Information processing method and device, electronic equipment, storage medium and product

Abstract

This disclosure presents an information processing method, apparatus, electronic device, storage medium, and product. The method includes: receiving a conversation message input by a first user for a virtual character; retrieving a first prompt word template for the virtual character; determining the interaction scenario corresponding to the conversation message based on virtual interaction requirements in the conversation message; filling the first prompt word template with a guiding prompt word corresponding to the interaction scenario; obtaining user attribute information corresponding to the first user; filling the first prompt word template with the user attribute information to determine a target prompt word; determining a model response result based on the target prompt word and a demand response model; and determining a modal response result corresponding to the conversation message based on the model response result and multimodal response rules. This technical solution achieves standardization, customization, and multi-scenario adaptation for virtual character interaction, and can dynamically generate multimodal responses that fit the requirements based on the character, scenario, and user attributes.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to an information processing method, apparatus, electronic device, storage medium, and product. Background Technology

[0002] With the rapid development of artificial intelligence technology, virtual characters and chatbots are widely used in social entertainment, online education, customer service and other fields.

[0003] In related technologies, the generation and configuration of virtual characters mainly rely on preset dialogue templates and rule engines. Basic attributes and response strategies of the virtual character are configured through predefined rules and templates, and keyword matching is used to trigger corresponding responses. This generation and configuration method has certain limitations, resulting in virtual characters with poor flexibility, difficulty adapting to multiple scenarios and complex interactions, and inability to achieve dynamic iteration and refined operation. Summary of the Invention

[0004] This disclosure provides an information processing method, apparatus, electronic device, storage medium, and product to achieve standardization, customization, and multi-scenario adaptation of virtual character interaction. It can dynamically generate multimodal responses that meet specific needs based on the character, scenario, and user attributes, significantly improving the flexibility, adaptability, and user experience of virtual character interaction.

[0005] According to one aspect of this disclosure, an information processing method is provided, the method comprising: The system receives a conversation message input by a first user for a virtual character and retrieves a first prompt word template corresponding to the virtual character. The first prompt word template includes the virtual character's role configuration information, scene configuration items, and user attribute configuration items. The role configuration information is related to the business domain to which the virtual character belongs. The scene configuration items are used to indicate the interaction scene corresponding to the conversation message. Determine the virtual interaction requirements in the session message; based on the virtual interaction requirements, determine the interaction scenario corresponding to the session message; and populate the scenario configuration item in the first prompt word template with the guidance prompt words corresponding to the interaction scenario; and, Obtain user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word based on the filled first prompt word template; Based on the target prompt and the demand response model, determine the model response result corresponding to the virtual interaction demand. Based on the model response result and the multimodal response rule corresponding to the virtual character, determine the modal response result corresponding to the session message, and return the modal response result as a session response message.

[0006] According to another aspect of this disclosure, an information processing apparatus is provided, the apparatus comprising: The conversation message receiving module is used to receive conversation messages input by a first user for a virtual character, and to retrieve a first prompt word template corresponding to the virtual character; wherein, the first prompt word template includes the virtual character's role configuration information, scene configuration items, and user attribute configuration items; the role configuration information is related to the business domain to which the virtual character belongs; the scene configuration items are used to indicate the interaction scene corresponding to the conversation message; The interaction scenario determination module is used to determine the virtual interaction requirements in the session message, determine the interaction scenario corresponding to the session message based on the virtual interaction requirements, and fill the scene configuration item in the first prompt word template with the guidance prompt word corresponding to the interaction scenario; The prompt word determination module is used to obtain user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word according to the filled first prompt word template; The response information determination module is used to determine the model response result corresponding to the virtual interaction requirement based on the target prompt word and the requirement response model, determine the modal response result corresponding to the session message based on the model response result and the multimodal response rule corresponding to the virtual character, and return the modal response result as a session response message.

[0007] According to another aspect of this disclosure, an electronic device is provided, the electronic device comprising: One or more processors; Storage device for storing one or more programs. When one or more programs are executed by one or more processors, the one or more processors implement any of the information processing methods described in the embodiments of this disclosure.

[0008] According to another aspect of this disclosure, a computer-readable storage medium is provided that stores computer instructions for causing a processor to execute and implement any of the information processing methods described in the embodiments of this disclosure.

[0009] According to another aspect of the present disclosure, a computer program product is provided, which, when executed by a processor, implements any of the information processing methods described in the embodiments of the present disclosure.

[0010] The technical solution of this disclosure, by receiving a conversation message input by a first user for a virtual character and retrieving a first prompt word template corresponding to the virtual character, can quickly obtain the basic configuration of the character and the user, avoid template mixing, and improve the stability of interaction initiation. Furthermore, by determining the virtual interaction requirements in the conversation message, determining the interaction scenario corresponding to the conversation message based on the virtual interaction requirements, and filling the scenario configuration item in the first prompt word template with the guiding prompt word corresponding to the interaction scenario, it can accurately match the user's current interaction scenario and dynamically fill in the corresponding guiding prompt word, making the prompt word template more closely match the actual interaction intent, effectively improving the virtual character's adaptability and response accuracy to multi-scenario interactions. Furthermore, by obtaining user attribute information corresponding to the first user, filling the user attribute information into the user attribute configuration item in the first prompt word template, and determining the target prompt word based on the filled first prompt word template, the target prompt word is made more closely match the user characteristics, effectively improving the degree of customization of the interaction and the accuracy of model understanding. Furthermore, by determining the model response result corresponding to the virtual interaction requirement based on the target prompt words and the demand response model, and determining the modal response result corresponding to the conversation message based on the model response result and the multimodal response rule corresponding to the virtual character, the modal response result is returned as the conversation response message. This achieves accurate generation of response content based on prompt words and output of multimedia interaction results according to multimodal rules. While ensuring that the response content fits the character's persona and business scenario, it improves the richness of the interaction and the user experience. The technical solution of this disclosure embodiment solves the problem of the limitations of virtual character configuration in related technologies, which leads to poor flexibility of the generated virtual characters and difficulty in adapting to multiple scenarios and complex interactions. It achieves the standardization, customization, and multi-scenario adaptation of virtual character interaction, and can dynamically generate multimodal responses that fit the requirements based on the character, scenario, and user attributes, significantly improving the flexibility, adaptability, and user experience of virtual character interaction.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1A flowchart illustrating an information processing method provided in an embodiment of this disclosure; Figure 2 A flowchart illustrating another information processing method provided in an embodiment of this disclosure; Figure 3 A flowchart illustrating another information processing method provided in an embodiment of this disclosure; Figure 4 This is a schematic diagram of the structure of an information processing device provided in an embodiment of the present disclosure; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0014] To enable those skilled in the art to better understand the present disclosure, the technical solutions of the present disclosure will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present disclosure, and not all embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present disclosure.

[0015] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0016] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0017] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to autonomously choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0018] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt to the user can be done via a pop-up window, where the prompt can be presented as text. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0019] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0020] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0021] With the rapid development of artificial intelligence and intelligent interaction technologies, virtual characters for various business scenarios have been widely applied in social networking, interactive entertainment, online services, and other business areas. Current virtual character implementations typically rely on pre-set dialogue templates and rule engines, responding to user interactions through pre-configured fixed reply scripts and keyword matching. This approach has significant limitations: virtual characters have low coupling with business scenarios, lacking a unified and scalable management mechanism in scenarios with parallel management of multiple businesses; the prompts and interaction strategies are fixed, unable to be dynamically adjusted based on real-time user interaction needs, user attributes, and historical interaction behavior, resulting in stiff character responses, poor scenario adaptability, and insufficient differentiation; the interaction format is limited, mainly text-based, lacking the ability to collaboratively output multimodal information such as text, images, audio, and emoticons, resulting in a limited human-like interactive experience; the configuration, testing, deployment, and iteration processes for virtual characters are rigid, lacking support for version management, canary releases, and hot updates, making it difficult to achieve rapid optimization while ensuring business stability, and also lacking a proactive recall mechanism based on user behavior, making it difficult to improve user retention and activity.

