Message processing method, electronic device, storage medium, and computer program product

Abstract

Embodiments of the present application provide a message processing method, an electronic device, a storage medium and a computer program product. The message processing method comprises: determining a target user conversation message in a group chat conversation window according to a task conversation message received by an artificial intelligence assistant in the group chat conversation window; determining a reference conversation message having a reference relationship with the target user conversation message; generating prompt information according to the target user conversation message and the reference conversation message; and sending the prompt information to a generative base model to enable the generative base model to perform task reply generation processing based on the prompt information. The technical solution of the embodiments of the present application can obtain prompt information according to the task conversation message, so that the generative base model can accurately implement task reply processing.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a message processing method, electronic device, storage medium, and computer program product. Background Technology

[0002] With the widespread use of social media and online collaboration platforms, group chats have become an important form of daily communication. To improve the user experience in group chats, solutions combining artificial intelligence with group chat message processing have emerged. For example, generative basic models such as large language models can be used to provide task responses to user requests in various application scenarios. However, these technical solutions often face problems such as chaotic multi-threading management and inaccurate context understanding when processing conversational messages in group chat scenarios, leading to a decline in user experience. Therefore, providing a new message processing solution for group chat scenarios has become a technical problem that needs to be solved. Summary of the Invention

[0003] This application provides a message processing scheme to at least partially solve the above-mentioned problems.

[0004] According to a first aspect of the embodiments of this application, a message processing method is provided, comprising: determining a target user session message in a group chat session window based on a task session message received by an AI assistant in the group chat session window; determining a reference session message that has a reference relationship with the target user session message; generating a prompt message based on the target user session message and the reference session message; and sending the prompt message to a generative basic model so that the generative basic model performs task response generation processing based on the prompt message.

[0005] According to a second aspect of the embodiments of this application, an electronic device is provided, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; the memory is used to store a computer program; and the processor is used to execute the message processing method described in the first aspect by running the computer program stored in the memory.

[0006] According to a third aspect of the embodiments of this application, a computer storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the message processing method as described in the first aspect.

[0007] According to a fourth aspect of the embodiments of this application, a computer program product is provided, including a computer program that, when executed by a processor, implements the message processing method as described in the first aspect.

[0008] According to the message processing scheme provided in the embodiments of this application, the target user's conversation message in the group chat window can be determined based on the task conversation message received by the AI ​​assistant. Prompt information is then generated based on the target user's conversation message and referenced conversation messages that have a reference relationship with the target user's conversation message. This allows for task response generation in multi-threaded chat sessions within a group chat, enabling the targeted acquisition of relevant conversation messages for a specific task conversation message to accurately generate prompt information. In complex group chat environments, it can more accurately capture information interactions between users, effectively supporting intelligent communication among multiple roles and users, reducing the adverse effects caused by the chaos of multi-threaded conversations, and improving the accuracy of the generated prompt information in instructing tasks. This allows the generative basic model to more accurately understand the task context targeted by the prompt information when processing task response generation based on the prompt information, resulting in more accurate task response generation and processing results that better meet task requirements, thereby improving user experience. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0010] Figure 1 This is a schematic diagram of a message processing system according to an embodiment of this application.

[0011] Figure 2 This is a flowchart of the steps of a message processing method according to an embodiment of this application.

[0012] Figure 3 This is a flowchart illustrating some optional steps for determining the reference session message in an embodiment of this application.

[0013] Figure 4 This is a flowchart illustrating some optional steps for determining the prompt information in the embodiments of this application.

[0014] Figure 5 This is an overall flowchart of an example message processing scheme according to an embodiment of this application.

[0015] Figure 6A and Figure 6B The diagram illustrates some examples of message processing schemes according to embodiments of this application.

[0016] Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in the embodiments of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art should fall within the protection scope of the embodiments of this application.

[0018] The specific implementation of the embodiments of this application will be further described below with reference to the accompanying drawings.

[0019] Figure 1 An exemplary system applicable to embodiments of this application is shown. For example... Figure 1 As shown, the system 100 may include a cloud server 102, a communication network 104, and / or one or more user devices 106. Figure 1 The example in the text is 106 user devices.

[0020] Optionally, user device 106 may include any one or more user devices suitable for displaying information, interacting with users, etc. In some embodiments, user device 106 may include any suitable type of device. For example, in some embodiments, user device 106 may include mobile devices, tablet computers, laptop computers, desktop computers, and / or any other suitable type of user device. Optionally, user device 106 may be equipped with an application capable of group chat sessions, for example, the application may be an instant messaging platform application.

[0021] Optionally, when the user equipment 106 implements the scheme of the embodiments of this application, in some embodiments, the user equipment 106 can be used to execute a message processing method. As an optional example, in some embodiments, the user equipment 106 can first determine the target user session message in the group chat session window based on the task session message received by the AI ​​assistant in the group chat session window; then, it can determine the reference session message that has a reference relationship with the target user session message; then, it can generate a prompt message based on the target user session message and the reference session message; and then send the prompt message to the generative basic model so that the generative basic model can perform task response generation processing based on the prompt message. Optionally, the generative basic model can be set on the cloud server 102, and the user equipment 106 can send the generated prompt message to the cloud server 102 so that the task response generation processing can be performed through the generative basic model.

[0022] In some embodiments, the communication network 104 can be any suitable combination of one or more wired and / or wireless networks. For example, the communication network 104 can include any one or more of the following: the Internet, an intranet, a wide area network (WAN), a local area network (LAN), a wireless network, a digital subscriber line (DSL) network, a frame relay network, an asynchronous transfer mode (ATM) network, a virtual private network (VPN), and / or any other suitable communication network. The user equipment 106 can be connected to the communication network 104 via one or more communication links (e.g., communication link 112), and the communication network 104 can be linked to the cloud server 102 via one or more communication links (e.g., communication link 114). The communication link can be any communication link suitable for transmitting data between the user equipment 106 and the cloud server 102, such as a network link, a dial-up link, a wireless link, a hardwired link, any other suitable communication link, or any suitable combination of such links.

