on the dialogue information between the initiating user and the target user.

[0054] The technical solution provided in this embodiment effectively realizes the operation of initiating a dialogue with the target user without the target user performing a wake-up operation by obtaining the dialogue task to be realized, determining the target user who needs to initiate an active dialogue based on the dialogue task, then determining the dialogue information corresponding to the target user, and initiating a dialogue with the target user based on the initiating user and the dialogue information. Moreover, the right to speak in the dialogue is not mastered by the target user but by the initiating user, which can guide the target user to have a dialogue, thereby effectively improving the application scope and usage efficiency of this method. In addition, after the dialogue with the target user is completed, the dialogue task can be completed based on the dialogue record with the target user, further improving the practicability of this method.

[0055] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without 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 is not strictly limited.

[0056] Figure 1 It is a schematic flowchart of a method for initiating an active dialogue provided by an embodiment of the present invention; Figure 2 It is a schematic diagram for initiating an active dialogue provided by an embodiment of the present invention; Refer to the attached Figure 1 - Figure 2As shown in the figure, this embodiment provides a method for initiating an active conversation. The execution subject of this method can be a device for initiating an active conversation, and this device for initiating an active conversation can be implemented as software, or a combination of software and hardware. Specifically, the method for initiating an active conversation can include the following steps:

[0057] Step S101: Obtain the conversation task to be implemented.

[0058] Step S102: Determine the target user who needs to have an active conversation based on the conversation task.

[0059] Step S103: Generate the initiating user for having an active conversation with the target user, and determine the conversation information corresponding to the target user.

[0060] Step S104: Initiate a conversation with the target user based on the initiating user and the conversation information, and obtain the conversation record corresponding to the target user, so as to complete the conversation task based on the conversation record.

[0061] The following is a detailed description of each of the above steps:

[0062] Step S101: Obtain the conversation task to be implemented.

[0063] Among them, the device for initiating an active conversation can be pre-configured with the conversation tasks to be implemented. In different application scenarios, the device for initiating an active conversation can be configured with different conversation tasks. For example: in the e-commerce application field, the conversation task can be to detect users (buyers or sellers) with illegal transaction risks; or, the conversation task can be to find potential users interested in a certain commodity, etc. In the article query field, the conversation task can be to find potential readers interested in a certain article, etc. And, the number of conversation tasks to be recognized obtained by the device for initiating an active conversation can be one or more. At the same moment, the device for initiating an active conversation can initiate multiple different active conversation operations for multiple different conversation tasks.

[0064] In addition, the dialogue task can be a pre-configured periodic task or a real-time task. For example, the dialogue task can be "every three months, detect users (buyers or sellers) at risk of illegal transactions", or the dialogue task can be "at the current moment, detect users (buyers or sellers) at risk of illegal transactions", etc. At this time, after the active dialogue initiation device obtains the dialogue task to be implemented, it can analyze and process the dialogue task to determine the trigger time of the dialogue task. Among them, some dialogue tasks can be triggered in real time, and some dialogue tasks can be triggered at a preset moment. For example, a manager can configure the dialogue task A to be implemented and the dialogue task B to be implemented in the active dialogue initiation device. Among them, the dialogue task A can be triggered in real time, that is, after obtaining the dialogue task A, an active dialogue operation can be directly initiated for the dialogue task A. The dialogue task B is triggered at a preset moment. When the time is T1, the dialogue task B is triggered, and then an active dialogue operation can be initiated for the dialogue task B, which effectively improves the flexibility of use of the method for initiating an active dialogue.

[0065] In addition, for the active dialogue initiation device, a manager can configure the dialogue task to be implemented in the active dialogue initiation device based on the requirements of the application scenario. Specifically, a human-machine interaction interface is set on the active dialogue initiation device, and the manager inputs an execution operation on the human-machine interaction interface, so that the dialogue task to be implemented configured by the manager can be obtained based on the execution operation. In some other instances, the dialogue task to be implemented can be stored in a preset area or other devices. In this way, by actively or passively accessing the preset area or other devices, the active dialogue initiation device can stably obtain the dialogue task to be implemented.

[0066] Step S102: Determine the target user for whom an active dialogue needs to be conducted based on the dialogue task.

[0067] After obtaining the dialogue task, it can be analyzed and processed to determine the target user for whom an active dialogue needs to be conducted. It should be noted that different dialogue tasks can determine different target users. For example, when the dialogue task is "detect users at risk of illegal transactions", the target user is a potential user at risk of illegal transactions; when the dialogue task is "find users interested in a certain item", the target user is a potential buyer user of a certain item.

[0068] In addition, the method for determining the target user in this embodiment is not limited, and those skilled in the art can set it according to specific application scenarios or application requirements. In some instances, after obtaining the dialogue task, keywords corresponding to the dialogue task can be obtained, and the target user who needs to initiate an active dialogue can be determined based on the keywords. In other instances, determining the target user who needs to initiate an active dialogue based on the dialogue task may include: obtaining the portrait information of alternative users; and determining the target user from the alternative users based on the dialogue task and the portrait information.

[0069] In order to accurately determine the target user, the portrait information of alternative users can be obtained. In some instances, the alternative users can be all users participating in a certain event. For example, when the dialogue task is "detecting users at risk of illegal transactions", the alternative users can be all users involved in the e-commerce transaction event; when the dialogue task is "finding users interested in a certain book", the alternative users are all users involved in events related to the above book (such as: browsing book-related introductions, viewing book-related advertisements, searching for books, etc.).

[0070] Specifically, the portrait information of alternative users can be stored in a preset area, and the portrait information of alternative users can be obtained by accessing the preset area. After obtaining the portrait information, the dialogue task and the portrait information can be analyzed and processed to determine the target user from the alternative users. In some instances, determining the target user who needs to initiate an active dialogue from the alternative users may include: obtaining the correlation or similarity between the dialogue task and the portrait information, determining the portrait information with a correlation or similarity greater than or equal to a preset threshold as the target portrait, and then determining the alternative user corresponding to the target portrait as the target user. At this time, the target user is at least a part of the alternative users.

[0071] Step S103: Generate an initiating user for an active dialogue with the target user, and determine the dialogue information corresponding to the target user.

[0072] After obtaining the target user, in order to make the dialogue with the target user closer to the real scenario, an initiating user for an active dialogue with the target user can be generated. The initiating user can be a virtual user. In some instances, generating an initiating user for an active dialogue with the target user may include: obtaining all historical users who have had a dialogue with the target user, and generating an initiating user for an active dialogue with the target user based on all historical users. The initiating user can be a virtual user similar to the historical users. In other instances, generating an initiating user for an active dialogue with the target user may include: randomly generating a username and user information for an active dialogue with the target user, and generating an initiating dialogue for an active dialogue with the target user based on the username and user information.

[0073] In some other examples, the initiating user who generates an active conversation with the target user may include: obtaining anthropomorphic features for defining the initiating user; and generating a virtual initiating user for an active conversation with the target user based on the anthropomorphic features.

[0074] Specifically, although the target user is a virtual user, in order to make the conversation between the initiating user and the target user more real and reliable, anthropomorphic features for defining the initiating user can be obtained. The anthropomorphic features can be feature information pre-configured and stored in a preset area or a preset device. In this way, by accessing the preset area or the preset device, the anthropomorphic features for defining the initiating user can be obtained. The anthropomorphic features may include at least one of the following: static features of the initiating user, dynamic features of the initiating user, and natural person features of the initiating user. Among them, the static features of the initiating user may include at least one of the following: the avatar of the conversation system account, male / female, nickname, birthday and other static features; the dynamic features of the initiating user may include at least one of the following: the login ip, status, platform footprint, etc. of the account corresponding to the sending user; the natural person features of the initiating user may include at least one of the following: typing speed, onomatopoeia, message read, message unread, asynchronous session, continuous reply and other natural person features. After obtaining the anthropomorphic features, a virtual initiating user for an active conversation with the target user can be generated based on the anthropomorphic features. The virtual initiating user has the above-mentioned anthropomorphic features, thus effectively ensuring the quality and efficiency of generating the initiating user.

[0075] It should be noted that different users of the same conversation task can correspond to the same or different initiating users. For example, after obtaining the same conversation task to be implemented, when the determined target users include User A and User B, in order to be able to perform an active conversation operation with User A, the initiating user A can be determined; similarly, in order to be able to perform an active conversation operation with User B, the initiating user B can be determined. The initiating user A and the initiating user B are different users. In addition, the initiating users generated by the same user of the same conversation task at different times can be the same or different. For example, after obtaining the same conversation task to be implemented, if the determined target user includes User A, at time T1, in order to be able to perform an active conversation operation with User A, the initiating user a can be determined, and at time T2, in order to be able to perform an active conversation operation with User A, the initiating user b can be determined. The initiating user a and the initiating user b are different users.

[0076] In addition, after determining the target user, the conversation information corresponding to the target user can also be determined. The conversation information may include at least one of the following: text information, voice information, picture information, etc. And the conversation information can be binary classification question information, such asFigure 2 As shown, the conversation information may include: "Hello, are you there?" and "Do you have knives with a blade length of 30 centimeters?" This can guide the response information of the target user within the preset conversation logic range, thus facilitating the guarantee of the authenticity and reliability of the active conversation with the target user.