[0022] To address the aforementioned issues, this disclosure proposes an information processing technology solution oriented towards business domains, supporting dynamic generation of prompts and multimodal responses. This solution resolves the problems of poor adaptability, insufficient personalization, and low iteration efficiency in existing technologies, enabling refined and unified management of virtual characters within business domains. It dynamically adapts to user interaction needs and attributes, provides a multimodal anthropomorphic interactive experience, and improves the interaction quality of virtual characters as well as user retention and activity.

[0023] Figure 1This is a flowchart illustrating an information processing method provided in an embodiment of this disclosure. This embodiment is applicable to situations involving interaction with virtual characters in any business domain. The method can be executed by an information processing device, which can be implemented in hardware and / or software and can be configured in electronic devices such as computers or servers. Figure 1 As shown, the method in this embodiment includes: S110. Receive the conversation message input by the first user for the virtual character, and retrieve the first prompt word template corresponding to the virtual character; wherein, the first prompt word template includes the virtual character's role configuration information, scene configuration items, and user attribute configuration items.

[0024] In this context, the first user can be understood as an end-user who operates the application platform and interacts with virtual characters belonging to the application platform. The first user can input conversational messages to the virtual character through the application platform and receive conversational response messages from the virtual character, thus being the application subject of the virtual character's interactive service. The first user can be an object with interaction records within the application platform to which the virtual character belongs, including at least one of registration records, login records, and interaction operation records. A virtual character can be understood as an intelligent interactive subject created based on business domain requirements and possessing fixed role settings; this intelligent interactive subject can be constructed based on artificial intelligence interaction technology. In simpler terms, the application platform can be an application or client that a user opens and uses, and the virtual character can be a virtual image or intelligent dialogue entity that chats, interacts, and provides interactive services to the user within this application. The virtual character belongs to and serves the application platform, and its role settings, dialogue style, and interaction methods are all matched to the business scenario adaptation range of the application platform. An application platform deploys one virtual character, which can be used to perform all interactive functions within the business scenario adaptation range of the application platform. For example, assuming the application platform is a social interaction application, the virtual characters within the platform can be virtual interactive entities such as chatbots, interactive characters, and voice assistants deployed within that application. Conversational messages can be understood as interactive information input by the first user to the virtual characters within the application platform. This interactive information can be used to characterize the first user's virtual interaction needs. Virtual interaction needs can be understood as the first user's intention or request to interact with the virtual characters, such as chatting, consulting, companionship, storytelling, and interactive games. Conversational messages can be input information that triggers subsequent operations. Conversational messages can be any form of information, optionally including at least one of the following: natural language text, voice information, images, and emoticons (e.g., emoji packs). The first prompt word template can be understood as a pre-configured prompt word framework for the virtual characters, used to construct the prompt content input to the model. The first prompt word template can include at least one of the following: character configuration information related to the virtual character itself, scene configuration items related to the interaction scenario, and user attribute configuration items related to user characteristics. The first prompt word template can generate prompt words that conform to the virtual character settings, the business scenario of the application platform, and the user interaction needs by combining the conversation messages entered by the first user.

[0025] The role configuration information can be understood as the configuration content used to define the basic attributes, background settings, interaction style, and behavioral characteristics of a virtual character, and is associated with the application platform to which the virtual character belongs. The role configuration information can be used to ensure that the virtual character's settings, expression methods, and interaction logic adapt to the business scenarios and positioning of the application platform. Optionally, the role configuration information includes at least one of the following: basic attribute information, background setting information, and behavioral characteristic information. Basic attribute information may include at least one of the following: character name, image setting (e.g., gentle companion, lively and cheerful), exclusive avatar, and character image materials. Background setting information includes at least one of the following: character identity (e.g., psychologist, chat assistant, game NPC, etc.) and business scenario adaptation scope (e.g., supporting casual conversation, story acting, interactive games, psychological counseling, etc.). Behavioral characteristic information may include at least one of the following: language style (e.g., colloquial / formal, aloof / friendly) and expression habits (e.g., whether to use interjections, preference for reply length). The scenario configuration item can be understood as a configuration item in the first prompt word template, used to represent the interaction scenario corresponding to the user's currently input conversation message, enabling the virtual character to respond according to the requirements of the current interaction scenario. The user attribute configuration item can be understood as the configuration part in the first prompt word template that reflects the user attribute characteristics. It is used to indicate the user attribute characteristics of the first user, so that the generated target prompt word can be combined with the user attributes to achieve more differentiated interaction.

[0026] In some cases, the provided solution may include receiving a conversation message input by a first user for a virtual character, which may include at least one of the following: receiving a text input operation by the first user for a text dialog box corresponding to the virtual character, and determining the input conversation message based on the text input operation; receiving a voice input operation by the first user for a voice input box corresponding to the virtual character, and determining the input conversation message based on the voice input operation; or receiving a control trigger operation by the first user for at least one message option control associated with the virtual character, and determining the input conversation message based on the selected message option control, wherein the message option control is associated with a preset conversation message.

[0027] In one implementation, a text dialog box corresponding to the virtual character is displayed. Upon receiving text input from a first user into the text dialog box, the text message entered in the dialog box is determined based on the text input, and the conversation message entered by the first user for the virtual character is determined based on the text message entered in the dialog box. Further, the virtual character's role identifier can be obtained, and a first prompt word template corresponding to the virtual character can be retrieved based on the role identifier.

[0028] S120. Determine the virtual interaction requirements in the session message, determine the interaction scenario corresponding to the session message based on the virtual interaction requirements, and fill the scene configuration item in the first prompt word template with the guidance prompt words corresponding to the interaction scenario.

[0029] Virtual interaction needs can be understood as the intention or demand of the first user to interact with the virtual character, such as chatting, consulting, companionship, storytelling, and interactive games. Interaction scenarios can be understood as the specific type or mode of interaction used in this dialogue, determined from multiple candidate interaction scenarios corresponding to the virtual character, based on the user's current virtual interaction needs. Interaction scenarios can be used to indicate the interaction method, dialogue style, and scope of interaction content between the virtual character and the first user. Optionally, interaction scenarios include at least one of the following: opening interaction, chatting, storytelling, and interactive games. Opening interaction can be understood as an interaction scenario, referring to the user's initial dialogue, opening interaction, or awakening interaction with the virtual character, also known as "icebreaker interaction." Chatting can be an interaction scenario, referring to daily conversations, emotional exchanges, psychological counseling, and companionship chats between the user and the virtual character. Storytelling can be an interaction scenario, referring to the virtual character engaging in story-based dialogue or plot progression according to a set story and character identity. Interactive games can be an interaction scenario, referring to entertaining interactions such as Q&A, mini-games, and fun interactions between the user and the virtual character. Guiding prompts can be understood as prompts corresponding to the currently defined interaction scenario. They instruct the model to generate responses that match the virtual character according to the settings, style, and logic of that interaction scenario, ensuring the interaction process conforms to the requirements of the current scenario. For example, assuming the interaction scenario is an opening interaction scenario, the corresponding guiding prompt could be, "Please start the conversation with the user in a friendly and relaxed manner, offering an opening greeting."; assuming the interaction scenario is a chat scenario, the corresponding guiding prompt could be, "Please engage in casual conversation with the user in a natural and friendly style, maintaining a smooth dialogue."; assuming the interaction scenario is a story-driven scenario, the corresponding guiding prompt could be, "Please conduct dialogue according to the established story background and character identity to advance the plot."; assuming the interaction scenario is an interactive game scenario, the corresponding guiding prompt could be, "Please engage in mini-game interactions with the user in a relaxed and fun manner, guiding the user to participate."

[0030] In this embodiment, determining virtual interaction needs in conversation messages includes at least one of the following: performing keyword retrieval on the conversation messages, and determining the user's current virtual interaction needs based on the correspondence between preset keywords and virtual interaction needs; parsing the semantic content of the conversation messages using a semantic understanding model to identify the user's true intent from a semantic level and obtain the corresponding virtual interaction needs; inputting the conversation messages into a trained intent classification model, and having the model output the corresponding interaction need category to determine the virtual interaction needs; and combining the current conversation messages with historical interaction content to determine the user's coherent interaction intent and thus determine the current virtual interaction needs.