[0023] The cloud server 102 can be any suitable device for storing information, data, programs, and / or any other suitable type of content, including but not limited to distributed storage system devices, server clusters, computing cloud server clusters, etc. In some embodiments, the cloud server 102 can perform any suitable function. In some embodiments, the cloud server 102 can send the result of the task response generation processing of the generative basic model to the user device 106 so that the user device 106 can display the result. For example, it can be displayed through a group chat window of the user device 106.

[0024] Based on the above system, this application provides a message processing scheme, which will be described below through several embodiments.

[0025] Figure 2 This is a flowchart illustrating the steps of a message processing method according to an embodiment of this application. According to a first aspect of an embodiment of this application, a message processing method is provided, referring to... Figure 2 As shown, the method consists of the following steps:

[0026] S202: Based on the task session message received by the AI ​​assistant in the group chat session window, determine the target user session message in the group chat session window.

[0027] In this embodiment, the group chat window can be a group chat window within an instant messaging platform, where users can engage in various activities such as chatting, information display, and content interaction through chat messages. For example, any user can send a chat message in the group chat window to communicate with other users on a specific topic. With the development of artificial intelligence (AI) technology, instant messaging platforms are introducing AI through methods such as AI assistants to interact with users and assist them in handling various tasks. However, currently, the interaction between AI, such as AI assistants, and users is mostly one-to-one, meaning there is no third user in the user's chat window with the AI ​​assistant. While this provides AI functionality, it fails to meet the needs of multiple users using the AI ​​assistant in the same chat environment. Furthermore, the efficiency of the AI ​​assistant is low in one-to-one interactions, preventing it from fully utilizing its functions.

[0028] Therefore, in this embodiment, an AI assistant is introduced into the group chat window (typically one AI assistant per group chat window) to provide services to multiple users simultaneously within the current group chat window. The AI ​​assistant can not only interact with users in non-task-related ways, such as regular chat, but also perform tasks or provide services required by users, including but not limited to data statistics, task or service information display, and task execution based on user instructions. Based on this, whether a user in the group chat window instructs the AI ​​assistant to perform a task or provide a service, or the AI ​​assistant automatically triggers the execution of a task or the provision of a service based on triggering conditions, the AI ​​assistant will receive the corresponding task session message. Then, the target user session message in the group chat window is determined. For example, it is necessary to identify user session messages among multiple user session messages that are related to the AI ​​assistant's task topic, so that the AI ​​assistant can perform the task (e.g., by generating the corresponding response result through a generative basic model in the background).

[0029] AI assistants can be of any type, and their functions can be selected as needed to perform a variety of tasks. For example, in some embodiments, an AI assistant can assist users in data statistics; it can also assist users in generating articles; it can also assist users in generating lyrics; and it can also assist users in generating a piece of music. Alternatively, AI assistants can be used for other functions; there is no single limitation.

[0030] Optionally, the AI ​​assistant can combine one or more machine learning models to achieve its functions. For example, in some embodiments, the aforementioned machine learning models may include generative base models. In the embodiments of this application, the generative base model may be a Large Language Model (LLM), which can be a type of natural language processing model with a large number of parameters. After being trained on large-scale data, it can understand and generate complex language text and is suitable for various application scenarios, including but not limited to dialogue systems, text summarization, and translation. Optionally, the generative base model can be deployed on the server side, and the AI ​​assistant can be deployed on the client side.

[0031] Task conversation messages can be messages received by the AI ​​assistant that request it to perform a specific task. For example, a user can send a task conversation message to the AI ​​assistant, or it can be sent through other means (such as automatic sending by the system after pre-setting). Taking user-sent task conversation messages as an example, users can send them to the AI ​​assistant in any way. For instance, a user can first activate the AI ​​assistant in a group chat window and then input the task conversation message through typing, voice input, or other methods. In some feasible scenarios, the AI ​​assistant can provide an input field for a one-on-one dialogue with the user, through which the user can input task conversation messages.

[0032] Optionally, after the AI ​​assistant receives the task session message, it can determine the target user session message from among the multiple existing session messages in the group chat window. For example, the earliest chronologically relevant user session message among the multiple user session messages can be identified as the target user session message. Alternatively, each existing user session message can be analyzed individually; if the semantic similarity between the user session message and the task session message reaches a predetermined threshold, then the user session message is considered relevant to the task indicated by the task session message. Another example is identifying task-related keywords from the task session message, then performing keyword matching among multiple user session messages to find user session messages with semantically similar or identical keywords, and identifying the earliest chronologically relevant user session message as the target user session message. Alternatively, other methods can be used to determine the target user session message, all of which are applicable to the solutions in this application embodiment.

[0033] To illustrate, consider this example: a user activates an AI assistant in a group chat window and enters a task message such as "Please help me count the number of people planning to participate in team building activities." Upon receiving this task message, the AI ​​assistant can determine the target user's session message from among the existing user session messages in the group chat window. For instance, if user session message A is "Those participating in team building activities, please let us know in the group," the AI ​​assistant can match the word "team building" in user session message A based on the task message. If user session message A is the earliest session in the sequence, then user session message A can be identified as the target user's session message. It should be understood that this example is not intended to limit the scope of the embodiments of this application.

[0034] S204: Identify the referenced session message that has a reference relationship with the target user's session message.

[0035] In this embodiment, by identifying the referencing session messages that have a reference relationship with the target user's session messages, the session messages related to the target user's session messages can be accurately and quickly determined, facilitating the accurate generation of subsequent prompt information. This makes the solution effectively adaptable to scenarios involving multi-threaded group chat sessions.