[0077] In addition, different target users can correspond to different conversation information. For example, when the target users determined by the same conversation task include target user A and target user B, since target user A is a seller of music CDs, the conversation information generated corresponding to target user A can be "User A, do you have CDs for sale?" Since target user B is a seller of medicines, the conversation information generated corresponding to target user B can be "User B, do you have XX medicine for sale?" This effectively expands the applicable range and usage frequency of the method for initiating an active conversation.

[0078] Furthermore, the implementation manner of determining the conversation information corresponding to the target user in this embodiment is not limited. In some instances, determining the conversation information corresponding to the target user may include: obtaining a conversation framework corresponding to the conversation task, and generating the conversation information corresponding to the target user based on the conversation framework and the target user. Of course, those skilled in the art can also use other methods to determine the conversation information corresponding to the target user. For example, a machine learning model for generating conversation information is preset, where the machine learning model is trained to generate conversation information; after obtaining the conversation task and the target user, the conversation task and the target user can be input into the machine learning model, so as to obtain the conversation information corresponding to the target user output by the machine learning model.

[0079] Step S104: Initiate a conversation with the target user based on the initiating user and the conversation information, and obtain the conversation record corresponding to the target user to complete the conversation task based on the conversation record.

[0080] After obtaining the initiating user and the conversation information, a conversation can be initiated with the target user based on the initiating user and the conversation information, thus effectively realizing the active conversation operation between the initiating user and the target user. After the active conversation between the initiating user and the target user is completed, the conversation record corresponding to the initiating user and the target user can be obtained. The conversation record may include the conversation information sent by the initiating user and the reply information sent by the target user. After obtaining the conversation record corresponding to the target user, the conversation record can be analyzed and processed to complete the conversation task based on the analysis and processing result, thus effectively ensuring the practicability of this method.

[0081] In some examples, after obtaining the conversation record corresponding to the target user, the method may further include: associatively storing the conversation record with the target user.

[0082] After obtaining the conversation record, the conversation record can be associatively stored with the target user, so that the target user can be analyzed and processed based on the conversation record stored associatively later, and then the conversation task can be completed based on the analysis and processing result, thus effectively ensuring the practicability of the method. It should be noted that for the same target user, when there are different initiating users, the conversation records associatively stored with the target user include multiple ones, and the multiple conversation records respectively correspond to different initiating users. When it is necessary to analyze and identify the target user, all the conversation records corresponding to the target user can be analyzed and processed, so that a more accurate analysis and processing result can be obtained, and then the accuracy rate of the realization of the conversation task is ensured. This not only makes the active conversation between the initiating user and the target user more vivid and authentic, but also ensures the accurate and reliable acquisition of the conversation record, facilitates the completion of the corresponding conversation task based on the stored conversation record, and further improves the accurate and reliable use of the method.

[0083] The method for initiating an active conversation provided in this embodiment realizes, by obtaining the conversation task to be realized, determining the target user who needs to have an active conversation based on the conversation task, then determining the conversation information corresponding to the target user, and initiating a conversation to the target user based on the initiating user and the conversation information, the function of actively initiating a conversation operation to the target user without the need for the target user to perform a wake-up operation. Moreover, the right to speak in the conversation is not mastered by the target user, but by the initiating user, so that the target user can be guided to have a conversation, thus effectively improving the application scope and use efficiency of the method. In addition, after the conversation with the target user is completed, the conversation task can be completed based on the conversation record with the target user, further improving the practicability of the method and being conducive to the market promotion and application.

[0084] Figure 3 It is a schematic flowchart of another method for initiating an active conversation provided by an embodiment of the present invention; referring to the appendix Figure 3 As shown, in order to improve the practicability of the method, the method in this embodiment may further include:

[0085] Step S401: Based on the conversation record, determine the probability information that the target user meets the conversation task.

[0086] Step S402: When the probability information is greater than or equal to a preset threshold, send the identity identifier of the target user and the conversation record to a preset platform, so that the preset platform manages the target user based on the conversation record and the identity identifier.

[0087] After obtaining the conversation record, in order to complete the conversation task based on the conversation record, the conversation record can be analyzed and processed to determine the probability information of the target user meeting the conversation task. In some examples, a machine learning model pre-trained for analyzing and processing the conversation record is used. The conversation record and the conversation task are input into the machine learning model, so that the probability information of the target user meeting the conversation task can be obtained. In other examples, based on the conversation record, determining the probability information of the target user meeting the conversation task may include: obtaining the conversation keywords corresponding to the conversation record and the task keywords corresponding to the conversation task, determining the correlation degree between the conversation keywords and the task keywords, and determining the probability information of the target user meeting the conversation task based on the correlation degree.

[0088] After obtaining the probability information, the probability information can be analyzed and compared with a preset threshold. When the probability information is less than the preset threshold, it indicates that the target user at this time is not the target user corresponding to the conversation task. Therefore, no operation needs to be performed on the conversation record. When the probability information is greater than or equal to the preset threshold, it indicates that the probability that the target user at this time is the target user corresponding to the conversation task is relatively high. At this time, the identity identifier of the target user and the conversation record can be sent to a preset platform. The preset platform can be a background server, a background processor, etc. After the preset platform obtains the conversation record and the identity identifier, it can manage the target user based on the conversation record and the identity identifier.

[0089] Specifically, the management operations for the target user may vary depending on the conversation task. For example, when the conversation task is to find potential illegal users, the management operations for the target user may include: identifying whether the target user is an illegal user; when the target user is an illegal user, muting the illegal user or closing the account; when the target user is a legal user, no processing operation may be performed on the legal user. When the conversation task is to find potential buyer users, the management operations for the target user may include: determining the intended product information corresponding to the target user; generating product recommendation information based on the intended product information; and sending the product recommendation information to the target user.

[0090] In this embodiment, based on the conversation record, the probability information of the target user meeting the conversation task is determined. When the probability information is greater than or equal to the preset threshold, the identity identifier of the target user and the conversation record are sent to the preset platform, so that the preset platform can manage the target user based on the conversation record and the identity identifier. Thus, when it is effectively identified on the preset platform that the target user is an illegal user, the illegal user can be muted or the account can be closed. When the target user is a potential buyer user, push information can be sent to the target user to increase the probability of product transactions, further improving the practicability of this method.

[0091] In some examples, after initiating a conversation with a target user based on the initiating user and conversation information, the method in this embodiment may further include: waiting to receive a reply message corresponding to the conversation information sent by the target user.

[0092] After initiating a conversation with a target user based on the initiating user and conversation information, since the conversation information is often a binary classification problem information, at this time, the target user often needs to reply based on the conversation information. In this way, for the initiating user, it is necessary to wait to receive a reply message corresponding to the conversation information sent by the target user. For example, if the initiating user sends conversation information A to the target user, after the target user obtains conversation information A, if the initiating user does not receive a reply message from the target user for conversation information A, the initiating user can keep waiting for the reply message sent by the target user.

[0093] Figure 4 It is a schematic flowchart of another method for initiating an active conversation provided by an embodiment of the present invention; refer to the appendix Figure 4 As shown, the method in this embodiment may further include:

[0094] Step S501: Receive a reply message corresponding to the conversation information sent by the target user.

[0095] After the initiating user initiates a conversation with the target user, the initiating user can wait to obtain a reply message corresponding to the conversation information sent by the target user. When the target user sends a reply message for the conversation information, the initiating user can receive the reply message corresponding to the conversation information sent by the target user. It should be noted that the reply messages corresponding to different conversation information from the same target user can be different.

[0096] Step S502: Determine a conversation strategy corresponding to the target user based on the reply message.

[0097] After obtaining the reply message, in order to be able to perform a more real and effective conversation operation with the target user, a conversation strategy corresponding to the target user can be determined. This conversation strategy is used to limit the conversation content between the initiating user and the target user. In some examples, different conversation strategies corresponding to different reply messages are pre-configured. At this time, determining a conversation strategy corresponding to the target user based on the reply message may include: obtaining the mapping relationship between the reply message and the conversation strategy, and determining the conversation strategy corresponding to the target user based on the reply message.

[0098] In still other examples, determining a conversation strategy corresponding to the target user based on the reply message may include: obtaining the meaning representation corresponding to the reply message; and determining the conversation strategy corresponding to the target user based on the meaning representation.

[0099] Specifically, after obtaining the reply information, a text processing algorithm or a pre-trained language model can be used to analyze and process the reply information to determine the corresponding meaning representation. Among them, the pre-trained language model (Pre-trained Language Model) can be a representation model constructed and trained on a large amount of corpus through autoencoding learning. The representation of each word in the pre-trained language model can change dynamically according to the context, and through further tuning in downstream tasks, it can often assist the model to achieve better results. Commonly used pre-trained language models include the BERT model (Bidirectional Encoder Representations from Transformers), the Funnel Transformer model, and so on.