[0031] Optionally, based on the virtual interaction requirements, the interaction scenario corresponding to the session message is determined, including: based on the virtual interaction requirements, the interaction scenario corresponding to the session message is determined from multiple candidate interaction scenarios related to the virtual character.

[0032] Among them, candidate interaction scenarios can be understood as a variety of interaction types that are pre-configured and available for selection for virtual characters, and are a set of alternatives for determining the current interaction scenario.

[0033] In this embodiment, the interaction scenario corresponding to the conversation message is determined from multiple candidate interaction scenarios related to the virtual character based on the virtual interaction requirements, including at least one of the following: pre-establishing a correspondence between virtual interaction requirements and candidate interaction scenarios; directly matching the corresponding interaction scenario based on the identified virtual interaction requirements; calculating the matching score with each candidate interaction scenario based on the content of the conversation message; determining the candidate interaction scenario with the highest score as the current interaction scenario; inputting the virtual interaction requirements into a trained scenario classification model; and having the model output the interaction scenario category that best matches the current conversation message to determine the final interaction scenario.

[0034] In one implementation, upon receiving a session message from a first user inputting a virtual character within a business domain, the received session message is input into a semantic understanding model. The model parses the semantic content of the session message to obtain the output virtual interaction requirements. Further, multiple candidate interaction scenarios related to the virtual character can be acquired, and a matching score between each virtual interaction scenario and each candidate interaction scenario can be determined. The candidate interaction scenario with the highest matching score is then identified as the interaction scenario corresponding to the session message. Further, a pre-configured guidance prompt word corresponding to the interaction scenario can be retrieved from a set of pre-configured guidance prompt words, and this retrieved prompt word is added to the scenario configuration item in the first prompt word template. Additionally, user attribute information corresponding to the first user can be acquired, and this acquired user attribute information is added to the user attribute configuration item in the first prompt word template. The filled first prompt word template is then identified as the target prompt word.

[0035] S130. Obtain the user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word based on the filled first prompt word template.

[0036] User attribute information can be understood as data used to characterize the first user's own features, interaction habits, or differentiated information, enabling virtual characters to interact with users more effectively. Optionally, user attribute information may include at least one of user basic attributes and user interaction attributes. Target prompt words can be understood as the final prompt content generated based on the first prompt word template and combined with the user's input conversational messages. Target prompt model information is used to guide the model to generate response results that conform to the application platform, role settings, interaction scenarios, and differentiated user needs.

[0037] In one implementation, user attribute information corresponding to the first user can be obtained, and the obtained user attribute information can be filled into the user attribute configuration items in the first prompt word template. Further, the filled first prompt word template can be determined as the target prompt word.

[0038] S140. Based on the target prompt words and the demand response model, determine the model response result corresponding to the virtual interaction demand. Based on the model response result and the multimodal response rules corresponding to the virtual role, determine the modal response result corresponding to the session message, and return the modal response result as a session response message.

[0039] The demand response model can be understood as an AI model that receives target prompts and generates preliminary response content based on the prompts. This AI model can understand and respond to the user's virtual interaction needs. The model response result can be understood as the preliminary response content output by the demand response model based on the target prompts. Multimodal response rules can be understood as pre-configured control rules for virtual characters, used to convert the model response results output by the model into multiple modal forms for display. These rules specify the presentation method, type, and combination of response content in different scenarios. Multimodal response rules can include modal response rules corresponding to multiple modal types; modal types can include at least one of text, voice, emoticons, interactive operations, images, videos, and animations. Simply put, multimodal response rules can be used to instruct the virtual character on the specific timing of responding with text, emoticons, voice, images, and animations. The modal response result can be understood as the final response content containing at least one form of expression obtained after converting the model response result according to the multimodal response rules. Optionally, modal response results include at least one of text response results, image response results, audio response results, and emoji response results. Text response results can be response content presented in text form, used to provide a text-based interactive reply to the first user. Image response results can be response content presented in the form of static or dynamic images, used to enrich the interactive display of the virtual character. Audio response results can be response content presented in audio form such as speech or sound effects, used to achieve voice interactive output from the virtual character. Emoji response results can be response content presented in the form of emojis, emotion icons, etc., used to assist in expressing the emotions and interactive states of the virtual character. A conversational response message can be understood as the final message directly returned to the first user for completing the interactive display with the virtual character. That is, a conversational response message can be a reply to the conversational message entered by the first user.

[0040] In this embodiment, the modal response result corresponding to the session message is determined based on the model response result and the multimodal response rule corresponding to the virtual character, including at least one of the following: matching the model response result with the multimodal response rule, and determining the modal response rule corresponding to the session message based on the matching result; determining the target modal type from multiple modal types according to the multimodal response rule based on user attribute information and / or the interaction scenario corresponding to the session message, and determining the modal response result corresponding to the session message based on the target modal type. The determination method will be explained in detail below.

[0041] Optionally, the multimodal response rules include modal response rules corresponding to multiple modal types; based on the model response results and the multimodal response rules corresponding to the virtual role, the modal response result corresponding to the session message is determined, including: matching the model response results with the modal response rules corresponding to multiple modal types in sequence according to the rule matching order corresponding to the multimodal response rules; if the matching result between the model response results and any modal type satisfies the preset conditions corresponding to the modal type, the modal type is determined as the target modal type, and the model response results are processed according to the target modal type to obtain the modal response result corresponding to the session message.

[0042] In this context, modality type can be understood as the form in which the response result is presented, such as text, image, audio, and emoji. Modal response rules can be understood as the processing rules set for each modality type, used to determine whether a response belongs to that modality type and how to generate the response content for that modality type. Rule matching order can be understood as the order in which matching is performed when selecting the target modality type for output. In other words, it's the order in which the model's response result matches each modal response rule. The rule matching order can be a default setting or can be adjusted by user input.

[0043] For example, the rule matching order can be: first, match the model response result with the modal response rule for audio type, then match it with the modal response rule for image type, then match it with the modal response rule for emoji type, and finally match it with the modal response rule for text type. The matching result can be understood as a quantitative value used to measure the fit, rationality, or applicability of the model response result with a certain modal type. Optionally, the matching result corresponding to the text type can be a text response probability value; the matching result corresponding to the image type can be an image content trigger probability; the matching result corresponding to the emoji type can be an emoji usage probability; and the matching result corresponding to the audio type can be a voice response ratio. Preset conditions can be understood as pre-set hit judgment criteria for each modal type. Optionally, preset conditions can include at least one of the following: the matching result is greater than a preset value, or the matching result is within a preset matching interval. The target modal type can be understood as the modal type selected from multiple modal types that meets the preset conditions and is ultimately used to generate the response.

[0044] In some cases, in order to accurately determine the target modality type that matches the model response result, modality rule matching can be performed through probability calculation in the provided solution. The specific execution process can be as follows: the model response result is matched sequentially with the modality response rules corresponding to multiple modality types, including: determining the hit probability between the model response result and the modality response rule corresponding to the modality type, and determining the determined hit probability as the matching result between the model response result and the modality type; the matching result between the model response result and any modality type meets the preset conditions corresponding to the modality type, including: the hit probability between the model response result and any modality type is located in the hit probability interval corresponding to the modality type.

[0045] The hit probability can be understood as the calculated numerical value of the likelihood that the current response will use that modality type. The hit probability can be a numerical value used to determine whether to trigger that modality type. For example, the hit probability could be 30%, 50%, or 80%. The hit probability range can be understood as a pre-defined effective probability range for each modality type. Each modality type corresponds to a different hit probability range. For example, the hit probability range for the voice modality is greater than or equal to 60%; the hit probability range for the emoji modality is 40%-80%.

[0046] In one implementation, the multimodal response rules may include modal response rules corresponding to multiple modal types. Upon obtaining the model response result, for each modal type, the model response result is sequentially matched with the modal response rules corresponding to the multimodal response rules in the order of rule matching. The hit probability between the model response result and the modal response rule corresponding to the modal type is determined, and the determined hit probability is defined as the matching result between the model response result and the modal type. Further, if the hit probability between the model response result and the modal type falls within the hit probability range corresponding to the modal type, it can be determined that the matching result between the model response result and that modal type satisfies a preset condition corresponding to that modal type. The rule matching process is then stopped, and that modal type is defined as the target modal type. Further, the model response result can be processed according to the target modal type, and the model response result can be presented in the target modal type to obtain the modal response result corresponding to the session message.