[0036] Generally, "quoting" a conversation message means referencing a specific conversation message in a group chat window as a reference or basis for a reply. Thus, in a multi-person group chat window, "quoting" a conversation message clearly indicates which message the user is replying to. However, this is not the only exception. In this embodiment, expressions of agreement or support based on a particular conversation message are also included in the scope of "quoting," such as the "OK" symbol or icon, or the "+1" symbol or icon posted by the user on a conversation message. For example, referencing a session message may include at least one of the following: user session messages that are referenced directly using the referencing function (e.g., in some optional examples, a user can long-press the target user session message and reference the message from the pop-up referencing function key), user session messages that explicitly indicate the message recipient as the publisher of the target user session message (e.g., in some optional examples, a user can send a message by @ the publisher of the target user session message to set the message recipient of the user session message as the publisher of the target user session message), and other reply session messages based on the target user session message (e.g., a user can send "+1", "-1", "OK", etc. to express a reply to the target user session message).

[0037] In some alternative embodiments, refer to Figure 3 As shown in the flowchart, step S204 determines the referencing session message through the following steps S2042 and S2044:

[0038] S2022: From the context conversation messages of the target user conversation message in the group chat conversation window, determine at least one user conversation message that belongs to the same reference chain as the target user conversation message.

[0039] As mentioned earlier, multiple conversation messages can exist in a group chat window. When identifying a target user's conversation message, the referencing chain is determined from the context conversation messages within the group chat window. The context conversation messages are those preceding and following the target user's conversation message.

[0040] Within the context of the session messages, the user session messages that reference the target user session message can be identified. For example, to illustrate, a user sends a target user session message, which is then referenced by other users. For instance, user session message 1 references the target user session message, user session message 2 references user session message 1 (equivalent to user session message 2 referencing the target user session message through referencing user session message 1), user session message 3 also references the target user session message, user session message 4 references user session message 2 (equivalent to user session message 2 referencing the target user session message through referencing both user session messages 2 and user session message 1), and so on. In this case, user session messages 1 through 4 can all be considered user session messages that reference the target user session message. Similar cases can be deduced in this way. It is understood that these user session messages and the target user session message belong to the same reference chain and have a mutual referencing relationship.

[0041] S2024: Determine the reference session message based on at least one identified user session message.

[0042] The reference session message can be determined by identifying at least one user session message belonging to the same reference chain as the target user session message, as determined in step S2022. For example, at least one user session message can be directly identified as the reference session message; alternatively, at least one user session message can be further processed, and the result of the processing can be identified as the reference session message.

[0043] Based on this, in the embodiments of this application, the optional solutions of steps S2022 to S2044 above can accurately and quickly determine the referenced session message by determining at least one user session message that belongs to the same reference chain as the target user session message, and then using the determined at least one user session message, so as to generate prompt information accurately and quickly in the future.

[0044] The specific implementation of step S2024 is not specifically limited in the embodiments of this application. In some optional embodiments, step S2024 may include: performing semantic analysis on at least one user session message, and determining the user session message that satisfies a first semantic similarity with the target user session message as a reference session message.

[0045] In some practical scenarios, although step S2022 has identified at least one user session message with a reference relationship to the target user session message, not all of these identified user session messages are necessarily related to the task targeted by the task session message. Some session messages may still be considered as interference. For example, continuing with the example mentioned earlier that "the target user session message is user session message A, 'Those who are participating in team building, please let us know in the group,'" suppose a user session message B identified as having a reference relationship to user session message A is "Are you going to the internet cafe to play games after get off work today?". Obviously, user session message B is unrelated to "team building" and also unrelated to the task of "counting the number of people who plan to participate in team building." Therefore, it can be considered as interference and removed from the list of reference session messages. The above-mentioned optional method in this application embodiment can accurately identify reference session messages and accurately exclude interference messages by performing semantic analysis on at least one user session message and then determining whether the user session message is a user session message with a first semantic similarity to the target user session message.

[0046] Therefore, in this embodiment of the application, semantic analysis is performed on these determined user session messages, and user session messages that meet the first semantic similarity with the target user session message are identified as reference session messages, while user session messages that do not meet the first semantic similarity are discarded. This can reduce the interference of messages that are not related to the task, making the subsequently generated prompt information more accurate and more conducive to the achievement of the task.

[0047] Optionally, satisfying the first semantic similarity can be greater than or equal to a preset first semantic similarity threshold. That is, when the semantic similarity between a determined user session message and a target user session message is greater than or equal to the first semantic similarity threshold, the two are considered to satisfy the first semantic similarity, and the user session message can be identified as a user session message and used as a reference session message. It should be understood that the first semantic similarity threshold can be set as needed and is not specifically limited here. For example, it can be 80%, 85%, 90%, 95%, etc.

[0048] In some optional embodiments, semantic similarity can be calculated using vector computation methods. For example, the target user session message can be vectorized to obtain the vector corresponding to the target user session message, and then the user session message can be vectorized to obtain the vector corresponding to the user session message. The semantic similarity between the two vectors can be calculated, and thus, based on the semantic similarity between the two vectors, it can be determined whether the user session message and the target user session message satisfy a first semantic similarity. It should be understood that in any optional scheme in the embodiments of this application, semantic similarity can be implemented using any similarity index, and no specific restrictions are imposed here.

[0049] By using the first semantic similarity, we can further filter and identify session messages that are closely related to the task indicated by the task session message, so as to provide an accurate basis for subsequent task processing.

[0050] Regarding the determination of referenced session messages, as mentioned above, although it can be determined by whether a referencing function or an indicator is used, in some cases, it can also be determined by other methods. For example, in some optional embodiments, the above-mentioned semantic analysis of at least one user session message, and the determination of user session messages that meet the first semantic similarity with the target user session message as referenced session messages, can be implemented as follows: determining whether at least one user session message contains preset characters and / or preset symbols used to express the same meaning as the target user session message; and determining user session messages containing preset characters and / or preset symbols as referenced session messages.