[0100] After obtaining the meaning representation, the meaning representation can be analyzed and processed to determine the corresponding dialogue strategy for the target user, thus effectively realizing the logical information for determining the dialogue with the target user based on the meaning representation of the target user. In some instances, determining the dialogue strategy corresponding to the target user based on the meaning representation may include: when the meaning representation is the first representation, determining the dialogue strategy as the first strategy, and the first strategy is used to identify continuing the dialogue with the target user based on the reply information sent by the target user; when the meaning representation is the second representation, determining the dialogue strategy as the second strategy, and the second strategy is used to identify the operation of jumping out of the dialogue with the target user; when the meaning representation is the third representation, determining the dialogue strategy as the third strategy, and the third strategy is used to identify replying to the target user a preset number of times and then jumping out of the dialogue operation with the target user; when the meaning representation is the fourth representation, determining the dialogue strategy as the fourth strategy, and the fourth strategy is used to identify waiting for the target user to send a reply information.

[0101] After obtaining the meaning representation, the meaning representation can be analyzed and processed to determine the type of the meaning representation. Specifically, multiple standard representation types are preset (for example: the first representation, the second representation, the third representation, and the fourth representation). After obtaining the meaning representation, the corresponding target standard representation can be determined, and then the target standard representation can be determined as the type of the meaning representation. Specifically, when the meaning representation is the first representation, the dialogue strategy can be determined as the first strategy, and the first strategy is used to identify continuing the dialogue with the target user based on the reply information sent by the target user, thus effectively realizing the ability to continuously dialogue with the target user based on the reply information of the target user until no reply information from the target user is received.

[0102] For example, after the initiating user sends a conversation message (e.g., "Hello, are you there?") to user A (i.e., the target user), the initiating user can receive a reply message sent by user A (e.g., "Yes"). Then, based on the reply message, the initiating user can determine the conversation strategy corresponding to user A. When the conversation strategy is the first strategy, the initiating user can continue the conversation with user A based on the reply message sent by user A (e.g., "Do you sell a certain item?"). When no reply message from user A is received within a preset time period (e.g., 5 minutes, 10 minutes, or 15 minutes, etc.), the conversation operation with user A is stopped; when a reply message from user A is received within the preset time period, the conversation with user A can continue until no reply message sent by user A is received.

[0103] When the meaning representation is the second representation, the conversation strategy can be determined as the second strategy. The second strategy is used to indicate jumping out of the conversation operation with the target user, thus effectively achieving that after obtaining the reply message of the target user, there is no need to continue the conversation operation on the reply message of the target user, and directly jump out of the conversation operation with the target user. For example, the initiating user sends a conversation message (e.g., "Do you sell a certain item?") to user B (i.e., the target user). After the initiating user receives the reply message sent by user B (e.g., "Yes, here is the link"), based on the reply message, the initiating user can determine the conversation strategy corresponding to user B. When the conversation strategy is the second strategy, the initiating user can directly jump out of the conversation operation with user B based on the reply message sent by user B, that is, end the conversation operation with user B.

[0104] When the meaning representation is the third representation, the conversation strategy can be determined as the third strategy. The third strategy is used to indicate replying to the target user a preset number of times (which can be 1 time, 2 times, or 3 times, etc.), and then jumping out of the conversation operation with the target user, thus effectively achieving that after obtaining the reply message of the target user, after having several rounds of conversations with the target user, jumping out of the conversation operation with the target user. For example, the initiating user sends a conversation message ("Hello, are you there?") to user C. After the initiating user receives the reply message sent by user C ("Yes"), based on the reply message ("Yes"), the initiating user can determine the conversation strategy corresponding to user C. When the conversation strategy is the third strategy, the initiating user can determine the second-round conversation message (e.g., "Do you sell a certain item?"). Then, the initiating user can obtain the second-round reply message of user C for the second-round conversation message again, and based on the second-round reply message, generate the third-round conversation message (e.g., "Could you please give me the link, okay?"). Then, the initiating user can obtain the third-round reply message corresponding to the third-round conversation message, and then jump out of the conversation operation with user C, that is, end the conversation operation with user C.

[0105] When the meaning representation is the fourth representation, the conversation strategy can be determined as the fourth strategy. The fourth strategy is used to indicate waiting for the target user to send a reply message, thus effectively enabling waiting for the reply message sent by the target user all the time when having a conversation with the target user. For example, after the initiating user sends a conversation message to user D (e.g., "Do you sell a certain item?"), if no reply message from user D is received, the conversation message with user D can be maintained until a reply message from user D is received.

[0106] Step S503: Manage the conversation with the target user based on the conversation strategy.

[0107] After obtaining the conversation strategy, the conversation with the target user can be managed based on the conversation strategy. Specifically, when the conversation strategy is the first strategy or the third strategy, the management of the conversation with the target user based on the conversation strategy in this embodiment may include: generating a second conversation message based on the conversation strategy and the meaning representation, where the second conversation message includes at least one conversation message corresponding to the reply message sent by the target user; and having a conversation with the target user based on the second conversation message.

[0108] Specifically, when the conversation strategy is the first strategy or the third strategy, since both the first strategy and the third strategy require continuing the conversation with the target user, a second conversation message can be generated based on the conversation strategy and the meaning representation of the reply message. The second conversation message includes at least one conversation message corresponding to the reply message sent by the target user, that is, the second conversation message may include one round or more than two rounds of conversation messages for having a conversation with the target user. For example, the conversation between the initiating user and the target user is: "Hello, are you there?" The reply message from the target user is "Yes". Then, a second conversation message can be generated based on the reply message. The second conversation message can be "We have XXX items with guaranteed quality. Do you want them?" If the reply message from the target user to the above second conversation message is "Send me the link to have a look", then a second conversation message can be generated based on this reply message. Specifically, the second conversation message can be "The XXX link is sent to you. You can consult me at any time if you have anything you want to know". The above second conversation messages include two rounds of conversation messages.

[0109] It should be noted that different application scenarios, different dialogue strategies, and different reply messages can generate different second dialogue messages. In this embodiment, the specific generation method of the second dialogue message is not limited, and those skilled in the art can configure it according to specific application scenarios or application requirements. In some instances, based on the dialogue strategy and the meaning representation, generating the second dialogue message may include: obtaining a database corresponding to the dialogue strategy and the meaning representation, and generating the second dialogue message based on the database and the meaning representation; and / or, obtaining a dialogue template corresponding to the dialogue strategy and the meaning representation, and generating the second dialogue message based on the dialogue template and the meaning representation; and / or, based on a machine learning model, processing the dialogue strategy and the meaning representation by the machine learning model to generate the second dialogue message. After obtaining the second dialogue message, a dialogue can be carried out with the target user based on the second dialogue message, thus effectively ensuring the initiative and authenticity of the dialogue with the target user.

[0110] Among them, when the second dialogue message is generated based on a database, a database for generating dialogue messages is pre-configured. The database includes multiple preset dialogue messages. Since different dialogue strategies and meaning representations can correspond to different databases, in order to accurately determine the second dialogue message, after obtaining the dialogue strategy and the meaning representation, a database corresponding to the dialogue strategy and the meaning representation can be obtained. The database can include at least one of the following: a text library, a voice library, a picture library, etc. After obtaining the database, the second dialogue message can be generated based on the database and the meaning representation, that is, selecting a dialogue message adapted to the meaning representation in the database as the second dialogue message, thus effectively realizing the accurate and reliable determination of the second dialogue message.

[0111] When the second dialogue message is generated based on a dialogue template, multiple dialogue templates for generating dialogue messages are pre-configured. Since different dialogue strategies and meaning representations can correspond to different dialogue templates, in order to accurately determine the second dialogue message, after obtaining the dialogue strategy and the meaning representation, a dialogue template corresponding to the dialogue strategy and the meaning representation can be obtained. The dialogue module can be: "Do you have

XX

XX

XX

XX

[0112] When the second conversation message is generated based on a machine learning model, the machine learning model is pre-trained. The machine learning model can be an embedded sequence-to-sequence model. The machine learning model can perform end-to-end text generation operations based on the user's meaning expression, so as to generate the second conversation message, further ensuring the quality and efficiency of the generation of the second conversation message. Among them, Sequence-to-sequence means mapping a sequence in a certain domain to a sequence in another domain through a model. For example, translating an English sentence into a Chinese sentence. End-to-end usually refers to a process or service that provides an output from the initial input to the final required result for a certain problem. As a complete executable solution, it usually does not require additional processing or additional third-party support.

[0113] In this embodiment, by receiving the reply message corresponding to the conversation message sent by the target user and determining the conversation strategy corresponding to the target user based on the reply message, the conversation with the target user can then be managed based on the conversation strategy. Thus, when different reply messages sent by the target user are received, different conversation strategies can be adopted to complete the conversation operation with the target user, which not only ensures the integrity and continuity of the conversation with the target user, but also ensures the initiative, flexibility, and guidance of the conversation, further improving the practicability of this method.

[0114] Figure 5 It is a schematic flowchart of another method for initiating an active conversation provided by an embodiment of the present invention; refer to the appendix Figure 5 In this embodiment, another implementation manner of having a conversation with the target user is provided. Specifically, the method in this embodiment may further include:

[0115] Step S601: Receive the reply message corresponding to the conversation message sent by the target user.

[0116] Among them, the specific implementation manner and implementation effect of this step in this embodiment are similar to those of step S501 in the above embodiment. For specific reference, please refer to the above content and will not be elaborated here.

[0117] Step S602: Generate the second conversation message corresponding to the target user based on the reply message.