[0047] It should be noted that during the conversational interaction between the first user and the virtual character, there may be multiple rounds of conversational interaction. In this case, to further ensure that the final conversational response information is closer to the actual interaction needs of the first user and to more accurately understand the differentiated interaction needs of the first user, the information processing method may optionally include: when the number of conversational interactions between the first user and the virtual character reaches a preset threshold, generating historical interaction features based on the acquired conversational messages and their corresponding conversational response information; and when the conversational message input by the first user to the virtual character is acquired and the target prompt word corresponding to the conversational message is determined, concatenating the historical interaction features to the target prompt word to update the target prompt word.

[0048] The number of session interactions can be understood as the number of interactions between the first user and the virtual character, including the completion of a single session message and the return of that message. The preset threshold can be a pre-defined number of interactions used to determine whether to incorporate historical interaction information. Historical interaction characteristics can be understood as feature information used to characterize user interaction behavior, interaction tendencies, and historical interaction content.

[0049] In this embodiment, historical interaction features are generated based on the received session messages and their corresponding session response information. This includes: semantic extraction of the received session messages and their corresponding session response information to extract information such as the interaction characteristics and content preferences of the first user, forming historical interaction features; statistical analysis of the received session messages and their corresponding session response information, using the statistical results as historical interaction features; and vectorization of the received session messages and their corresponding session response information using an encoding model to generate feature vectors that can characterize the overall interaction content, which are then used as historical interaction features.

[0050] In one implementation, during the conversational interaction between the first user and the virtual character, the number of conversational interactions between the first user and the virtual character can be detected. Further, if the number of conversational interactions between the first user and the virtual character reaches a preset threshold, the received conversational messages and their corresponding conversational response messages can be input into an encoding model. The encoding model then vectorizes the received conversational messages and their corresponding conversational response messages to obtain the output historical interaction features. Further, upon receiving subsequent conversational messages input by the first user for the virtual character and determining the target prompt word corresponding to the conversational message, the generated historical interaction features are concatenated into the target prompt word, thereby obtaining an updated target prompt word containing the historical interaction features. Further, based on the updated target prompt word and the demand response model, a model response result corresponding to the virtual interaction demand can be determined. Based on the model response result and the multimodal response rules corresponding to the virtual character, a modal response result corresponding to the conversational message can be determined, and the modal response result can be returned as a conversational response message.

[0051] It should also be noted that during the interaction between the first user and the virtual character, there may be instances where the first user has not interacted or logged in for an extended period. In such cases, to increase the frequency and activity of user interaction with the virtual character and its associated application platform, and to maintain a continuous interactive relationship between the user and the virtual character, a recall message can be proactively pushed to the user to reactivate the dormant user. Optionally, the information processing method also includes: in response to meeting preset recall conditions, determining a recall message based on the received session messages and their corresponding session response information, and pushing the recall message to the first user through the virtual character.

[0052] The preset recall conditions can be understood as pre-set judgment conditions used to trigger the execution of user recall operations. Optionally, the preset recall conditions include at least one of the following: the duration for which no session messages input by the first user have been received reaches a first preset duration; the duration for which the first user has not logged into the application platform to which the virtual character belongs reaches a second preset duration. It can be understood that the duration for which no session messages have been received can be the time interval from the last time a session message from the first user was received to the current time. The first preset duration can refer to a pre-set time threshold used to determine whether the user has not sent session messages for an extended period. The duration for which the user has not logged into the application platform can refer to the time interval from the last time the first user logged out of the application platform to which the virtual character belongs to the current time. The second preset duration can refer to a pre-set time threshold used to determine whether the user has not logged in for an extended period. The application platform can be understood as a terminal application or service platform (such as a standalone application, web application, or embedded application module) used to provide interactive services to the virtual character, and is the carrier through which the virtual character provides interactive services to the user. That is to say, the user can interact with their virtual character through the application platform. In general, to further enable intelligent interaction between users and the application platform, virtual characters are often configured for the platform. These virtual characters then provide users with further interactive services. A recall message can be understood as a push notification generated based on the user's historical conversations, used to wake the user up and encourage them to return to the platform and continue the conversation with the virtual character, especially if the user has been inactive or logged in for an extended period. For example, if the user has previously chatted or asked questions with the virtual character and hasn't spoken for a long time, the generated recall message could be, "Long time no see! We haven't finished our previous conversation, come and continue interacting with me!" If the user has engaged in multiple rounds of task-based interaction with the virtual character and hasn't logged in for a long time, the generated recall message could be, "You haven't been here for a long time! You can continue experiencing the unfinished interactive tasks."

[0053] In one implementation, if the duration for which no session message input from the first user is received reaches a first preset duration, it is determined that the preset recall condition is met. Further, the acquired session messages and their corresponding session response information can be analyzed, and based on the analysis results, a recall message can be determined from a pre-configured pool of candidate recall messages. The determined recall message is then pushed to the first user through a virtual role.

[0054] It should be noted that, in cases involving multiple business domains, in order to achieve orderly archiving and isolated management of the interaction data of virtual characters under different business domains, the information processing may optionally include: classifying and storing the acquired session messages and their corresponding session response messages according to modal type to resource storage space associated with the business domain.

[0055] In this context, resource storage space associated with a business domain can refer to an independent storage space specifically belonging to and serving the business domain where the current virtual role resides, used to achieve data isolation and unified management. Resource storage space can refer to a storage area allocated to the corresponding business domain for storing content such as role configurations, interaction data, and multimodal resources.

[0056] The technical solution of this disclosure, by receiving a conversation message input by a first user for a virtual character and retrieving a first prompt word template corresponding to the virtual character, can quickly obtain the basic configuration of the character and the user, avoid template mixing, and improve the stability of interaction initiation. Furthermore, by determining the virtual interaction requirements in the conversation message, determining the interaction scenario corresponding to the conversation message based on the virtual interaction requirements, and filling the scenario configuration item in the first prompt word template with the guiding prompt word corresponding to the interaction scenario, it can accurately match the user's current interaction scenario and dynamically fill in the corresponding guiding prompt word, making the prompt word template more closely match the actual interaction intent, effectively improving the virtual character's adaptability and response accuracy to multi-scenario interactions. Furthermore, by obtaining user attribute information corresponding to the first user, filling the user attribute information into the user attribute configuration item in the first prompt word template, and determining the target prompt word based on the filled first prompt word template, the target prompt word is made more closely match the user characteristics, effectively improving the degree of customization of the interaction and the accuracy of model understanding. Furthermore, by determining the model response result corresponding to the virtual interaction requirement based on the target prompt words and the demand response model, and determining the modal response result corresponding to the conversation message based on the model response result and the multimodal response rule corresponding to the virtual character, the modal response result is returned as the conversation response message. This achieves accurate generation of response content based on prompt words and output of multimedia interaction results according to multimodal rules. While ensuring that the response content fits the character's persona and business scenario, it improves the richness of the interaction and the user experience. The technical solution of this disclosure embodiment solves the problem of the limitations of virtual character configuration in related technologies, which leads to poor flexibility of the generated virtual characters and difficulty in adapting to multiple scenarios and complex interactions. It achieves the standardization, customization, and multi-scenario adaptation of virtual character interaction, and can dynamically generate multimodal responses that fit the requirements based on the character, scenario, and user attributes, significantly improving the flexibility, adaptability, and user experience of virtual character interaction.

[0057] Figure 2 This is a schematic flowchart illustrating another information processing method provided in this embodiment. The technical solution of this embodiment can be combined with other embodiments; for the same or related parts, descriptions of other embodiments can be used, and will not be repeated here. Figure 2 As shown, the method in this embodiment may specifically include: S210. Determine the role configuration information based on the business association information of the business domain to which the virtual role belongs, and create a basic virtual role under the business domain based on the role configuration information.