[0051] In actual group chat sessions, when the first user wants to express a message with the same meaning as a message posted by the second user, they don't need to type the same message content as the second user's message. Instead, they can use a more concise form to express the same meaning, such as preset characters and / or preset symbols. For example, preset characters might be "+1" or "OK," and preset symbols might be symbols representing "+1" or "OK." It's understood that in this case, the actual semantics of the message posted by the first user and the message posted by the second user are essentially the same (for example, in this optional embodiment, "essentially the same semantics" can be understood as the semantic similarity meeting a first semantic similarity threshold, such as a semantic similarity greater than or equal to a first semantic similarity threshold, for example, 95%, 96%, 97%, 98%, 99%, 100%, etc.). Therefore, in some embodiments, preset characters and / or preset symbols in the user chat message used to express the same meaning as the target user's chat message can be identified as referenced chat messages.

[0052] For example, if user A posts "Weekend team building registration" as the target user's conversation message, and then user B posts "+1", user C also posts "+1", and user D posts an "OK" emoticon, then all three user conversation messages posted by users B, C, and D belong to the same reference chain as the target user's conversation message. Here, "+1" can be considered a preset character, and the "OK" emoticon can be considered a preset symbol. It can be determined that the messages posted by users B, C, and D all include preset characters and / or preset symbols used to express the same meaning as the target user's conversation message (i.e., participating in team building). Therefore, these three user conversation messages can all be identified as reference conversation messages. It should be understood that the above is only an example; preset characters and / or preset symbols can be set as needed, and this application embodiment does not impose specific limitations.

[0053] It should be understood that in the embodiments of this application, by determining whether the user session message contains preset characters and / or preset symbols used to express the same meaning as the target user session message, and then determining the user session message containing the preset characters and / or preset symbols as the reference session message, it is possible to avoid the user session message with the same meaning as the target user session message being ignored, thereby more accurately and comprehensively determining the reference session message, so as to accurately and comprehensively determine the prompt information, and enable the generative basic model to accurately perform task response generation processing.

[0054] It should be noted that in practical applications, user session messages in the referencing chain may also contain characters such as "-1" or symbols or other characters with meanings different from the target user session message. These user session messages can be effectively filtered out using the first semantic similarity test, thereby improving the efficiency of subsequent task processing.

[0055] In another feasible approach, as described above, the referenced session messages can be determined based on those user session messages that carry an identifier. Specifically, in some optional embodiments, the aforementioned step S2022, which determines from the context session messages of the target user session message in the group chat session window that belong to at least one user session message in the same reference chain as the target user session message, can be implemented as follows: from the context session messages of the target user message in the group chat session window, determine at least one user session message corresponding to the target user session message that carries an identifier for specifying a user, wherein the identifier for specifying a user is an identifier used to indicate the user to whom the user session message needs to be delivered.

[0056] For example, the identifier used to specify a user can be, but is not limited to, text identifiers such as @ and #. Taking @ as an example, when @ing a user's username in a group chat window, the sent user chat message can be specifically delivered to that user. As another example, the identifier used to indicate a user can also be any identifier that can indicate a user, such as an icon.

[0057] It should be understood that by identifying at least one user session message that corresponds to the target user session message and carries an identifier for the specified user in the context message, it is possible to accurately and conveniently identify at least one user session message that belongs to the same reference chain as the target user session message, so as to accurately and conveniently identify the referenced session message.

[0058] In this case, optionally, the aforementioned step S2024, which determines the reference session message based on the determined at least one user session message, can be determined in the following manner: performing semantic analysis on at least one user session message corresponding to the target user session message and carrying an identifier for the specified user; and determining the user session message whose semantic similarity to the target user session message satisfies a first semantic similarity based on the semantic analysis results.

[0059] In this embodiment of the application, semantic analysis is performed on at least one user session message that corresponds to the target user session message and carries an identifier for specifying the user. Then, user session messages whose semantic similarity with the target user session message meets the first semantic similarity are identified as reference session messages, while user session messages that do not meet the first semantic similarity are discarded. This can reduce the interference of messages that are not related to the task, making the subsequently generated prompt information more accurate and more conducive to the achievement of the task.

[0060] For example, suppose user A posts a "Weekend team building registration" message as the target user's conversation message. User B replies to the target user's conversation message using the referencing function, forming a conversation message like "@userA, I'm attending." User C also replies to the target user's conversation message using the referencing function, forming a conversation message like "@userA, I can't attend this weekend." Although both user B's and user C's messages carry the identifier "@" to specify the user, it's clear that user B's conversation message and the target user's conversation message satisfy the first semantic similarity requirement, while user C's conversation message does not. Therefore, user B's conversation message will be identified as the final referencing conversation message. This method can filter out valid referencing conversation messages, improving the efficiency of subsequent task processing.

[0061] S206: Generate a prompt message based on the target user's session message and the referenced session message.

[0062] The prompt information can be used to process the task response generation for the generative base model, so as to obtain the generation result that meets the task requirements.

[0063] The specific implementation of step S206 is not limited in the embodiments of this application. In some optional embodiments, refer to Figure 4 In the flowchart shown, step S206 can determine the prompt information through the following steps S2062 and S2064:

[0064] S2062: According to the preset weight information, at least the target user session messages and the referenced session messages are weighted.

[0065] In this embodiment of the application, corresponding weights can be assigned to the target user session message and the reference session message respectively, and at least the two can be weighted or merged.

[0066] Weighting information can be set as needed. For example, in some embodiments, the weight corresponding to the target user's session message can be set to be greater than the weight corresponding to the referencing session message. This ensures that the target user's session message is retained and identified more effectively, while the referencing session message can mainly serve as a reference message, making the subsequently generated prompts more accurate and reliable, and enabling the generative base model to perform task response generation more accurately.

[0067] For example, the weight of the target user's session message can be set to 60%, while the weight of the reference session message can be set to 40%.

[0068] S2064: Generate a prompt message based on the weighted target user session message and the referenced session message.

[0069] Optionally, the weighted processing result of the target user's session message and the referenced session message can be used to generate a prompt message in a predetermined format to meet the requirements.

[0070] Based on this, the implementation of steps S2062 to S2064 in this application embodiment can effectively perform weighted processing based on the target user session message and the reference session message, and conveniently generate prompt information that meets the requirements, so as to facilitate the subsequent task response generation processing of the generative basic model.

