Asynchronous AI interactive accompanying mode based on relation

Abstract

The invention discloses a relation-based asynchronous AI interactive accompanying mode, which specifically comprises the following steps: S1, a user logs in an APP, selects an AI role, interacts with a note published by the AI role, and constructs a relation model according to recording and analysis of the note published by the user, behaviors, preferences and interaction history; and S2, designing an adaptive asynchronous AI interaction mode for the state, thought or daily recorded behavior issued by the user based on the relation model. The invention belongs to the technical field of asynchronous AI interaction, and aims to solve the problems that in the prior art, a real-time dialogue interaction mode easily brings psychological pressure to a user, and meanwhile, high-frequency interaction possibly causes high-concurrency calling of an AI model and increases system burden. The method has the technical effects that the asynchronous interaction mode is adopted, the user does not need to keep immediate response in AI interaction, meanwhile, the reasoning requests of the AI model are dispersed, and high-frequency synchronous calling is avoided.

Description

Technical Field

[0001] The present invention relates to the field of asynchronous AI interaction technology, and in particular to a relationship-based asynchronous AI interaction companionship method. Background Art

[0002] With the continuous development of AI technology, especially breakthroughs in natural language processing, speech recognition, image recognition and other fields, the integration of AI and social scenarios has become increasingly close. This integration has brought users a new social experience. Among them, AI companionship has become the "standard configuration" of social apps. The application of AI companionship in emotional support has become more and more common. Users can not only customize their own AI characters, but also interact deeply with AI in carefully designed plots. The AI ​​system learns users' preferences and emotions through algorithms and can provide personalized interactive experiences. This interaction makes users feel listened to and understood, especially when facing a busy and lonely life, AI becomes their emotional sustenance.

[0003] However, there are currently a large number of AI companion products on the market, and their main form of interaction is synchronous AI chat, that is, after the user selects a preset AI character, they have a real-time conversation with it. This type of interaction can easily bring psychological pressure to users, and as chat records accumulate, the AI ​​model may expose problems, thereby reducing its emotional value. In addition, in the synchronous interaction mode, high-frequency interactions may lead to high-concurrency calls to the AI ​​model, increase the system burden, and thus affect the user experience. Summary of the invention

[0004] To this end, the present invention provides a relationship-based asynchronous AI interactive companionship method to solve the above-mentioned problems in the prior art.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] According to a first aspect of the present invention, a relationship-based asynchronous AI interactive companionship method specifically includes the following steps:

[0007] S1: The user logs in to the APP, selects an AI character, and interacts with the AI ​​character by posting notes. The relationship model is built based on the records and analysis of the user's posted notes, behaviors, preferences, and interaction history;

[0008] S2: Design an adaptive asynchronous AI interaction mode based on the relationship model based on the user's own status, thoughts or daily records;

[0009] S3: The AI ​​role generates personalized response content based on the user's relationship model and the adapted asynchronous AI interaction mode and the current context, and formulates the AI ​​response time through the expectation management module;

[0010] S4: Monitor the user's interaction frequency and AI's response time, and dynamically adjust the AI's response interval to avoid frequent interactions, reduce the user's psychological burden, and make the interaction appear more natural;

[0011] S5: Set AI response expectations based on user interaction frequency, including response time and response format, so that users have reasonable expectations of AI behavior;

[0012] S6: AI generates personalized response content based on the expected response value, and pushes the response content to the user's APP client through the server. After receiving the push notification, the user opens the APP to view the AI's reply, completing a complete asynchronous interaction.

[0013] Furthermore, the specific steps of constructing the relationship model are as follows:

[0014] Step 1: Record and collect the interaction data between users and AI characters and the user's activity patterns through the APP;

[0015] Step 2: Classify the recorded and collected interaction data into different behaviors and dimensions, calculate the intimacy and trust between the user and the AI, and record the frequency of interaction between the user and the AI;

[0016] Step 3: Extract key features from different types of interaction data, train, calculate and predict key feature data based on the logistic regression mathematical model to infer the future interaction mode and response timing between users and AI;

[0017] Step 4: The logistic regression mathematical model will be updated regularly based on the user's latest behavioral data to ensure that the AI's response content and interaction methods always meet the user's current emotional needs and behavioral characteristics.