[0118] After obtaining the reply information, the reply information can be analyzed and processed to generate second dialogue information corresponding to the target user. In this embodiment, the generation method of the second dialogue information is not limited, and those skilled in the art can set it according to specific application scenarios or application requirements. In some examples, generating second dialogue information corresponding to the target user based on the reply information may include: obtaining the meaning representation corresponding to the reply information, determining the database corresponding to the meaning representation, and generating the second dialogue information based on the database and the meaning representation. In some other examples, generating second dialogue information corresponding to the target user based on the reply information may include: obtaining the meaning representation corresponding to the reply information, determining the dialogue template corresponding to the meaning representation, and generating the second dialogue information based on the dialogue template and the meaning representation. In still other examples, generating second dialogue information corresponding to the target user based on the reply information may include: obtaining the meaning representation corresponding to the reply information, determining the machine learning model corresponding to the reply information, and processing the meaning representation based on the machine learning model to generate the second dialogue information.

[0119] Step S603: Determine the dialogue strategy corresponding to the target user.

[0120] In order to ensure the authenticity of the dialogue with the target user, during the process of actively communicating with the target user, the dialogue strategy corresponding to the target user can be determined. In some examples, the dialogue strategy corresponding to the target user can be determined based on the reply information. Specifically, determining the dialogue strategy corresponding to the target user based on the reply information may include: obtaining the meaning representation corresponding to the reply information; determining the dialogue strategy corresponding to the target user based on the meaning representation. In some other examples, determining the dialogue strategy corresponding to the target user based on the meaning representation may include: when the meaning representation is the first representation, determining the dialogue strategy as the first strategy, where the first strategy is used to indicate continuing the dialogue with the target user based on the reply information sent by the target user; when the meaning representation is the second representation, determining the dialogue strategy as the second strategy, where the second strategy is used to indicate jumping out of the dialogue operation with the target user; when the meaning representation is the third representation, determining the dialogue strategy as the third strategy, where the third strategy is used to indicate replying to the target user a preset number of times and then jumping out of the dialogue operation with the target user; when the meaning representation is the fourth representation, determining the dialogue strategy as the fourth strategy, where the fourth strategy is used to indicate waiting for the target user to send a reply message.

[0121] It should be noted that the specific determination method of the dialogue strategy in this embodiment is similar to the implementation method of step S502 in the above embodiment, and the specific content can be referred to the above description, so it will not be elaborated here.

[0122] Step S604: Based on the dialogue strategy, determine whether to communicate with the target user based on the second dialogue information.

[0123] Since the dialogue strategy may be a strategy that requires continuing the dialogue with the target user or a strategy that does not require continuing the dialogue with the target user, in order to ensure the authenticity and reliability of the dialogue with the target user, after obtaining the dialogue strategy, the dialogue strategy can be analyzed and processed to determine whether to perform a dialogue operation with the target user based on the second dialogue information. In some examples, determining whether to perform a dialogue operation with the target user based on the dialogue strategy may include: when the dialogue strategy is the first strategy or the third strategy, performing a dialogue with the target user based on the second dialogue information; where the first strategy is used to identify continuing the dialogue with the target user based on the reply information sent by the target user, and the third strategy is used to identify replying to the target user a preset number of times and then exiting the dialogue with the target user; when the dialogue strategy is the second strategy or the fourth strategy, clearing the second dialogue information, where the second strategy is used to identify exiting the dialogue operation with the target user, and the fourth strategy is used to identify waiting for the target user to send a reply message.

[0124] Specifically, when the dialogue strategy is the first strategy or the third strategy, since both the first strategy and the third strategy require continuing the dialogue with the target user, a dialogue can be performed with the target user based on the previously generated second dialogue information, and the second dialogue information includes at least one dialogue information corresponding to the reply information sent by the target user, that is, the second dialogue information can include one round or more than two rounds of dialogue information for communicating with the target user. When the dialogue strategy is the second strategy or the fourth strategy, since the second strategy and the fourth strategy do not require continuing the dialogue with the target user, the pre-generated second dialogue information can be cleared, thereby effectively realizing the effective management and adjustment operation of the dialogue corresponding to the target user.

[0125] To understand the technical solution in this embodiment, the following will Figure 4 the technical solution of the corresponding embodiment and Figure 5 the technical solution of the corresponding embodiment are compared and described:

[0126] Figure 4 The implementation manner of the corresponding embodiment is: when the initiating user initiates an active dialogue operation with the target user, the target user can reply based on the dialogue information, and the initiating user can obtain the reply information. After obtaining the reply information, the dialogue strategy corresponding to the target user can be determined first. When the dialogue strategy requires continuing the dialogue with the target user, the second dialogue information for communicating with the target user can be generated, and a dialogue operation can be performed with the target user based on the second dialogue information; when the dialogue strategy does not require continuing the dialogue with the target user, there is no need to generate the second dialogue information for communicating with the target user.

[0127] Figure 5 The implementation method of the corresponding embodiment is as follows: when the initiating user initiates an active conversation operation with the target user, the target user can reply based on the conversation information, and the initiating user can obtain the reply information. After obtaining the reply information, the initiating user can first generate second conversation information corresponding to the reply information. At this time, the second conversation information can be displayed in the information display area to be sent to the target user. Then, determine the conversation strategy corresponding to the target user. When the conversation strategy requires continuing the conversation with the target user, the initiating user can perform a conversation operation with the target user based on the pre-generated second conversation information, that is, send the pre-generated second conversation information in the information display area to the target user. When the conversation strategy does not require continuing the conversation with the target user, the initiating user can clear the pre-generated second conversation information, that is, delete the second conversation information located in the information display area.

[0128] In this embodiment, by receiving the reply information corresponding to the conversation information sent by the target user, generating the second conversation information corresponding to the target user based on the reply information, determining the conversation strategy corresponding to the target user, and then judging whether to perform a conversation with the target user based on the second conversation information according to the conversation strategy, it effectively realizes that when it is necessary to have a conversation with the target user, the initiating user can perform a conversation operation with the target user based on the pre-generated second conversation information, and when it is not necessary to have a conversation with the target user, the pre-generated second conversation information can be cleared, thereby effectively improving the flexible conversation operation with the target user and further enhancing the practicality of this method.

[0129] Figure 6 It is a schematic flowchart of another method for initiating an active conversation provided by an embodiment of the present invention; referring to the attached Figure 6 As shown, in order to improve the security and reliability of the use of this method, after initiating a conversation with the target user based on the initiating user and the conversation information, the method in this embodiment may further include:

[0130] Step S701: Obtain the conversation status with the target user.

[0131] Among them, when the initiating user has a conversation with the target user, the conversation with the target user can correspond to different conversation statuses. For example: normal status and end status. When the conversation between the initiating user and the target user is in the normal status, it means that the initiating user and the target user can perform normal conversation operations. When the conversation between the initiating user and the target user is in the end status, it means that the initiating user and the target user have ended the conversation.

[0132] In order to ensure the safety and reliability of the dialogue operation, the dialogue state with the target user can be obtained. In some instances, identification information for identifying the dialogue state is pre-configured, and the dialogue state with the target user is determined based on the identification information. For example, when the identification information is "1", it can be determined that the dialogue state with the target user is the normal state; when the identification information is "0", it can be determined that the dialogue state with the target user is the end state. Or, in some other instances, during the process of having a dialogue with the target user, the dialogue strategy corresponding to the target user can be obtained. When the dialogue strategy is used to indicate continuing the dialogue with the target user, it can be determined that the dialogue state with the target user is the normal state; when the dialogue strategy is used to indicate jumping out of the dialogue operation with the target user, it can be determined that the dialogue state with the target user is the end state.

[0133] Step S702: When the dialogue state is the end state, control the initiating user who is used to have a dialogue with the target user to prohibit receiving any information.

[0134] After obtaining the dialogue state, management operations can be performed on the initiating user who has a dialogue with the target user based on the dialogue state. Specifically, when the dialogue state is the end state, the dialogue between the initiating user and the target user can be ended. At this time, in order to ensure the safety and reliability of the data, the initiating user who is used to have a dialogue with the target user can be controlled to prohibit receiving any information. In this way, it is effectively realized that after ending the dialogue with the target user, the initiating user will not receive any information. For example, it will not receive push information, advertising information, information sent by other users, etc. In this way, the safety and reliability of information processing are effectively ensured.

[0135] In some other instances, the method in this embodiment may further include:

[0136] Step S801: When the dialogue state is the normal state, allow the initiating user to accept a third-party intervention request.

[0137] Step S802: Manage the initiating user based on the third-party intervention request.

[0138] Specifically, when the dialogue information is in the normal state, the initiating user can be allowed to accept a third-party intervention request. The third-party intervention request may include at least one of the following: a manual intervention request, a management platform intervention request, a system push message, a system reminder message, etc. After obtaining the third-party intervention request, the initiating user can be managed based on the third-party intervention request, thereby effectively realizing that the initiating user, who was originally a virtual user, can be switched to a real user, making the dialogue with the target user more authentic and reliable.