[0071] In some optional embodiments, the referenced session messages determined in step S204 include multimodal messages. For example, the multimodal messages may include at least one of image messages, rich text messages, merged session history messages, and group chat card messages.

[0072] Therefore, the embodiments of this application can adapt to the processing of various multimodal messages, and can facilitate the accurate and comprehensive determination of prompt information, so that the subsequent generative basic model can accurately perform task response generation processing.

[0073] Optionally, step S206 may include: processing the multimodal message to obtain the multimodal message processing result; and generating a prompt message based on the target user message, other referencing session messages besides the multimodal message, and the multimodal message processing result.

[0074] It should be understood that by processing multimodal messages, prompt information can be accurately and comprehensively generated based on the target user's session messages, other reference session messages besides multimodal messages, and the processing results of multimodal messages, enabling the subsequent generative base model to accurately perform task response generation processing.

[0075] Optionally, different processing methods can be used for different types of multimodal messages to obtain the required multimodal processing results.

[0076] For example, if the multimodal message includes an image message, the link corresponding to the image message is obtained, and the link is converted into a first multimodal processing result in a predetermined format. Optionally, the predetermined format can be set to a format readable by the generative base model.

[0077] It should be understood that in this embodiment of the application, by converting the link corresponding to the image message into a first multimodal processing result in a predetermined format, the prompt information generated subsequently based on the first multimodal message processing result can be better identified and processed by the generative basic model, enabling the subsequent generative basic model to accurately perform task response generation processing.

[0078] For example, if the multimodal message includes a rich text message, the rich text message is decomposed to obtain a second multimodal message processing result containing text information and image information.

[0079] Rich text is a text format that includes formatted information. Compared to plain text, rich text includes not only text content but also text style, size, font, color, and embedded multimedia elements. Rich text enhances the expressiveness of text, intuitively conveys user intent, and improves readability.

[0080] It should be understood that by splitting the rich text message in this embodiment, a second multimodal message processing result containing text and image information can be effectively obtained. The prompt information generated based on the second multimodal message processing result can be better recognized and processed by the generative basic model, enabling the subsequent generative basic model to accurately perform task response generation processing.

[0081] For example, if the multimodal message includes a merged session record message, the merged session record message is decomposed to obtain a third multimodal message processing result containing multiple session records from the merged session record message.

[0082] It should be understood that in this embodiment of the application, splitting the merged session record message can effectively obtain a third multimodal message processing result containing multiple session records in the merged session record message. The prompt information generated based on the third multimodal message processing result can be better identified and processed by the generative basic model, enabling the subsequent generative basic model to accurately perform task response generation processing.

[0083] For example, if the multimodal message includes group chat card messages, then the event tracking data is generated based on the group chat card messages to obtain the fourth multimodal message processing result containing the event tracking data.

[0084] In this embodiment, a group chat card message is a card message that displays certain information in the group chat window, and the card can contain any content. Optionally, the group chat card message can be pushed to the group chat window by an AI assistant for display. In some cases, the group chat window may include one or more group chat card messages.

[0085] In this embodiment, "embedded data" refers to capturing, processing, and sending relevant data in group chat card messages by setting data collection points (i.e., "embedded points") within the messages. By generating embedded data based on the group chat card messages, relevant data within the messages can be captured to generate a fourth multimodal message processing result, which can then be used to generate prompt information based on the fourth multimodal message processing result. This allows the generative base model to better identify and process the data, enabling the subsequent generative base model to accurately generate task responses.

[0086] It should be understood that the above-mentioned processing methods for image messages, rich text messages, merged conversation record messages, and group chat card messages can be selected as needed, and no specific limitation is made here.

[0087] In some optional embodiments, the message processing method of this application further includes: obtaining, from the group chat window where the target user's session message is located, session messages whose publication time is within a preset time range, wherein the publication time of both the target user's session message and the referencing session message is within the preset time range; determining, from the obtained session messages, other messages besides the referencing session message; and determining, from the determined other messages, at least one similar session message that satisfies a second semantic similarity with the target user's session message. Optionally, step S206 may include: generating a prompt message based on the target user's session message, the referencing session message, and at least one similar session message.

[0088] In some cases, user conversation messages sent by some users in a group chat window may not be sent in various referencing formats (neither containing identifiers for specific users nor preset characters or symbols). For example, taking user conversation message A, "Those who want to participate in the team building activity, please let us know in the group," as an example, suppose a user conversation message C, which is determined to be unrelated to user conversation message A, is "Please leave a message as soon as possible if other students who want to participate in the team building activity do so." Obviously, user conversation message C is related to "team building" and also to the task of "counting the number of people who plan to participate in the team building activity." Therefore, it can be recalled to generate a notification message.

[0089] Therefore, this application also introduces a semantic relevance recall mechanism. This mechanism can recall at least one similar session message that satisfies a second semantic similarity with the target user's session message, thereby generating prompt information. It should be understood that the prompt information generated using the target user's session message, the referenced session message, and at least one similar session message is more accurate and comprehensive, enabling the subsequent generative base model to accurately perform task response generation processing.

[0090] Optionally, the second semantic similarity can be greater than or equal to a preset second semantic similarity threshold. That is, when the semantic similarity between a determined user session message and a target user session message is greater than or equal to the second semantic similarity threshold, they are considered to meet the second semantic similarity requirement, and the user session message can be identified as a similar session message. It should be understood that the second semantic similarity threshold can be set as needed and is not specifically limited here; for example, it could be 80%, 85%, 90%, 95%, etc.

[0091] In some optional embodiments, semantic similarity can be calculated using vector computation methods. For example, the target user session message can be vectorized to obtain the vector corresponding to the target user session message, and then the user session message can be vectorized to obtain the vector corresponding to the user session message. The semantic similarity between the two vectors can be calculated, and thus, based on the semantic similarity between the two vectors, it can be determined whether the user session message and the target user session message satisfy a second semantic similarity. It should be understood that in any optional scheme in the embodiments of this application, semantic similarity can be implemented using any similarity index, and no specific restrictions are imposed here.