[0058] In this context, a business domain can be understood as the business area, application scenario, or business category to which a virtual character belongs. A business domain can be any managed business domain. When multiple business domains are managed, virtual characters under different business domains can operate independently. Generally, an application platform can provide a runtime environment, user entry point, interactive capabilities, and publishing channels for a business domain; one application platform can host at least one business domain. Business-related information can be understood as information related to the business domain, describing business attributes, business scenarios, business scope, business objects, business functions, business processes, business permissions, and business objectives, providing a basis for the virtual character's basic positioning, functional scope, and interaction methods. For example, if the business domain is a customer service consultation domain, the business-related information may include at least one of the following: service scenario, consultation type, response scope, user group, etc.; if the business domain is an education and learning domain, the business-related information may include at least one of the following: subject type, learning stage, teaching style, interaction form, etc. A basic virtual character can be understood as an initial virtual character pre-created based on role configuration information, not yet bound to prompt word templates and interaction rules, serving as the basic carrier for subsequent configuration.

[0059] It should be noted that the technical solution provided in this embodiment can create one virtual role under a business domain, or multiple virtual roles under different business domains. Whether it is one or multiple roles, the technical solution provided in this embodiment can be used for creation.

[0060] In practical applications, virtual roles under different business domains are usually managed independently. Different business domains often need to maintain their own virtual roles independently, resulting in high development and operation costs and low resource utilization.

[0061] To address the above issues, this embodiment centrally manages virtual roles across multiple business domains, uniformly configures and creates basic virtual roles based on business association information between business domains, supports flexible configuration and version management of prompt word templates, interaction rules, etc., and then publishes the configured virtual roles to the corresponding application platform. This enables the reuse and unified management of virtual roles across multiple business domains, reduces development and operation costs, and improves resource utilization.

[0062] In one implementation, business association information corresponding to the business domain to which the virtual character belongs is obtained, and based on the obtained business association information, role configuration information for defining the basic attributes, identity positioning, functional scope and interaction style of the virtual character is determined. Furthermore, according to the determined role configuration information, an initial virtual character with basic business capabilities is created for the corresponding business domain, that is, the basic virtual character under the business domain.

[0063] S220: Receive the prompt word template configuration operation for the basic virtual character, and determine the first prompt word template for the basic virtual character in the first version based on the character configuration information and the prompt word template configuration operation.

[0064] The prompt template configuration operation can be understood as the configuration operations performed on the basic virtual character regarding the prompt structure, content, format, constraints, and other related conditions. The first version can be understood as a specific version of the basic virtual character during the configuration and iteration process, used to distinguish virtual characters at different stages and with different configurations. For example, the first version could be the first-generation version of the virtual character, i.e., v1. The first prompt template could be the prompt template corresponding to the basic virtual character in the first version, used to constrain the virtual character's interaction logic, response style, and output content.

[0065] Optionally, based on the role configuration information and the prompt word template configuration operation, determine the first prompt word template for the basic virtual character in the first version, including: determining the third prompt word template corresponding to the basic virtual character based on the template configuration operation, wherein the third prompt word template includes role configuration items, scene configuration items, and user attribute configuration items; and filling the role configuration information corresponding to the basic virtual character into the role configuration items to obtain the first prompt word template for the basic virtual character in the first version.

[0066] The third prompt word template can be understood as the original prompt word template framework, determined based on template configuration operations, without specific business and role information, and used as the configuration basis for the first prompt word template. Role configuration items can be understood as configuration fields in the third prompt word template used to carry information such as the basic virtual role's identity, location, and functional scope.

[0067] In one implementation, upon receiving a template configuration operation for a basic virtual character, a third prompt word template corresponding to the basic virtual character is determined based on the template configuration operation. This third prompt word template is an initial configuration template, which predefines structured fields such as role configuration items, scene configuration items, and user attribute configuration items. Furthermore, after determining the third prompt word template, the role configuration information pre-configured for the basic virtual character is filled into the role configuration items in the third prompt word template to complete the assignment of role-related information, thereby generating and obtaining the first prompt word template used by the basic virtual character in the first version.

[0068] S230, Receive interaction rule configuration operation for basic virtual character, and determine the interaction rules corresponding to basic virtual character; the interaction rules include at least multimodal response rules.

[0069] The interaction rule configuration operation can be understood as the configuration operations related to the interaction flow, response method, and output rules performed on the basic virtual character. Interaction rules can be understood as a set of rules pre-configured for the basic virtual character to constrain and regulate its interaction behavior, response method, output format, and processing logic with the user. Interaction rules at least indicate the basis for determining the modal response result. In other words, interaction rules are a set of rules that can be used to control how the virtual character interacts with the user. Interaction rules can be used to ensure that the virtual character performs dialogue, replies, and multimodal outputs according to preset requirements within its business domain. Optionally, interaction rules include multimodal response rules, dialogue flow rules, and business constraint rules, etc.

[0070] In one implementation, the system receives the interaction rule configuration operation performed by the configuration personnel for the basic virtual role. Based on the interaction constraints, response processing logic, output format requirements, and business domain-related rules set in the interaction rule configuration operation, the system determines the interaction rules corresponding to the basic virtual role, so as to provide a rule basis for generating session response messages for the subsequent virtual role according to the determined interaction rules.

[0071] S240. Bind the first prompt word template and interaction rules to the basic virtual role to obtain the virtual role under the business domain.

[0072] In this embodiment, after obtaining the configuration and interaction rules of the first prompt word template corresponding to the basic virtual role of the domain, the configured first prompt word template and interaction rules are associated and bound with the basic virtual role, establishing a correspondence between the first prompt word template, interaction rules and the basic virtual role. This enables the basic virtual role to determine the target prompt word based on the first prompt word template during operation, and to perform dialogue processing and multimodal response output according to the interaction rules. In this way, the basic virtual role, which only has basic attributes, is improved into a virtual role that can officially provide interactive services to the outside world and belongs to the corresponding business domain.

[0073] S250. Determine the first user that matches the first version, and bind the virtual character to the first user so that the first user can interact with the virtual character based on the first prompt word template.

[0074] The first user can be a target user authorized to use the current first version of the virtual role, matching the first version, and is the service recipient of the virtual role under the first version. The terminal to which the first user belongs can be an electronic device used by the first user that can install, run, and access the application platform, such as a mobile phone, tablet, or computer. The application platform to which the business domain belongs can be the software platform that carries the corresponding business domain, provides business services, and serves as the operating environment for the virtual role; it is the carrier through which the virtual role provides interactive services to the user. The binding of the virtual role to the first user can be understood as assigning the first user a virtual role of a corresponding version. The first user will exclusively use this version of the virtual role and its corresponding first prompt template for each subsequent interaction, and will not switch to other roles or other versions.

[0075] In one implementation, a first user matching the first version of the virtual character is selected and determined based on the version's applicability, user permissions, and business scenario requirements. Further, the configured virtual character is published and associated with the application platform of the corresponding business domain on the terminal used by the first user, establishing a binding relationship between the virtual character and the application platform. This allows the first user to access and use the corresponding virtual character within the application platform. During interaction, the virtual character can generate target prompts based on the first prompt word template of the first version and provide session response services to the first user according to the configured interaction rules.

[0076] S260. Receive the conversation message input by the first user for the virtual character, and retrieve the first prompt word template corresponding to the virtual character; wherein, the first prompt word template includes the virtual character's role configuration information, scene configuration items, and user attribute configuration items.

[0077] S270. Determine the virtual interaction requirements in the session message, determine the interaction scenario corresponding to the session message based on the virtual interaction requirements, and fill the scene configuration item in the first prompt word template with the guidance prompt words corresponding to the interaction scenario.

[0078] S280. Obtain the user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word based on the filled first prompt word template.

[0079] S290. Based on the target prompt words and the demand response model, determine the model response result corresponding to the virtual interaction demand. Based on the model response result and the multimodal response rules corresponding to the virtual role, determine the modal response result corresponding to the session message, and return the modal response result as a session response message.