[0139] For example, during the process of initiating a conversation between the initiating user and the target user, the conversation state of the conversation between the initiating user and the target user can be obtained. When the conversation state is in a normal state, during the process of the initiating user and the target user having a conversation, an artificial intervention request or a management platform intervention request sent by the system or the platform can be received. After obtaining the above artificial intervention request or management platform intervention request, an artificial management operation can be initiated for the initiating user based on the artificial intervention request and the management platform intervention request, that is, switching the virtual user having a conversation with the target user to a real person, thus effectively realizing the artificial switching operation.

[0140] Alternatively, during the process of the initiating user and the target user having a conversation, when a push message sent by the system or the platform is received, a control instruction replied by the initiating user can be obtained based on the push message, and a corresponding operation can be executed based on the control instruction. For example: when the push message is a fraud reminder message, the initiating user can generate a marking instruction for the fraud reminder message, and add identification information of the fraud reminder to the conversation between the initiating user and the target user based on the marking instruction; when the push message is an advertisement message, the initiating user can generate a closing instruction for the advertisement message, close the advertisement message based on the closing instruction, and then the conversation state between the initiating user and the target user can be maintained, thus effectively improving the authenticity of the initiating user.

[0141] In this embodiment, by obtaining the conversation state with the target user, when the conversation state is in an end state, the initiating user used to have a conversation with the target user is controlled to prohibit receiving any information. When the conversation state is in a normal state, the initiating user is allowed to accept a third-party intervention request, and the initiating user is managed based on the third-party intervention request, thus effectively improving the authenticity and anthropomorphism of the initiating user, enabling the target user to have a more real conversation with the initiating user, facilitating the realization of the conversation task, and further improving the practicability of this method.

[0142] In specific applications, this application embodiment provides a method for initiating active dialogue for human-computer dialogue. The method can be applied to a dialogue robot. The dialogue goal of the dialogue robot is "to obtain enough information for downstream tasks in the shortest possible dialogue process" and has anthropomorphic characteristics, that is, the initiating user needs to simulate the interaction characteristics of a natural person as much as possible. The right to speak in the dialogue is controlled by the dialogue robot, that is, the dialogue robot can not only actively initiate a conversation, but also guide the user to conduct a conversation to obtain the required information. Since the dialogue robot does not rely on the user's active triggering operation, the application scope and use efficiency of the method are expanded. In addition, the conversation initiated by the dialogue robot in this application embodiment is mainly dominated by the algorithm, which realizes complete control of the algorithm interface, avoids the key issues of algorithm governance such as third parties calling the algorithm interface to counter attacks, model theft, capability abuse, and violation of user personal privacy, and further improves the security and reliability of the use of this method. Specifically, the method in this embodiment may include the following steps:

[0143] Step 1: Get the conversation task with the conversation robot.

[0144] In different application scenarios, the conversation robot can be configured with different conversation tasks. For example, the conversation task can be "used to identify suspected users of illegal operations". Specifically, in the application scenario of e-commerce, since the risk identification operation of the e-commerce platform is mainly for static information, the risk identification operation for dynamic information is mainly achieved through manual detection, which greatly increases the cost and efficiency of manual identification. Therefore, in order to be able to perform legality detection operations on dynamic information, the conversation robot can act as an initiating user that simulates the buyer's interaction characteristics and initiate an active conversation operation to the target user who needs to perform an active conversation operation. At this time, the conversation robot is configured with an interface for actively obtaining information from the outside to realize active conversation operations. In specific implementation, the conversation robot at this time can be a task-based robot with the characteristics of a chat robot.

[0145] In addition, in the application scenario of e-commerce, since e-commerce platforms mainly use advertisements, recommendation systems, activities, etc. to dig out and guide buyers to consume, most of the above methods require passively waiting for consumers to receive information, and platform merchants do not all have the funds and qualifications to carry out the above promotions, and active advertising push often causes user disgust. Therefore, in order to solve the above technical problems, the dialogue task in this embodiment can be "used to determine potential users with purchasing tendencies". At this time, the dialogue robot can be used as an initiating user simulating the seller's interactive characteristics, and initiate active dialogue operations to the buyer user. The dialogue robot can be configured with an interface for actively reaching buyers, which can be defined as a task-based robot with the characteristics of a chat robot, which effectively realizes the low-threshold product push operation.

[0146] Step 2: Determine the target users for proactive conversations.

[0147] Before initiating a proactive conversation, user portrait information can be obtained based on the conversation task. This portrait information can include static information in the preset system. Based on the portrait information and the conversation task, the target users for proactive conversations are determined. A target user list can be constructed based on the user's portrait information. This target user list can include multiple target users, and different conversation tasks can correspond to different target users. For example, when the conversation task is to identify suspected users of illegal operations, the target users can be merchants with agreements to sell illegal goods. At this time, by conducting proactive conversation operations with the above target users, hidden transactions of the target users can be discovered and evidence can be obtained. When the conversation task is to identify potential users with purchase needs, the purchase tendency of the target users is determined based on the conversation information with the target users. When the purchase tendency is greater than or equal to the preset threshold, push information about the product can be sent to the target users.

[0148] Step 3: Generate the initiating user for the conversation operation with the target user.

[0149] Specifically, since the chatbot needs to have anthropomorphic features, in order to make the conversation more real and reliable, anthropomorphic features for defining the initiating user can be obtained. The anthropomorphic features include but are not limited to: (a) static features such as the avatar, male / female, nickname, birthday, etc. of the conversation account; (b) dynamic features such as the login IP, status, and platform footprint of the conversation account; (c) natural person features such as typing speed, onomatopoeia, message read status, asynchronous conversation, and continuous reply. After obtaining the anthropomorphic features, the initiating user for the conversation operation with the target user can be generated based on the anthropomorphic features. This initiating user is a virtualized user.

[0150] Step 4: Determine the conversation information corresponding to the target user. Based on the conversation information and the initiating user, initiate a conversation with the target user to obtain the conversation record corresponding to the target user, so as to complete the conversation task based on the conversation record.

[0151] When completing a dialogue task based on a dialogue record, corresponding execution operations can be performed based on the dialogue record. For example, when it is determined based on the dialogue record that the target user is a user at risk of illegal operations, the dialogue record can be provided to a preset platform for manual review operations, so that the preset platform can review the target user again based on the dialogue record, which can avoid misreview operations on the target user. When it is determined that the target user is an illegal user, corresponding operations can be performed on the illegal user. For example, silencing processing, store closing processing, etc. The execution subject for performing the corresponding operations on the illegal user can be a preset platform, such as an e-commerce platform, a management platform, etc. This effectively ensures the legality and security of data operation.

[0152] Among them, when determining the dialogue information corresponding to the target user, the method in this embodiment may include the following steps:

[0153] Step 4.1: Initiate a session.

[0154] The dialogue robot does not need to be awakened by the user, but actively pushes a session to the target user in the target user list according to the initiating user and the dialogue information. The forms of the session include but are not limited to text, voice, and pictures. It should be noted that the session content may vary according to different target users. For example, when the target user is a platform merchant, the actively initiated dialogue information can be of text type, and the specific content can include "Excuse me, is the boss there?"

[0155] Step 4.2: Make a reply.

[0156] After obtaining the reply information sent by the target user, the initiating user can make a dialogue reply based on the reply information sent by the target user. The forms of the dialogue reply include but are not limited to text, voice, and pictures.

[0157] Step 4.3: Wait for information.

[0158] When the dialogue robot sends dialogue information to the target user, it can always wait for the information for the target user to reply based on the dialogue information, that is, wait for the user to reply and keep the session going.

[0159] Step 5: End the session, extract and analyze information according to the dialogue history to complete the dialogue task.

[0160] When ending a conversation between the initiating user and the target user, the conversation record can be stored in a preset database for analysis and use in downstream conversation tasks. For example, when the conversation goal is to determine merchant risk and complete evidence collection, the merchant risk can be determined based on the conversation record, and key information can be extracted from the conversation history as the basis for judgment. Specifically, if a suspected merchant with the violation of selling controlled knives replies "Sorry, we don't sell knives this long. It's not allowed by the platform", the merchant risk can be determined as low, and "don't sell" is used as the basis for judgment. On the contrary, if the suspected merchant replies "Add me for private chat. You take a product and change the price with a note", the merchant risk can be determined as high, and "Add me for private chat" and "You take a product" are used as the basis for judgment, thus completing the conversation task.

[0161] After obtaining the reply information of the target user, the method in this embodiment may further include:

[0162] Step 11: Parse the reply information sent by the target user to obtain the meaning representation corresponding to the reply information.

[0163] After obtaining the reply information, natural language processing algorithms, speech recognition algorithms, image recognition and other technologies can be used to obtain the meaning representation corresponding to the reply information. The meaning representation may include, but is not limited to, text content, text representation of speech content, text representation of picture content, emotion and intention. For example: when the reply information of the target user is text information, the content may be "Yes, what do you need?", at this time, by parsing the reply information, it can be determined that the meaning representation of the target user is "interrogative sentence, positive attitude, asking for needs".

[0164] It should be noted that the meaning representation in this application embodiment can be an explicit expression for easy understanding. In practice, the meaning representation also includes custom ways such as implicit representation. For example: natural language processing technology can be used to convert "Yes, what do you need?" into a hidden state represented by a tensor as the input information for downstream modules.

[0165] In addition, after the initiating user initiates a session or makes a reply, it is necessary to wait for the reply of the target user. If no reply information sent by the target user is received within the given waiting time window period, it can be determined that the conversation times out at this time, and then the conversation with the target user can be ended.