[0092] As can be seen from the above, in one example, the message ultimately used to generate the prompt message may include, in addition to the target user session message, at least one of the following: a similar session message; a reference session message containing preset characters and / or preset symbols for expressing the same meaning as the target user session message; and a reference session message corresponding to the target user session message, carrying an identifier for specifying the user, and whose semantic similarity to the target user session message satisfies a first semantic similarity.

[0093] S208: Send the prompt message to the generative base model so that the generative base model can perform task response generation processing based on the prompt message.

[0094] After determining the prompt information in step S206 above, the prompt information can be sent to the generative base model. The generative base model can perform task response generation processing based on the prompt information to obtain the required result.

[0095] Optionally, the generative base model can be deployed on the server side, and the task response generation process can be performed by sending the prompt information to the server.

[0096] Based on this, the message processing scheme of steps S202 to S208 in this embodiment can determine the target user's conversation message in the group chat window based on the task conversation message received by the AI ​​assistant, and generate prompt information based on the target user's conversation message and the reference conversation message that has a reference relationship with the target user's conversation message. This can adapt to task reply generation in multi-threaded chat sessions in group chats, and can obtain relevant conversation messages for a specific task conversation message to accurately generate prompt information. In complex group chat environments, it can more accurately capture information interaction between users, effectively support intelligent communication among multiple roles and users, reduce the adverse effects caused by the chaos of multi-threaded sessions, and improve the accuracy of the generated prompt information in instructing tasks. This allows the generative basic model to more accurately understand the task context targeted by the prompt information when processing task reply generation based on the prompt information, thereby making the result of task reply generation processing more accurate and better in line with task requirements, thus improving the user experience.

[0097] Optionally, the server can deploy different types of generative base models, each capable of performing different types of generative processing. For example, some can be used to generate lyrics, while others can be used to generate articles, and so on.

[0098] In some optional embodiments, in step S206, when generating the prompt message, the type of the generative base model to be used can be determined based on the target user session message and the reference session message, and a prompt message matching the type of the generative base model to be used can be generated. In step S208, the prompt message can be sent to the determined generative base model to be used.

[0099] For example, based on the target user's session messages and the referenced session messages, it can be determined that the type of generative base model to be used is the type of generating lyrics. Then, a prompt message that matches the type of generative base model that generates lyrics can be generated, and the generated prompt message can be sent to the generative base model that generates lyrics, instead of sending it to the generative base model that generates articles.

[0100] Therefore, based on the target user's session messages and the referenced session messages, the appropriate generative base model can be accurately determined to the type, and appropriate prompts can be generated. This allows the appropriate generative base model to perform targeted task response generation based on the appropriate prompts, thereby enabling a more accurate understanding of the task context targeted by the prompts. As a result, the task response generation is more accurate and better meets the task requirements, which in turn improves the user experience.

[0101] In some optional embodiments, the message processing method in this application embodiment further includes: generating a target group chat card message carrying the result and responding to the task session message through an artificial intelligence assistant based on the result of the generative basic model, and displaying the target group chat card message in the group chat session window.

[0102] Based on this, by generating a target group chat card message carrying the result in response to the task session message through an AI assistant, and displaying the target group chat card message in the group chat session window, it is convenient for each user in the group chat to view the result through the target group chat card message in the group chat session window, thereby realizing visual feedback on the task completion result and improving the user experience.

[0103] For example, refer to Figure 5 A general flowchart of an example message processing scheme according to an embodiment of this application is shown. Figure 5 As shown, users can interact with an AI assistant to obtain task conversation messages. They can then identify the target user's conversation message and recall reference conversation messages based on it. Next, based on the target user's conversation message and the reference conversation messages, a prompt message is generated (for example, the target user's conversation message and the reference conversation messages can be weighted according to preset weight information, and the prompt message is generated based on the weighted target user's conversation message and the reference conversation messages). This prompt message is sent to a generative base model (which can be an LLM). The generative base model performs task response generation processing based on the prompt message, obtaining the result of the task response generation processing. Afterwards, the AI ​​assistant can generate a target group chat card message carrying the result, responding to the task conversation message, and displaying the target group chat card message in the group chat conversation window. It should be understood that... Figure 5 The examples provided are merely simple illustrations for ease of understanding and are not intended to limit the scope of the embodiments described in this application.

[0104] The following is combined Figure 6A and Figure 6B The illustrated scenario provides an example of an application scenario for the message processing scheme of this application embodiment. This example can be understood by substituting it with a real-world scenario where a user uses an AI assistant to count the number of participants in a team-building activity.

[0105] like Figure 6AAs shown, the group chat window contains multiple user conversation messages, sent by users 1 through 6. Any user can invoke the AI ​​assistant within the group chat window; here, we'll use user 2 as an example. User 2 can interact with the AI ​​assistant through the input field. Taking the task conversation message sent by user 2 as "Please help me count the number of people planning to participate in the team building activity" as an example, the user clicks send to send the task conversation message to the AI ​​assistant. Then, the group chat window can identify user 2's message "Those participating in the team building activity, please let us know in the group" as the target user conversation message. Next, it identifies reference conversation messages that have a referencing relationship with the target user conversation message. Specifically, the user conversation messages from users 3, 4, and 5 are all reference conversation messages. Furthermore, it can be determined that user 6's conversation message is a similar conversation message. Finally, based on the target user conversation message, reference conversation messages, and similar conversation messages, a prompt message can be generated.

[0106] like Figure 6B As shown, after generating the prompt message, it can be sent to the generative base model (which can be an LLM). The generative base model can then perform task response generation processing to obtain the result. The result can then be processed by an AI assistant. The AI ​​assistant can call its card service to generate a target group chat card message carrying the result and responding to the task session message, and then display the target group chat card message in the group chat session window.