[0018] Furthermore, the specific steps of designing the asynchronous AI interaction mode are as follows:

[0019] 1): Establish an asynchronous message queue for each user. When a user posts a note, the note will be added to the queue. The content in the queue is analyzed through the relational model to determine the time and content of the reply;

[0020] 2): Set different priorities according to the intimacy calculated by the user's relationship model;

[0021] 3): Preset AI response time based on user activity patterns;

[0022] 4): AI generates highly personalized replies based on the content of the user's notes, contextual information, and relationship models. The replies include text, pictures, and voice, and will be dynamically adjusted based on the user's preferences;

[0023] 5): After replying, AI can conduct multiple rounds of interaction based on the user's further feedback. This multiple rounds of interaction can make the user feel a sense of continuous companionship;

[0024] 6): Regularly collect user feedback on AI interactions and use the feedback data to update the relationship model and optimize the AI ​​response content.

[0025] Furthermore, the interaction frequency records the average interaction time interval between the user and the AI. According to the interaction frequency between the user and the AI, the response time and content complexity of the AI ​​can be adjusted to match the interaction rhythm of the user.

[0026] Furthermore, the classification of different behaviors and dimensions includes the length of the user's interactive dialogue, the user's content preference, and the user's emotional tendency in the dialogue.

[0027] Furthermore, the interactive data includes the frequency of users posting notes, content type, user reactions after AI replies, and other user activities in the APP. User reactions include likes and replies, and other activities include browsing history and like habits.

[0028] The present invention has the following advantages:

[0029] 1. By adopting an asynchronous interaction mode similar to WeChat Moments, users do not need to respond immediately to AI interactions, reducing the psychological pressure that frequent conversations may bring. At the same time, users can post notes at a time convenient for them and receive AI replies at an appropriate time, making the interaction rhythm more natural and close to real human social habits.

[0030] 2. By building a relationship model and generating personalized responses based on the strength of the relationship between the user and AI, the interaction of AI is made more authentic and trustworthy. At the same time, the response time and form of AI are reasonably arranged to simulate more natural interpersonal interaction and enhance the user's emotional experience.

[0031] 3. Through relationship models and context analysis, AI can generate diverse and highly personalized response content, making each interaction fresh and targeted, avoiding user boredom caused by repetitive conversation content in traditional AI interactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A workflow diagram of a relationship-based asynchronous AI interactive companionship method provided for some embodiments of the present invention.

[0033] Figure 2 A flowchart of a relationship-based asynchronous AI interactive companionship method provided for some embodiments of the present invention. DETAILED DESCRIPTION

[0034] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] Example 1

[0036] like Figure 1 to Figure 2 As shown, a relationship-based asynchronous AI interactive companionship method in an embodiment of the first aspect of the present invention specifically includes the following steps:

[0037] S1: The user logs in to the APP, selects an AI character, and interacts with the AI ​​character by posting notes. A relationship model is built based on the records and analysis of the notes, behaviors, preferences, and interaction history posted by the user. The notes, behaviors, preferences, and interaction history posted by the user are parsed by the server and stored in the database for subsequent processing.

[0038] The specific steps to build a relationship model are as follows:

[0039] Step 1: Use the APP to record and collect the interaction data between users and AI characters and the user's activity patterns. The interaction data includes the frequency of users posting notes, the content type, the user's reaction after the AI ​​replies, and other user activities in the APP; user reactions include likes and replies, and other activities include browsing history and like habits;

[0040] Step 2: Classify the recorded and collected interaction data into different behaviors and dimensions, calculate the intimacy and trust between the user and AI, and record the frequency of interaction between the user and AI. The classification of different behaviors and dimensions includes the length of the user's interactive conversation, the user's content preference, and the user's emotional tendency in the conversation; intimacy indicates the closeness of the relationship between the user and the AI ​​character. The system dynamically calculates the intimacy score by analyzing the frequency of interaction between the user and AI, the length of the conversation, emotional expression, etc. Users with higher intimacy may receive faster responses and more in-depth interactive content.