[0080] The technical solution of this disclosure embodiment determines role configuration information based on business association information of the business domain to which the virtual role belongs, and creates a basic virtual role under the business domain based on the role configuration information; further, it receives a prompt word template configuration operation for the basic virtual role, and determines a first prompt word template for the basic virtual role in a first version based on the role configuration information and the prompt word template configuration operation; further, it receives an interaction rule configuration operation for the basic virtual role, and determines the interaction rules corresponding to the basic virtual role; the interaction rules include at least multimodal response rules; further, it binds the first prompt word template and the interaction rules to the basic virtual role to obtain a virtual role under the business domain; further, it determines a first user matching the first version, binds the virtual role to the first user, so that the first user can interact with the virtual role based on the first prompt word template. This achieves standardized creation, flexible configuration, and precise user binding of virtual roles across business domains, enabling the rapid construction of virtual roles adapted to different scenarios and stable delivery to corresponding users, improving the efficiency of virtual role management and the reliability of interaction implementation.

[0081] Figure 3 This is a schematic flowchart illustrating another information processing method provided in this embodiment. The technical solution of this embodiment can be combined with other embodiments; for the same or related parts, descriptions of other embodiments can be used, and will not be repeated here. Figure 3 As shown, the method in this embodiment may specifically include: S310. Receive the conversation message input by the first user for the virtual character, and retrieve the first prompt word template corresponding to the virtual character; wherein, the first prompt word template includes the virtual character's role configuration information, scene configuration items, and user attribute configuration items.

[0082] S320. Determine the virtual interaction requirements in the session message, determine the interaction scenario corresponding to the session message based on the virtual interaction requirements, and fill the scene configuration item in the first prompt word template with the guidance prompt words corresponding to the interaction scenario.

[0083] S330. Obtain the user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word based on the filled first prompt word template.

[0084] S340. Based on the target prompt words and the demand response model, determine the model response result corresponding to the virtual interaction demand. Based on the model response result and the multimodal response rules corresponding to the virtual role, determine the modal response result corresponding to the session message, and return the modal response result as a session response message.

[0085] S350. If the second prompt word template of the virtual character in the second version is obtained, update the virtual character according to the second prompt word template and determine the second user that matches the second version.

[0086] The second version can be understood as a new version formed after iterative updates to the virtual character's prompt word template following the first version. Its release date is later than the first version corresponding to the first prompt word template. For example, assuming the first version is v1, the second version could be v2, with v2's release date later than v1. The second prompt word template can be understood as a prompt word template formed by updating the first prompt word template. In other words, compared to the first prompt word template, the second prompt word template is a new version of the prompt word template formed by modifying, optimizing, or expanding the content, structure, or configuration items of the first prompt word template, used to achieve iterative upgrades to the virtual character's interactive capabilities. The second user refers to the target user who matches the second version of the virtual character and is authorized to use the second version, and can interact with the virtual character under the second version. The second user is the user who performs gray-scale verification of the second prompt word template under the second version. It should be noted that, to avoid conflicts in interactive effects between different versions of the prompt word template for the same user, affecting business stability or user experience, the second user does not include the first user.

[0087] In one implementation, after obtaining the second prompt word template corresponding to the virtual character in the second version through configuration operations, the prompt word template information of the virtual character is replaced and updated based on the second prompt word template, enabling the virtual character to possess the interactive capabilities and configuration content corresponding to the second version. Furthermore, based on the applicable scope of the second version, user permissions, business scenarios, and version isolation requirements, a second user matching the second version is selected and determined, and the second user does not overlap with the first user corresponding to the first version, thereby achieving user-specific isolation verification of prompt word templates for different versions.

[0088] S360: Bind the updated virtual character to the second user so that the second user can interact with the virtual character based on the second prompt word template.

[0089] In this embodiment, the updated virtual character can be associated with a second user matching the second version, establishing a correspondence between the second user and the new version's virtual character. Furthermore, when the second user initiates an interaction, the corresponding second prompt word template for the second version can be directly invoked. Based on this template, target prompt words are generated, and model responses and multimodal outputs are completed. This ensures that the second user interacts stably and independently with the virtual character throughout the entire process, according to the new version configuration, achieving the goals of version isolation and gray-scale testing.

[0090] It should be noted that the second user of the second version of the application does not include the first user of the first version. In this case, in order to upgrade the capabilities of the virtual character used by the first user, the running data of the second version can be detected during the interaction between the second user and the virtual character through the application platform based on the second prompt word template. Furthermore, if the second version is running stably, the newly added configuration information in the second version can be updated to the first prompt word template of the first version without affecting the first user's usage.

[0091] Optionally, the information processing method further includes: during the interaction between the second user and the virtual character based on the second prompt word template, obtaining the running data of the second version, evaluating the running data based on preset evaluation indicators, and obtaining evaluation results; if the evaluation results meet the preset launch standards, obtaining the newly added configuration information in the second prompt word template, and updating the newly added configuration information to the first prompt word template applied by the first user through hot update; wherein, the newly added configuration information is configuration information not included in the first prompt word template.

[0092] The operational data can be understood as the dialogue data, response effects, scene matching degree, user feedback, and other data generated during the interaction between the second user and the virtual character in the second version, used to evaluate the quality of the version. The preset evaluation indicators can be understood as pre-set evaluation criteria used to determine the usability of the second prompt word template, such as interaction accuracy, response rationality, and scene adaptability. The evaluation results can be understood as the conclusions drawn from analyzing the operational data of the second version based on the evaluation indicators, used to determine whether the launch conditions are met. The preset launch standards can be understood as the pre-set qualified conditions that the second prompt word template of the second version needs to meet for official promotion. New configuration information can be understood as configuration content included in the second prompt word template but not in the first prompt word template, which is the core improvement of this version upgrade. Hot update can be understood as a method of directly and in real-time updating configuration information while the system is running, without restarting the service or redeploying the application.

[0093] In practical applications, when a second prompt word template for a virtual character is configured in version 2, it is usually released to all users on the application platform where the virtual character is registered, so that all registered users can use the updated prompt word template. However, this may result in the new prompt word template being released to all users without sufficient verification, leading to unstable interaction effects, unexpected responses, and abnormal business scenario adaptation for some users, thus affecting the overall user experience and business operation stability.

[0094] To address the above situation, this embodiment adopts a release strategy of first performing gray-scale verification and then full update. The second prompt word template is only made available to second users who match the second version and do not overlap with the first user. After verifying the stability and reliability of the second version through runtime data, the verified new configuration information is then merged into the first prompt word template through hot update. This avoids the interactive anomalies caused by releasing the new version in full without sufficient verification, and improves the stability of the virtual character update process and the user experience.

[0095] In one implementation, during the interaction between the second user and the second prompt word template based on the virtual character, operational data generated by the second prompt word template under the second version is collected in real time. This operational data is then evaluated and its rationality verified according to pre-set evaluation indicators. Furthermore, when the evaluation results meet the preset deployment criteria, new configuration information not present in the first prompt word template is extracted from the second prompt word template. This new configuration information is then seamlessly and without restarting, synchronously updated to the first prompt word template used by the first user using a hot update method, achieving a secure upgrade of the original official version prompt word template.

[0096] It should be noted that when the newly added configuration information is updated to the first prompt word template used by the first user, the first prompt word template used by the first user will still be the first version, and will not be switched to the second version. This update only optimizes and supplements the configuration content of the first prompt word template, and does not change the version type, applicable objects, or overall architecture of the first prompt word template. In this way, while maintaining the stable operation of the first version, the prompt word template content is smoothly enhanced, avoiding fluctuations in the interactive experience for the first user due to version switching.

[0097] The technical solution of this disclosure, by configuring a second prompt word template for a virtual character in a second version, updates the virtual character according to the second prompt word template and determines a second user matching the second version; furthermore, the updated virtual character is bound to the second user so that the second user can interact with the virtual character according to the second prompt word template. This achieves version-isolated updates and user-based gray-scale releases of the virtual character prompt word template, and safely conducts new version verification without affecting the normal interaction of existing users, thereby improving the stability and security of the virtual character iteration process.