[0107] It should be understood that Figures 6A-6B The examples provided are merely simple illustrations for ease of understanding and are not intended to limit the scope of the embodiments described in this application.

[0108] In summary, the message processing scheme of this application embodiment can determine the target user's conversation message in the group chat window based on the task conversation message received by the AI ​​assistant, and generate prompt information based on the target user's conversation message and the reference conversation message that has a reference relationship with the target user's conversation message. This can adapt to task reply generation in multi-threaded chat sessions in group chats, and can obtain relevant conversation messages for a specific task conversation message to accurately generate prompt information. In complex group chat environments, it can more accurately capture information interaction between users, effectively support intelligent communication among multiple roles and users, reduce the adverse effects caused by the chaos of multi-threaded sessions, and improve the accuracy of the generated prompt information in instructing tasks. This allows the generative basic model to more accurately understand the task context targeted by the prompt information when processing task reply generation based on the prompt information, thereby making the result of task reply generation processing more accurate and better in line with task requirements, thus improving the user experience.

[0109] It is understood that the message processing scheme in this application embodiment can effectively realize multi-threaded session management in group chats, track and process multiple concurrent dialogues in a group chat environment, ensure efficient allocation of computing resources, and enable each individual dialogue thread to proceed continuously and correctly, thus ensuring high-efficiency dialogue management. Furthermore, the message processing scheme in this application embodiment can effectively achieve accurate semantic recognition of multiple roles, accurately understand and distinguish the intentions and needs of different participants (users) in a group chat, improve the targeting and accuracy of group chat interactions, and enhance the personalized service level of the dialogue. In addition, the message processing scheme in this application embodiment can dynamically inject referenced session messages, automatically integrate and input relevant user session information of the current group chat environment during group chat dialogue, generate prompt information and send it to the generative base model, enhance the generative base model's ability to understand context, and improve dialogue coherence and quality.

[0110] It should also be understood that the message processing scheme of this application embodiment introduces an artificial intelligence assistant into the group chat window to provide services to multiple users in the current group chat window simultaneously. The artificial intelligence assistant can not only interact with users in non-task-type ways, such as regular chat, but also perform tasks or provide services required by users, including but not limited to data statistics, task or service information display, and task execution based on user instructions. Based on this, whether a user in the group chat window instructs the artificial intelligence assistant to perform a task or provide a service, or the artificial intelligence assistant automatically triggers the execution of a task or the provision of a service based on triggering conditions, the artificial intelligence assistant will receive the corresponding task session message. Furthermore, the target user session message can be determined based on the task session message received by the artificial intelligence assistant, and then the referencing session message that has a reference relationship with the target user session message can be determined. By determining at least one user session message corresponding to the target user session message and carrying an identifier for the specified user in the context message, at least one user session message belonging to the same reference chain as the target user session message can be accurately and conveniently determined, thus facilitating the accurate and convenient determination of the referencing session message. The prompts generated based on the target user's session messages and the referenced session messages can then provide more accurate instructions for the task. This makes the results of the generative base model in generating task responses based on the prompts more accurate and better meet the task requirements, thereby improving the user experience.

[0111] It is understood that the foregoing description of the message processing method is merely an exemplary description of the embodiments of this application and is not intended to limit the embodiments of this application in any way.

[0112] According to a second aspect of the embodiments of this application, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; the memory is used to store a computer program; and the processor is used to execute the message processing method described in the first aspect by running the computer program stored in the memory.

[0113] Figure 7 A structural block diagram of an optional electronic device according to an embodiment of this application is shown. This application does not limit the specific implementation of the electronic device 1000; however, as an example, reference is made to... Figure 7 The electronic device 1000 provided in this application embodiment includes: a processor 1002, a communication interface 1004, a memory 1006, and a communication bus 1008. Wherein:

[0114] The processor 1002, communication interface 1004, and memory 1006 communicate with each other via communication bus 1008.

[0115] Communication interface 1004 is used to communicate with other electronic devices or servers.

[0116] The processor 1002 is used to execute the computer program 1010, specifically the relevant steps in any of the aforementioned message processing method embodiments.

[0117] Specifically, computer program 1010 may include program code that includes computer operation instructions.

[0118] The processor 1002 may be a CPU, a GPU (Graphics Processing Unit), an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0119] Memory 1006 is used to store computer program 1010. Memory 1006 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0120] Specifically, computer program 1010 can be used to cause processor 1002 to execute the message processing method in any of the foregoing embodiments.

[0121] The specific implementation of each step in computer program 1010 can be found in the corresponding descriptions of the steps and units in any of the aforementioned message processing method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding process descriptions in the aforementioned method embodiments, and will not be repeated here.

[0122] The electronic device 1000 in this application embodiment has been described in detail in the foregoing message processing method embodiment. Therefore, its related content and beneficial effects can be understood by referring to the above method embodiment, and will not be repeated here.

[0123] According to a fourth aspect of the embodiments of this application, the embodiments of this application also provide a computer storage medium storing a computer program thereon, which, when executed by a processor, implements the message processing method described in any of the foregoing method embodiments. The computer storage medium includes, but is not limited to, compact disc read-only memory (CD-ROM), random access memory (RAM), floppy disk, hard disk, or magneto-optical disk, etc.

[0124] According to a fifth aspect of the embodiments of this application, the embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the message processing method as described in any of the embodiments of the plurality of method embodiments described above.

[0125] The electronic device 1000 / computer storage medium / computer program product embodiments in this application have been described in detail in the foregoing message processing method embodiments. Therefore, their related content and beneficial effects can be understood by referring to the above method embodiments, and will not be repeated here.

[0126] Furthermore, it should be noted that the user-related information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to sample data used for training the model, data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0127] It should be noted that, depending on the implementation needs, the various components / steps described in the embodiments of this application can be broken down into more components / steps, or two or more components / steps or parts of the operation of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of this application.