[0166] Step 12: Based on the meaning representation, perform a judgment operation on the conversation logic between the sending user and the target user.

[0167] After obtaining the indication of intention, techniques such as embedded rules, decision models, or neural networks can be used to analyze and process the indication of intention to determine the conversation logic between the sending user and the target user, so as to judge the next action, including but not limited to: replying to the user, replying to the user and exiting the conversation, not replying to the user, and not replying to the user and exiting the conversation. Specifically:

[0168] (a) When the conversation logic requires replying to the target user, step 13 can be executed, that is, continue the conversation operation with the target user based on the reply information of the target user.

[0169] (b) When the conversation logic requires replying to the target user and exiting the conversation, step 13 can be executed, that is, generate the second conversation information based on the reply information of the target user. After sending the second conversation information to the target user, the conversation operation can be exited.

[0170] (c) When the conversation logic does not require replying to the target user, wait for the target user to reply and keep the conversation going.

[0171] (d) When the conversation logic does not require replying to the target user and exiting the conversation, the conversation operation can be directly exited.

[0172] For example, when the reply information sent by the target user is text information and the content is "Yes, what do you need?", by analyzing and processing the reply information, the indication of intention of the target user can be determined, that is, "interrogative sentence, positive attitude, asking for needs", and then the next action can be determined as "reply to the user and wait for the user to reply". On the contrary, when the reply information sent by the target user is text information and the content is "Not in, this is an automatic reply", by analyzing and processing the reply information, the indication of intention of the target user can be determined as "affirmative sentence, neutral attitude, no needs", and then the next action can be determined as "not reply to the user and exit the conversation".

[0173] It should be noted that for the initiating user, after the conversation between the initiating user and the target user ends, the initiating user can be controlled to block the conversations actively initiated by ordinary users, such as advertisement pushes, etc. At the same time, when performing the conversation operation between the initiating user and the target user, the initiator of the current conversation can be verified. If it is not initiated by the system, then "do not reply to the user and exit the conversation".

[0174] Step S13: Generate the second conversation information based on the conversation logic and the indication of intention of the reply information, so as to continue the conversation operation with the target user based on the second conversation information.

[0175] Based on the dialogue logic and expression of intention, through technologies such as natural language processing, speech recognition, and image recognition, the second dialogue information is obtained. The second dialogue information may include, but is not limited to, text, speech, and pictures. The content of the second dialogue information can serve the downstream task objectives according to the embedded conversation logic. For example, when the target user is a merchant suspected of illegally selling controlled knives, the dialogue task can be to determine the merchant's risk and complete evidence collection. If the expression of intention in the current reply message is "interrogative sentence, positive attitude, asking for needs", then the next action is "reply to the user and wait for the user to reply". In this case, the content of the second dialogue information can be "Hello, does the boss sell kitchen knives with a blade length of more than 24 centimeters?".

[0176] It should be noted that a major feature of the dialogue robot in this application embodiment is to master the right to speak in the dialogue. Therefore, the guidance of the dialogue topic and the posing of questions are mainly carried out by the dialogue robot (initiating user). Therefore, in order to master the right to speak in the open-domain dialogue scenario, the dialogue information (dialogue questions) thrown by the dialogue robot (initiating user) will try to narrow the reply space of the target user, such as using the form of dichotomous questions, that is, "Do you have?", "Can you?", "Do you sell?", etc. It should be noted that if the reply information of the target user jumps out of the preset dialogue logic framework, the dialogue with the target user can be directly terminated.

[0177] Generally speaking, the dialogue robot provided in this application embodiment can circle the target users (including potential sellers and buyers) based on static information, and then can actively initiate a conversation with the target users to obtain additional information; evaluate whether the target users are users who meet the dialogue tasks according to the additional information. Specifically, when the target user is a potential seller, the dialogue robot provided in this application embodiment can perform the following operations: (a) Supplement the risk identification means of the static risk control link, build an active risk discovery mechanism, and actively obtain additional information in advance for risk confirmation, identification, and disposal for high-risk goods under hidden transactions (such as controlled knives, firearms and ammunition, etc.). Reduce public opinion exposure, contain potential illegal transactions prior to supervision, and avoid being punished by interviews due to ineffective governance. (b) At major security time nodes, such as 535 (such as century-old commodity customization, banner customization, etc.), the dialogue logic can be quickly customized according to special features, push inspections to high-risk and highly suspicious merchants, discover potential risks in advance and dispose of them, and at the same time can actively trigger reminders to avoid unconscious violations by merchants. (c) Through algorithmic automated inspections, as a participant in private domain information, the algorithm is involved throughout the process, and only a limited number of merchants judged as black by the algorithm are subject to human review, thus avoiding legal risks. When the target user is a potential buyer, a conversation can be initiated with the potential buyer to obtain additional information and recommend relevant products, and the purchase needs and consumption tendencies of the potential buyer can be evaluated according to the additional information.

[0178] The technical solution provided by the embodiment of this application. The dialogue robot in this embodiment does not rely on the user's active trigger. It can actively reach the user by an algorithm to initiate a conversation. Besides the existing passive dialogue system, it supplements the active dialogue system, completes the human-computer interaction method, reduces the dependence of the dialogue system on the user's subjective initiative, expands the application scope of the dialogue system, improves the usage efficiency of the dialogue technology, and effectively realizes the autonomy and controllability of the technology application. In addition, since the dialogue robot does not need to be actively awakened by the user, the interface can not be opened to the user, and the interface can be opened as needed, realizing the full control of the dialogue robot. For example, the user cannot unilaterally trigger the dialogue robot actively, ensuring that the capabilities of the dialogue robot will not be misused and related technologies will not be stolen. Further, because the dialogue system does not need to monitor the input end, it avoids the risk of violating laws such as the protection of user personal information, thus effectively improving the security and reliability of the method.

[0179] Figure 8 It is a schematic structural diagram of an active dialogue initiation device provided by an embodiment of the present invention; refer to the attached Figure 8 As shown, this embodiment provides an active dialogue initiation device, and this active dialogue initiation device is used to execute the above Figure 1 shown active dialogue initiation method. Specifically, this active dialogue initiation device may include:

[0180] The first acquisition module 11 is used to acquire the dialogue task to be realized;

[0181] The first determination module 12 is used to determine the target user who needs to have an active dialogue based on the dialogue task;

[0182] The first determination module 12 is used to generate the initiating user for having an active dialogue with the target user and determine the dialogue information corresponding to the target user;

[0183] The first processing module 13 is used to initiate a dialogue with the target user based on the initiating user and the dialogue information, obtain the dialogue record corresponding to the target user, and complete the dialogue task based on the dialogue record.

[0184] In some instances, when the first determination module 12 determines the target user who needs to have an active dialogue based on the dialogue task, the first determination module 12 is used to execute: acquire the portrait information of the alternative users; determine the target user among the alternative users based on the dialogue task and the portrait information.

[0185] In some instances, the first processing module 13 is used to associate and store the dialogue record with the target user.

[0186] In some instances, when the first processing module 13 generates an initiating user for an active conversation with a target user, the first processing module 13 is configured to perform: obtaining anthropomorphic features for defining the initiating user; and generating a virtual initiating user for an active conversation with the target user based on the anthropomorphic features.

[0187] In some instances, the anthropomorphic features include at least one of the following: static features of the initiating user, dynamic features of the initiating user, and natural-person features of the initiating user.

[0188] In some instances, the first determination module 12 and the first processing module 13 in this embodiment are configured to perform the following steps:

[0189] The first determination module 12 is configured to determine probability information that the target user meets a conversation task based on a conversation record;

[0190] The first processing module 13 is configured to, when the probability information is greater than or equal to a preset threshold, send the identity identifier and the conversation record of the target user to a preset platform, so that the preset platform manages the target user based on the conversation record and the identity identifier.

[0191] In some instances, after initiating a conversation with the target user based on the initiating user and conversation information, the first processing module 13 in this embodiment is configured to perform: waiting to receive a reply message corresponding to the conversation information sent by the target user.

[0192] In some instances, the first obtaining module 11, the first determination module 12, and the first processing module 13 in this embodiment are configured to perform the following steps:

[0193] The first obtaining module 11 is configured to receive a reply message corresponding to the conversation information sent by the target user;

[0194] The first determination module 12 is configured to determine a conversation strategy corresponding to the target user based on the reply message;

[0195] The first processing module 13 is configured to manage the conversation with the target user based on the conversation strategy.

[0196] In some instances, when the first determination module 12 determines a conversation strategy corresponding to the target user based on the reply message, the first determination module 12 is configured to perform: obtaining a meaning representation corresponding to the reply message; and determining a conversation strategy corresponding to the target user based on the meaning representation.

[0197] In some instances, when the first determination module 12 determines the dialogue strategy corresponding to the target user based on the expression of intention, the first determination module 12 is configured to perform: when the expression of intention is the first expression, determining the dialogue strategy as the first strategy, where the first strategy is used to identify continuously conducting a dialogue with the target user based on the reply information sent by the target user; when the expression of intention is the second expression, determining the dialogue strategy as the second strategy, where the second strategy is used to identify the operation of jumping out of the dialogue with the target user; when the expression of intention is the third expression, determining the dialogue strategy as the third strategy, where the third strategy is used to identify replying to the target user a preset number of times and then jumping out of the dialogue operation with the target user; when the expression of intention is the fourth expression, determining the dialogue strategy as the fourth strategy, where the fourth strategy is used to identify waiting for the target user to send a reply message.