[0098] The technical solution provided in this disclosure can be applied to a virtual character processing system, which can process virtual characters across multiple business domains. The virtual character processing system may include: a virtual character configuration module, a prompt word configuration module, a version management module, a real-time debugging module, and a virtual character application module. The virtual character configuration module can be used to configure the basic attributes and manage multimedia resources of virtual characters, including building a standardized character attribute model to uniformly manage the character's basic information, background settings, and behavioral characteristics; simultaneously, it establishes an independent multimedia resource library for each virtual character, classifying and storing resources such as images, audio, and emoticons according to tags and scenarios, and managing them hierarchically, providing basic configuration support for virtual characters.

[0099] The prompt word configuration module can be used to build and manage a multi-dimensional and multi-level prompt template system, supporting scenario-based and personalized prompt word template design and parameterized configuration; it can dynamically adapt prompt templates based on user attributes and dialogue history, support A / B testing of multiple versions of prompt templates, and achieve instant configuration through a hot update mechanism, providing standardized and flexibly adjustable prompt word capabilities for the interactive generation of virtual characters.

[0100] The version management module can be used to perform full lifecycle version control on the configuration information of virtual roles and prompt templates, realize unique version management and change record tracking based on timestamps; support the distribution of different version configurations to corresponding user groups according to the canary release strategy, and provide a fast version rollback function based on configuration snapshots, so as to realize the secure release, iteration and rollback of configuration versions while ensuring uninterrupted service.

[0101] The real-time debugging module provides an isolated and secure debugging and verification environment for virtual character configuration and prompt word optimization, avoiding impact on production operations through an independent debugging environment; it supports various simulated dialogue tests such as single-turn, multi-turn, batch, and stress tests, and evaluates the debugging effect in real time based on multi-dimensional evaluation indicators; it also supports rapid configuration modification and hot update to achieve rapid iteration and optimization of virtual character interaction effects.

[0102] The virtual character application module can receive conversation messages input by the user, generate and return conversation responses that match the user's needs based on the configured virtual character information, prompt templates and interaction rules; determine the target prompt words according to the prompt template, call the demand response model to obtain the model response results, and output modal response results such as text, images, audio, and emoticons according to multimodal response rules, and finally return them to the user, realizing stable and personalized interaction with the user.

[0103] Figure 4 This is a schematic diagram of the structure of an information processing device provided in an embodiment of this disclosure. Figure 4As shown, the information processing device includes: a conversation message receiving module 410, an interaction scenario determination module 420, a prompt word determination module 430, and a response information determination module 440. The conversation message receiving module 410 receives a conversation message input by a first user for a virtual character and retrieves a first prompt word template corresponding to the virtual character. The first prompt word template includes the virtual character's role configuration information, scenario configuration items, and user attribute configuration items. The role configuration information is related to the business domain to which the virtual character belongs. The scenario configuration items indicate the interaction scenario corresponding to the conversation message. The interaction scenario determination module 420 determines the virtual interaction requirements in the conversation message, determines the interaction scenario corresponding to the conversation message based on the virtual interaction requirements, and fills in the prompt words corresponding to the interaction scenario. The prompt word determination module 430 is used to obtain user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word according to the filled first prompt word template; the response information determination module 440 is used to determine the model response result corresponding to the virtual interaction requirement according to the target prompt word and the requirement response model, determine the modal response result corresponding to the session message according to the model response result and the multimodal response rule corresponding to the virtual character, and return the modal response result as a session response message.

[0104] The technical solution of this disclosure, by receiving a conversation message input by a first user for a virtual character and retrieving a first prompt word template corresponding to the virtual character, can quickly obtain the basic configuration of the character and the user, avoid template mixing, and improve the stability of interaction initiation. Furthermore, by determining the virtual interaction requirements in the conversation message, determining the interaction scenario corresponding to the conversation message based on the virtual interaction requirements, and filling the scenario configuration item in the first prompt word template with the guiding prompt word corresponding to the interaction scenario, it can accurately match the user's current interaction scenario and dynamically fill in the corresponding guiding prompt word, making the prompt word template more closely match the actual interaction intent, effectively improving the virtual character's adaptability and response accuracy to multi-scenario interactions. Furthermore, by obtaining user attribute information corresponding to the first user, filling the user attribute information into the user attribute configuration item in the first prompt word template, and determining the target prompt word based on the filled first prompt word template, the target prompt word is made more closely match the user characteristics, effectively improving the degree of customization of the interaction and the accuracy of model understanding. Furthermore, by determining the model response result corresponding to the virtual interaction requirement based on the target prompt words and the demand response model, and determining the modal response result corresponding to the conversation message based on the model response result and the multimodal response rule corresponding to the virtual character, the modal response result is returned as the conversation response message. This achieves accurate generation of response content based on prompt words and output of multimedia interaction results according to multimodal rules. While ensuring that the response content fits the character's persona and business scenario, it improves the richness of the interaction and the user experience. The technical solution of this disclosure embodiment solves the problem of the limitations of virtual character configuration in related technologies, which leads to poor flexibility of the generated virtual characters and difficulty in adapting to multiple scenarios and complex interactions. It achieves the standardization, customization, and multi-scenario adaptation of virtual character interaction, and can dynamically generate multimodal responses that fit the requirements based on the character, scenario, and user attributes, significantly improving the flexibility, adaptability, and user experience of virtual character interaction.

[0105] In some embodiments of this disclosure, optionally, the information processing device further includes: a basic virtual character creation module, a prompt word template configuration module, an interaction rule configuration module, a virtual character creation module, and a virtual character publishing module; wherein, the basic virtual character creation module is used to determine role configuration information based on the business association information of the business domain to which the virtual character belongs, and to create a basic virtual character under the business domain based on the role configuration information; the prompt word template configuration module is used to receive a prompt word template configuration operation for the basic virtual character, and to determine a first prompt word template for the basic virtual character in a first version based on the role configuration information and the prompt word template configuration operation; An interaction rule configuration module is used to receive interaction rule configuration operations for the basic virtual character and determine the interaction rules corresponding to the basic virtual character; wherein, the interaction rules are at least used to indicate the basis for determining the modal response result; the interaction rules include at least multimodal response rules; a virtual character creation module is used to bind the first prompt word template and the interaction rules to the basic virtual character to obtain a virtual character under the business domain; a virtual character publishing module is used to determine a first user matching the first version and bind the virtual character to the first user so that the first user interacts with the virtual character according to the first prompt word template.

[0106] In some embodiments of this disclosure, optionally, the interaction scenario determination module 420 is specifically used to determine the interaction scenario corresponding to the session message from multiple candidate interaction scenarios related to the virtual character based on the virtual interaction requirements.

[0107] In some embodiments of this disclosure, the interactive scenarios may optionally include at least one of the following: initiation interaction, chat, storyline interpretation, and interactive game.

[0108] In some embodiments of this disclosure, optionally, the multimodal response rules include modal response rules corresponding to multiple modal types; the response information determination module 420 includes: a rule matching unit and a modal response result determination unit; wherein, the rule matching unit is used to match the model response result sequentially with the modal response rules corresponding to multiple modal types according to the rule matching order corresponding to the multimodal response rules; the modal response result determination unit is used to determine the modal type as the target modal type when the matching result between the model response result and any modal type satisfies the preset condition corresponding to the modal type, and process the model response result according to the target modal type to obtain the modal response result corresponding to the session message.

[0109] In some embodiments of this disclosure, optionally, the rule matching unit is specifically used to determine the hit probability between the model response result and the modal response rule corresponding to the modal type, and to determine the determined hit probability as the matching result between the model response result and the modal type; the modal response result determination unit is specifically used to determine that the hit probability between the model response result and any modal type is located in the hit probability interval corresponding to the modal type.

[0110] In some embodiments of this disclosure, the information processing device may optionally further include: a historical interaction feature generation module and a prompt word update module. The historical interaction feature generation module is used to generate historical interaction features based on the acquired conversation messages and their corresponding conversation response information when the number of conversation interactions between the first user and the virtual character reaches a preset threshold. The prompt word update module is used to, upon acquiring a conversation message input by the first user to the virtual character and determining a target prompt word corresponding to the conversation message, concatenate the historical interaction features to the target prompt word to update the target prompt word.