[0128] The methods described in the embodiments of this application can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code downloaded over a network that is originally stored in a remote recording medium or a non-transitory machine-readable medium and will be stored in a local recording medium. Thus, the methods described herein can be stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an Application Specific Integrated Circuit (ASIC) or a Field Programmable Gate Array (FPGA)). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., Random Access Memory (RAM), Read-Only Memory (ROM), Flash Memory, etc.) capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for executing the methods shown herein.

[0129] Those skilled in the art will recognize that the units and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for specific applications, but such implementations should not be considered beyond the scope of the embodiments of this application.

[0130] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". It should be noted that the concepts of "first", "second", etc., mentioned in the embodiments of this application are only used to distinguish different devices, modules, or units, and are not used to limit the order of functions performed by these devices, modules, or units or their interdependencies. It should be noted that the modifications of "a" and "a plurality" mentioned in the embodiments of this application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0131] The above embodiments are only used to illustrate the embodiments of this application, and are not intended to limit the embodiments of this application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the embodiments of this application. Therefore, all equivalent technical solutions also fall within the scope of the embodiments of this application, and the patent protection scope of the embodiments of this application should be defined by the claims.

Claims

1. A message processing method, comprising: Based on the task conversation messages received by the AI ​​assistant in the group chat conversation window, determine the target user conversation messages in the group chat conversation window; Identify the referenced session messages that have a reference relationship with the target user's session messages; Based on the target user session message and the referenced session message, a prompt message is generated; The prompt information is sent to the generative base model so that the generative base model can perform task response generation processing based on the prompt information.

2. The method according to claim 1, wherein, The step of determining the referenced session message that has a reference relationship with the target user session message includes: From the context conversation messages of the target user conversation message in the group chat conversation window, determine at least one user conversation message that belongs to the same reference chain as the target user conversation message; The reference session message is determined based on at least one identified user session message.

3. The method of claim 2, wherein, Determining the referenced session message based on at least one determined user session message includes: Semantic analysis is performed on the at least one user session message, and the user session message that satisfies the first semantic similarity with the target user session message is identified as the referenced session message.

4. The method of claim 3, wherein, The step of performing semantic analysis on the at least one user session message, and determining the user session message that satisfies a first semantic similarity with the target user session message as the referenced session message, includes: Determine whether the at least one user session message contains preset characters and / or preset symbols used to express the same meaning as the target user session message; User session messages containing the preset characters and / or preset symbols are identified as the referenced session messages.

5. The method of claim 3, wherein, The step of determining, from the context session messages of the group chat window, at least one user session message belonging to the same reference chain as the target user session message includes: From the context conversation messages of the group chat conversation window, determine at least one user conversation message corresponding to the target user conversation message and carrying an identifier for specifying the user, wherein the identifier for specifying the user is an identifier used to indicate the user to whom the user conversation message needs to be delivered.

6. The method of claim 5, wherein, The step of performing semantic analysis on the at least one user session message, and determining the user session message that satisfies a first semantic similarity with the target user session message as the referenced session message, includes: Semantic analysis is performed on at least one user session message corresponding to the target user session message, which carries an identifier for specifying the user. Based on the semantic analysis results, user session messages whose semantic similarity to the target user session message satisfies the first semantic similarity are identified as the referenced session messages.

7. The method of any one of claims 1-6, wherein, The determined reference session messages include multimodal messages; The step of generating a prompt message based on the target user session message and the referenced session message includes: The multimodal messages are processed to obtain the multimodal message processing results; Based on the target user message, other reference session messages besides the multimodal message, and the multimodal message processing result, a prompt message is generated.

8. The method of claim 7, wherein, The multimodal messages include at least one of image messages, rich text messages, merged session history messages, and group chat card messages.

9. The method of claim 8, wherein, The processing of the multimodal message to obtain the multimodal message processing result includes at least one of the following: If the multimodal message includes an image message, then obtain the link corresponding to the image message and convert the link into a first multimodal message processing result in a predetermined format; If the multimodal message includes a rich text message, then the rich text message is decomposed to obtain a second multimodal message processing result containing text information and image information; If the multimodal message includes a merged session record message, then the merged session record message is decomposed to obtain a third multimodal message processing result containing multiple session records from the merged session record message; If the multimodal message includes a group chat card message, then the event tracking data is generated based on the group chat card message to obtain a fourth multimodal message processing result containing the event tracking data.

10. The method of any one of claims 1-6, wherein, The method further includes: From the group chat window where the target user's session message is located, obtain session messages whose publication time is within a preset time range, wherein the publication time of both the target user's session message and the referenced session message is within the preset time range; From the obtained session messages, determine other messages besides the referenced session message; From the determined other messages, identify at least one similar session message that satisfies the second semantic similarity with the target user session message; The step of generating a prompt message based on the target user session message and the referenced session message includes: A prompt message is generated based on the target user session message, the referenced session message, and at least one similar session message.

11. The method of any one of claims 1-6, wherein, The step of generating a prompt message based on the target user session message and the referenced session message includes: According to preset weight information, at least the target user session message and the referenced session message are weighted. The prompt message is generated based on the weighted target user session message and the referenced session message.

12. The method of any one of claims 1-6, wherein, The method further includes: Based on the result of the generative basic model generation process, the AI ​​assistant generates a target group chat card message that carries the result and responds to the task session message, and displays the target group chat card message in the group chat session window.

13. The method according to any one of claims 1-6, wherein, The step of generating prompt information based on the target user session message and the reference session message includes: determining the type of generative base model to be used based on the target user session message and the reference session message, and generating prompt information that matches the type of generative base model to be used; Sending the prompt information to the generative base model includes: sending the prompt information to the determined generative base model to be used.

14. An electronic device comprising: The processor, the communication interface, the memory, and the communication bus are provided, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus. The memory is used to store computer programs; The processor is configured to perform the method of any one of claims 1-13 by running the computer program stored in the memory.

15. A computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of claims 1-13.

16. A computer program product comprising a computer program that, when executed by a processor, implements the method as described in any one of claims 1-13.

Images