[0198] In some instances, when the dialogue strategy is the first strategy or the third strategy, when the first processing module 13 manages the dialogue with the target user based on the dialogue strategy, the first processing module 13 is configured to perform the following steps: generating second dialogue information based on the dialogue strategy and the expression of intention, where the second dialogue information includes at least one dialogue information corresponding to the reply information sent by the target user; conducting a dialogue with the target user based on the second dialogue information.

[0199] In some instances, when the first processing module 13 generates the second dialogue information based on the dialogue strategy and the expression of intention, the first processing module 13 is configured to perform: obtaining a database corresponding to the dialogue strategy and the expression of intention, and generating the second dialogue information based on the database and the expression of intention; and / or, obtaining a dialogue template corresponding to the dialogue strategy and the expression of intention, and generating the second dialogue information based on the dialogue template and the expression of intention; and / or, based on a machine learning model, processing the dialogue strategy and the expression of intention by the machine learning model to generate the second dialogue information.

[0200] In some instances, the first obtaining module 11, the first determination module 12, and the first processing module 13 in this embodiment are configured to perform the following steps:

[0201] The first obtaining module 11 is configured to receive the reply information sent by the target user corresponding to the dialogue information;

[0202] The first processing module 13 is configured to generate second dialogue information corresponding to the target user based on the reply information;

[0203] The first determination module 12 is configured to determine the dialogue strategy corresponding to the target user;

[0204] The first processing module 13 is configured to determine whether to conduct a dialogue with the target user based on the second dialogue information according to the dialogue strategy.

[0205] In some examples, when the first processing module 13 determines whether to have a conversation with the target user based on the second conversation information according to the conversation policy, the first processing module 13 is configured to perform: when the conversation policy is the first policy or the third policy, have a conversation with the target user based on the second conversation information; wherein, the first policy is used to indicate continuously having a conversation with the target user based on the reply information sent by the target user, and the third policy is used to indicate replying to the target user a preset number of times and then exiting the conversation with the target user; when the conversation policy is the second policy or the fourth policy, clear the second conversation information, wherein the second policy is used to indicate an operation of exiting the conversation with the target user, and the fourth policy is used to indicate waiting for the target user to send a reply message.

[0206] In some examples, after initiating a conversation with the target user based on the initiating user and the conversation information, the first obtaining module 11 and the first processing module 13 in this embodiment are configured to perform the following steps:

[0207] The first obtaining module 11 is configured to obtain the conversation state with the target user;

[0208] The first processing module 13 is configured to, when the conversation state is the end state, control the initiating user for having a conversation with the target user to prohibit receiving any information.

[0209] In some examples, the first processing module 13 in this embodiment is configured to perform the following steps: when the conversation state is the normal state, allow the initiating user to accept a third-party intervention request; manage the initiating user based on the third-party intervention request.

[0210] Figure 8 The shown device can execute Figure 1 - Figure 7 the method of the shown embodiment. For parts not described in detail in this embodiment, reference can be made to the relevant descriptions of the Figure 1 - Figure 7 shown embodiment. The execution process and technical effects of this technical solution are referred to the descriptions in the Figure 1 - Figure 7 shown embodiment and will not be elaborated here.

[0211] In a possible design, Figure 8 the structure of the shown device for initiating an active conversation can be implemented as an electronic device, which can be various devices such as a mobile phone, a tablet computer, a robot, an electronic device, a server, etc. As Figure 9 shown, the electronic device can include: a first processor 21 and a first memory 22. Wherein, the first memory 22 is used to store a program for the corresponding electronic device to execute the method for initiating an active conversation in the Figure 1 - Figure 7 shown embodiment, and the first processor 21 is configured to execute the program stored in the first memory 22.

[0212] The program includes one or more computer instructions. When the one or more computer instructions are executed by the first processor 21, the following steps can be implemented:

[0213] Obtain the dialogue task to be implemented;

[0214] Based on the dialogue task, determine the target user who needs to have an active dialogue;

[0215] Generate an initiating user for an active dialogue with the target user, and determine the dialogue information corresponding to the target user;

[0216] Initiate a dialogue with the target user based on the initiating user and the dialogue information, and obtain the dialogue record corresponding to the target user, so as to complete the dialogue task based on the dialogue record.

[0217] Furthermore, the first processor 21 is also used to execute all or part of the steps in the foregoing Figure 1 - Figure 7 illustrated embodiments.

[0218] Wherein, the structure of the electronic device may further include a first communication interface 23, which is used for the electronic device to communicate with other devices or communication networks.

[0219] In addition, an embodiment of the present invention provides a computer storage medium for storing computer software instructions used by an electronic device, which includes a program involved in the method for initiating an active dialogue in the foregoing Figure 1 - Figure 7 illustrated method embodiment.

[0220] In addition, an embodiment of the present invention provides a computer program product, including: a computer program, when the computer program is executed by a processor of an electronic device, enabling the processor to execute the steps in the foregoing Figure 1 - Figure 7 illustrated method for initiating an active dialogue.

[0221] Figure 10 is a flowchart of another method for initiating an active dialogue provided by an embodiment of the present invention; refer to the appendix Figure 10As shown in the figure, this embodiment provides another method for initiating an active conversation. The execution subject of this method can be a device for initiating an active conversation, and the device for initiating an active conversation can be implemented as software, or a combination of software and hardware. Specifically, the device for initiating an active conversation can be implemented as a wearable device. That is, at this time, the method for initiating an active conversation can be applied to a wearable device. The wearable device can be: a head-mounted display device based on Augmented Reality (AR), a head-mounted display device based on Virtual Reality (VR) or Mixed Reality (MR), a head-mounted display device based on (Hybrid Reality, HR) or Cinematic Reality (CR), etc. For example, VR glasses. Specifically, the method for initiating an active conversation can include the following steps:

[0222] Step S1001: Obtain the image information displayed by the wearable device.

[0223] Among them, when the user uses the wearable device, virtual image information can be displayed through the display module of the wearable device. The image information can be a scene image included in the virtual scene, and the scene image can be a two-dimensional scene virtual image or a three-dimensional scene virtual image. It should be noted that the image information displayed in the wearable device can change with the adjustment of the posture of the wearable device, that is, the image information displayed by the wearable device in different postures. In addition, the image information displayed by the wearable device can also change with time, that is, at different times, the image information displayed by the wearable device can also change.

[0224] Step S1002: When the image information includes a preset object or a preset area, generate a user who initiates an active conversation with the user, and determine the conversation information corresponding to the user, where the conversation information corresponds to the preset object or the preset area.

[0225] After obtaining the image information, the image information can be analyzed and processed to identify whether the image information includes a preset object or a preset area. Among them, the preset object can be an object included in the virtual scene that is pre-configured. For example, the preset object can be food, items, transportation tools, animals, plants, and buildings, etc. included in the virtual scene. The preset area can be the area where the door is located, the area where the wall is located, the area where the house is located, the area where the vehicle is located, etc. included in the virtual scene. It should be noted that in different application scenarios, different preset objects and / or preset areas can be configured.

[0226] In some instances, analyzing and processing image information to identify whether the image information includes a preset object or a preset region may include: obtaining a pre-trained machine learning model, which is trained to identify whether an image includes a preset object or a preset region. After obtaining the image information, the image information can be input into the machine learning model, so as to identify whether the image information includes a preset object or a preset region. In other instances, analyzing and processing image information to identify whether the image information includes a preset object or a preset region may include: identifying the object contour included in the image information, and analyzing and comparing the object contour with a preset standard object contour. When there is a standard object contour corresponding to the object contour, it is determined that the image information includes a preset object or a preset region; when there is no standard object contour corresponding to the object contour, it is determined that the image information does not include a preset object or a preset region.

[0227] When the image information includes a preset object or a preset region, an initiating user for initiating an active conversation with the user can be generated, and the conversation information corresponding to the user is determined. It should be noted that the conversation information corresponds to the preset object or the preset region. Specifically, the implementation manner and implementation effect of the above steps in this embodiment are similar to those of step S103 in the above embodiment. For specific reference, please refer to the description content and will not be elaborated here.

[0228] Step S1003: Initiate a conversation based on the initiating user and the conversation information to perform an active conversation operation with the user through the wearable device.

[0229] After obtaining the initiating user and the conversation information, a conversation can be initiated to the user based on the initiating user and the conversation information included in the wearable device, thus effectively realizing that an active conversation operation can be performed between the initiating user and the user. After the active conversation between the initiating user and the user is completed, a conversation record corresponding to the initiating user and the user can be obtained. The conversation record may include the conversation information sent by the initiating user and the reply information sent by the user. After obtaining the conversation record corresponding to the user, the conversation record can be analyzed and processed to complete the conversation operation based on the analysis and processing result, thus effectively ensuring the practicability of this method.