[0111] In some embodiments of this disclosure, the information processing device may optionally further include a recall message push module. The recall message push module is configured to, in response to satisfying preset recall conditions, determine a recall message based on the received session messages and their corresponding session response information, and push the recall message to the first user through the virtual character; wherein the preset recall conditions include at least one of the following: the duration for which no session message input by the first user has been received reaches a first preset duration; the duration for which the first user has not logged into the application platform to which the virtual character belongs reaches a second preset duration.

[0112] In some embodiments of this disclosure, the information processing device may optionally further include: a virtual character update module and a virtual character binding module. The virtual character update module is configured to update the virtual character according to the second prompt word template when a second prompt word template for the virtual character in a second version is configured, and to determine a second user matching the second version. The virtual character binding module is configured to bind the updated virtual character to the second user, so that the second user can interact with the virtual character according to the second prompt word template. The second version's release date is later than the release date of the first version corresponding to the first prompt word template.

[0113] In some embodiments of this disclosure, the information processing device may optionally further include: a runtime data acquisition module and a prompt word template update module. The runtime data acquisition module is used to acquire runtime data of the second version during the interaction between the second user and the virtual character based on the second prompt word template, and to evaluate the runtime data according to preset evaluation indicators to obtain an evaluation result. The prompt word template update module is used to acquire newly added configuration information from the second prompt word template when the evaluation result meets preset online criteria, and to update the newly added configuration information to the first prompt word template applied by the first user via hot update; wherein the newly added configuration information is configuration information not included in the first prompt word template.

[0114] In some embodiments of this disclosure, the modal response result may optionally include at least one of text response result, image response result, audio response result, and emoji response result.

[0115] The information processing apparatus provided in this disclosure can execute the information processing method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects for executing the method.

[0116] It is worth noting that the various units and modules included in the above-mentioned information processing device are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of this disclosure.

[0117] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0118] like Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0119] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0120] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as information processing methods.

[0121] In some embodiments, the information processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via read-only memory (ROM) 12 and / or communication unit 19. When the computer program is loaded into random access memory (RAM) 13 and executed by processor 11, one or more steps of the information processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the information processing method by any other suitable means (e.g., by means of firmware).

[0122] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0123] Computer programs used to implement the information processing methods of this disclosure may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0124] This disclosure provides a computer-readable storage medium storing computer instructions for causing a processor to execute an information processing method, comprising: receiving a session message input by a first user for a virtual character, and retrieving a first prompt word template corresponding to the virtual character; wherein the first prompt word template includes role configuration information, scene configuration items, and user attribute configuration items of the virtual character; the role configuration information is related to the business domain to which the virtual character belongs; the scene configuration items are used to indicate an interaction scene corresponding to the session message; determining virtual interaction requirements in the session message, and determining, based on the virtual interaction requirements, a corresponding interaction scene with the session message. The system identifies the interaction scenario corresponding to the message and fills the scene configuration item in the first prompt word template with the guidance prompt word corresponding to the interaction scenario; it also obtains the user attribute information corresponding to the first user, fills the user attribute information into the user attribute configuration item in the first prompt word template, and determines the target prompt word based on the filled first prompt word template; it determines the model response result corresponding to the virtual interaction requirement based on the target prompt word and the requirement response model, determines the modal response result corresponding to the session message based on the model response result and the multimodal response rule corresponding to the virtual character, and returns the modal response result as a session response message.

[0125] In the context of this disclosure, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0126] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0127] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0128] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0129] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of embodiments of this disclosure.

[0130] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements an information processing method according to any embodiment of this disclosure.

[0131] In implementing a computer program product, computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0132] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0133] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An information processing method characterized by comprising: include: The system receives a conversation message input by a first user for a virtual character and retrieves a first prompt word template corresponding to the virtual character. The first prompt word template includes the virtual character's role configuration information, scene configuration items, and user attribute configuration items. The role configuration information is related to the business domain to which the virtual character belongs. The scene configuration items are used to indicate the interaction scene corresponding to the conversation message. Determine the virtual interaction requirements in the session message; based on the virtual interaction requirements, determine the interaction scenario corresponding to the session message; and populate the scenario configuration item in the first prompt word template with the guidance prompt words corresponding to the interaction scenario; and, Obtain user attribute information corresponding to the first user, fill the user attribute information into the user attribute configuration item in the first prompt word template, and determine the target prompt word based on the filled first prompt word template; Based on the target prompt and the demand response model, determine the model response result corresponding to the virtual interaction demand. Based on the model response result and the multimodal response rule corresponding to the virtual character, determine the modal response result corresponding to the session message, and return the modal response result as a session response message.

2. The information processing method according to claim 1, characterized by, Also includes: Based on the business association information of the business domain to which the virtual character belongs, determine the character configuration information, and create a basic virtual character under the business domain based on the character configuration information; Receive a prompt word template configuration operation for the basic virtual character, and determine the first prompt word template for the basic virtual character in the first version based on the character configuration information and the prompt word template configuration operation; The system receives an interaction rule configuration operation for the basic virtual character and determines the interaction rules corresponding to the basic virtual character; wherein the interaction rules are at least used to indicate the basis for determining the modal response results; the interaction rules include at least multimodal response rules. The first prompt word template and the interaction rules are bound to the basic virtual role to obtain the virtual role under the business domain; A first user matching the first version is identified, and the virtual character is bound to the first user so that the first user can interact with the virtual character based on the first prompt word template.

3. The information processing method according to claim 1, characterized by, The step of determining the interaction scenario corresponding to the session message based on the virtual interaction requirements includes: Based on the virtual interaction requirements, the interaction scenario corresponding to the session message is determined from multiple candidate interaction scenarios related to the virtual character.

4. The information processing method according to claim 1, characterized by, The multimodal response rules include modal response rules corresponding to multiple modal types; determining the modal response result corresponding to the session message based on the model response result and the multimodal response rules corresponding to the virtual character includes: The model response results are matched sequentially with the modal response rules corresponding to the various modal types according to the rule matching order corresponding to the multimodal response rules. When the matching result between the model response result and any of the modal types satisfies the preset conditions corresponding to the modal type, the modal type is determined as the target modal type, and the model response result is processed according to the target modal type to obtain the modal response result corresponding to the session message.

5. The information processing method according to claim 4, characterized by, The step of matching the model response results sequentially with modal response rules corresponding to multiple modal types includes: Determine the hit probability between the model response result and the modal response rule corresponding to the modal type, and determine the determined hit probability as the matching result between the model response result and the modal type; The matching result between the model response result and any of the modal types satisfies the preset conditions corresponding to the modal type, including: The hit probability between the model response result and any of the modal types lies within the hit probability interval corresponding to the modal type.

6. The information processing method according to claim 1, characterized by, Also includes: When the number of conversation interactions between the first user and the virtual character reaches a preset threshold, historical interaction features are generated based on the acquired conversation messages and their corresponding conversation response information. Upon obtaining the conversation message input by the first user for the virtual character and determining the target prompt word corresponding to the conversation message, the historical interaction features are concatenated with the target prompt word to update the target prompt word.

7. The information processing method according to claim 1, characterized by, Also includes: In response to the fulfillment of preset recall conditions, a recall message is determined based on the received session messages and their corresponding session response information, and the recall message is pushed to the first user through the virtual character; The preset recall conditions include at least one of the following: the duration for which the first user has not received a session message has reached a first preset duration; or the duration for which the first user has not logged into the application platform to which the virtual character belongs has reached a second preset duration.

8. The information processing method according to claim 1, characterized by, Also includes: Once the second prompt word template for the virtual character in the second version is obtained, the virtual character is updated according to the second prompt word template, and a second user matching the second version is determined; The updated virtual character is bound to the second user, so that the second user can interact with the virtual character based on the second prompt word template.

9. The information processing method according to claim 8, characterized in that, Also includes: During the interaction between the second user and the virtual character based on the second prompt word template, the running data of the second version is obtained, and the running data is evaluated according to the preset evaluation indicators to obtain the evaluation result; When the evaluation result meets the preset online criteria, the newly added configuration information in the second prompt word template is obtained, and the newly added configuration information is updated to the first prompt word template applied by the first user through hot update; wherein, the newly added configuration information is configuration information not included in the first prompt word template.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the information processing method according to any one of claims 1-9.

Images