[0230] It should be noted that this embodiment may also include Figure 1 - Figure 7 the method of the embodiment shown. For parts not described in detail in this embodiment, reference can be made to the relevant description of the Figure 1 - Figure 7 embodiment shown. The execution process and technical effect of this technical solution can be seen in the description in the Figure 1 - Figure 7 embodiment shown and will not be elaborated here.

[0231] The method for initiating an active conversation provided in this embodiment obtains the image information displayed on the wearable device. When the image information includes a preset object or a preset area, it generates an initiating user for an active conversation with the user, determines the conversation information corresponding to the user, and then initiates a conversation based on the initiating user and the conversation information. This effectively realizes the operation of initiating an active conversation with the user through the wearable device without the need for the target user to perform a wake-up operation. Moreover, the initiative of the conversation is not in the hands of the user but in the hands of the initiating user, which can guide the user to have a conversation, thereby effectively improving the application scope and usage efficiency of this method. In addition, after the conversation with the user is completed, certain processing operations can be completed based on the conversation record with the user, further improving the practicality of this method and facilitating market promotion and application.

[0232] Figure 11 It is a schematic structural diagram of a device for initiating an active conversation provided by an embodiment of the present invention; refer to the appendix Figure 11 As shown, this embodiment provides a device for initiating an active conversation, and this device can execute the above Figure 10 corresponding method for initiating an active conversation. In some instances, the device for initiating an active conversation can be implemented as a wearable device. Specifically, this device may include:

[0233] A second acquisition module 31, configured to acquire the image information displayed on the wearable device;

[0234] A second generation module 32, configured to generate an initiating user for an active conversation with the user and determine the conversation information corresponding to the user when the image information includes a preset object or a preset area, and the conversation information corresponds to the preset object or the preset area;

[0235] A second processing module 33, configured to initiate a conversation based on the initiating user and the conversation information to perform an active conversation operation with the user through the wearable device.

[0236] Figure 11 The device shown can execute Figure 10 the method of the embodiment shown. For parts not described in detail in this embodiment, reference can be made to the relevant descriptions of Figure 10 the embodiment shown. The execution process and technical effects of this technical solution are referred to the descriptions in Figure 10 the embodiment shown and will not be elaborated here.

[0237] In a possible design, Figure 11 the structure of the device for initiating an active conversation shown can be implemented as an electronic device, and this electronic device can be various devices such as a mobile phone, a tablet computer, a robot, an electronic device, a server, a wearable device, etc. As Figure 12As shown, the electronic device may include: a second processor 41 and a second memory 42. Among them, the second memory 42 is used to store a program for the corresponding electronic device to execute the method of initiating an active conversation in the above Figure 10 shown embodiment, and the second processor 41 is configured to execute the program stored in the second memory 42.

[0238] The program includes one or more computer instructions. When the one or more computer instructions are executed by the second processor 41, the following steps can be implemented:

[0239] Obtain the image information displayed by the wearable device;

[0240] When the image information includes a preset object or a preset area, an initiating user for initiating an active conversation with the user is generated, and the conversation information corresponding to the user is determined. The conversation information corresponds to the preset object or the preset area;

[0241] Initiate a conversation based on the initiating user and the conversation information to perform an active conversation operation with the user through the wearable device.

[0242] Furthermore, the second processor 41 is also used to execute all or part of the steps in the foregoing Figure 10 shown embodiment.

[0243] Among them, the structure of the electronic device may further include a second communication interface 43 for the electronic device to communicate with other devices or communication networks.

[0244] In addition, an embodiment of the present invention provides a computer storage medium for storing computer software instructions used by the electronic device, which includes a program involved in executing the method of initiating an active conversation in the Figure 10 shown method embodiment.

[0245] In addition, an embodiment of the present invention provides a computer program product, including: a computer program, when the computer program is executed by the processor of the electronic device, the processor is caused to execute the steps in the method of initiating an active conversation as described above Figure 10 shown.

[0246] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0247] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of adding a necessary general hardware platform, and of course, it can also be implemented by a combination of hardware and software. Based on such an understanding, the above technical solution, in essence, or the part that contributes to the prior art can be embodied in the form of a computer product. The present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) that contain computer-usable program codes.

[0248] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0249] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0250] These computer program instructions can also be loaded onto a computer or other programmable device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0251] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0252] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM) and / or non-volatile memory such as read-only memory (ROM) or flash RAM. The memory is an example of computer-readable media.

[0253] Computer-readable media includes both permanent and non-permanent, removable and non-removable media implemented by any method or technology for information storage. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transitory media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0254] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. However, such modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for initiating an active conversation, characterized in that, it includes: Obtain the conversation task to be implemented; Based on the conversation task, determine the target user who needs to have an active conversation. When the conversation task is used to determine users with purchase needs, the target user is determined by purchase propensity; when the conversation task is used to determine users with illegal operations, the target user is determined by selling behavior; Generate an initiating user for having an active conversation with the target user, and determine the conversation information corresponding to the target user; Initiate a conversation with the target user based on the initiating user and the conversation information, and obtain the conversation record corresponding to the target user, so as to complete the conversation task based on the conversation record; The method further includes: Based on the conversation record, determine the probability information that the target user meets the conversation task; When the probability information is greater than or equal to a preset threshold, send the identity identifier of the target user and the conversation record to a preset platform, so that the preset platform manages the target user based on the conversation record and the identity identifier.

2. The method according to claim 1, characterized in that, Determining the target user who needs to have an active conversation based on the conversation task includes: Obtain the portrait information of alternative users; Based on the conversation task and the portrait information, determine the target user among the alternative users.

3. The method according to claim 1, characterized in that, Generating an initiating user for having an active conversation with the target user includes: Obtain the anthropomorphic features for defining the initiating user; Based on the anthropomorphic features, generate a virtual initiating user for having an active conversation with the target user.

4. The method according to claim 1, characterized in that, After initiating a conversation with the target user based on the initiating user and the conversation information, the method further includes: Wait to receive the reply information corresponding to the conversation information sent by the target user.

5. The method according to claim 4, characterized in that, The method further includes: Receive the reply information corresponding to the conversation information sent by the target user; Based on the reply information, determine the conversation strategy corresponding to the target user; Manage the conversation with the target user based on the conversation strategy.

6. The method according to claim 5, characterized in that, Determining the conversation strategy corresponding to the target user based on the reply information includes: Obtain the meaning representation corresponding to the reply information; Based on the meaning representation, determine the conversation strategy corresponding to the target user.

7. The method according to claim 6, characterized in that, Determining the conversation strategy corresponding to the target user based on the meaning representation includes: When the meaning representation is the first representation, determine the conversation strategy as the first strategy, and the first strategy is used to indicate continuing the conversation with the target user based on the reply information sent by the target user; When the indication of intention is the second indication, determine that the conversation strategy is the second strategy, and the second strategy is used to identify the operation of jumping out of the conversation with the target user; When the indication of intention is the third indication, determine that the conversation strategy is the third strategy, and the third strategy is used to identify replying to the target user a preset number of times and then jumping out of the conversation operation with the target user; When the indication of intention is the fourth indication, determine that the conversation strategy is the fourth strategy, and the fourth strategy is used to identify waiting for the target user to send a reply message.

8. The method according to claim 7, wherein, when the conversation strategy is the first strategy or the third strategy, manage the conversation with the target user based on the conversation strategy, including: generate second conversation information based on the conversation strategy and the indication of intention, where the second conversation information includes at least one conversation information corresponding to the reply message sent by the target user; carry out a conversation with the target user based on the second conversation information.

9. The method according to claim 8, wherein, generating second conversation information based on the conversation strategy and the indication of intention includes: obtain a database corresponding to the conversation strategy and the indication of intention, and generate the second conversation information based on the database and the indication of intention; and / or, obtain a conversation template corresponding to the conversation strategy and the indication of intention, and generate the second conversation information based on the conversation template and the indication of intention; and / or, based on a machine learning model, process the conversation strategy and the indication of intention by the machine learning model to generate the second conversation information.

10. The method according to claim 4, wherein, the method further includes: receive a reply message sent by the target user corresponding to the conversation information; generate second conversation information corresponding to the target user based on the reply message; determine a conversation strategy corresponding to the target user; judge whether to carry out a conversation with the target user based on the second conversation information according to the conversation strategy.

11. The method according to claim 10, wherein, judging whether to carry out a conversation with the target user based on the second conversation information according to the conversation strategy includes: when the conversation strategy is the first strategy or the third strategy, carry out a conversation with the target user based on the second conversation information; wherein, the first strategy is used to identify continuously carrying out a conversation with the target user based on the reply message sent by the target user, and the third strategy is used to identify replying to the target user a preset number of times and then jumping out of the conversation with the target user; when the conversation strategy is the second strategy or the fourth strategy, clear the second conversation information, wherein the second strategy is used to identify the operation of jumping out of the conversation with the target user, and the fourth strategy is used to identify waiting for the target user to send a reply message.

12. The method according to claim 11, wherein, the method further includes: when the conversation state is the normal state, allow the initiating user to accept a third-party intervention request; Manage the initiating user based on the third-party intervention request.

13. An electronic device, characterized in that it includes: a memory and a processor; wherein, the memory is used to store one or more computer instructions, and when the one or more computer instructions are executed by the processor, the method for initiating an active conversation according to any one of claims 1-12 is implemented.

Images