[0041] Trust indicates the degree of trust a user has in an AI character. Trust is calculated based on the user's acceptance of the AI's responses, emotional changes after interaction, and long-term interaction data between the user and the AI ​​character.

[0042] The interaction frequency records the average interaction time interval between users and AI. Based on the interaction frequency between users and AI, the AI's response time and content complexity can be adjusted to match the user's interaction rhythm.

[0043] Step 3: Extract key features from different types of interaction data, train, calculate and predict key feature data based on the logistic regression mathematical model to infer the future interaction mode and response timing of users and AI, and optimize the user relationship model through continuous training to make it more accurate;

[0044] Step 4: The logistic regression mathematical model will be regularly updated based on the user's latest behavioral data to ensure that the AI's response content and interaction methods always meet the user's current emotional needs and behavioral characteristics;

[0045] S2: Design an adaptive asynchronous AI interaction mode based on the relationship model for users’ posted status, thoughts or daily records. The asynchronous interaction mode is similar to the asynchronous interaction of friends in social networks. Users can post their status, thoughts or daily records at any time. The AI ​​character will respond at the right time according to the user’s relationship model and interaction rules. Users do not need to respond immediately, which reduces the psychological pressure that frequent conversations may bring. At the same time, users can post notes at a time convenient for them and receive AI replies at the right time, making the interaction rhythm more natural and close to real human social habits.

[0046] The specific steps for designing an asynchronous AI interaction model are as follows:

[0047] 1): Establish an asynchronous message queue for each user. When a user posts a note, the note will be added to the queue. The content in the queue is analyzed through the relational model to determine the time and content of the reply;

[0048] 2): Different priorities are set according to the intimacy calculated by the user's relationship model. Users with high intimacy will receive faster responses, while users with low intimacy will need to wait longer;

[0049] 3): According to the user's activity pattern, the AI ​​response time is preset. For users who are more active during the day, the AI ​​will reply in the time period that the user is accustomed to. For users who are active at night, the AI's reply will be postponed until night time. At the same time, the AI ​​response time includes a delay strategy. The AI ​​will not reply to each note immediately, but will choose an appropriate time to reply based on the user's relationship model;

[0050] 4): AI generates highly personalized replies based on the content of the notes posted by users, contextual information and relationship models. The replies include text, pictures and voice, and will be dynamically adjusted according to user preferences.

[0051] 5): After replying, AI can conduct multiple rounds of interaction based on the user's further feedback. This multiple rounds of interaction can make the user feel a continuous sense of companionship.

[0052] 6): Regularly collect user feedback on AI interactions and use the feedback data to update the relationship model and optimize the AI ​​response content. Use natural language processing technology to analyze the emotional changes of users during the interaction, so as to further adjust the AI ​​interaction strategy to make it more in line with the emotional needs of users;

[0053] S3: The AI ​​role generates personalized response content based on the user's relationship model and the adapted asynchronous AI interaction mode and in connection with the current context, and sets the AI's response time through the expectation management module; ensuring that users have clear expectations of the AI's response time and form, so that even in the context of asynchronous interaction, users can still enjoy a coherent and consistent companionship experience, thereby enhancing user satisfaction and dependence. The content and time of the response are determined by the strength of the relationship between the user and the AI, the frequency of interaction, and the nature of the user's content. For example, for users with a close relationship, the AI ​​may reply with a warm comment in a short time to avoid users from feeling anxious due to not receiving a reply for a long time. For users with a newly established relationship, the AI ​​may delay responding for a few hours or even a few days, thereby creating a more authentic and more socially friendly interactive experience.

[0054] S4: Monitor the user's interaction frequency and AI's response time, and dynamically adjust the AI's response interval to avoid frequent interactions, reduce the user's psychological burden, and make the interaction appear more natural. For example, after a user posts multiple statuses, AI may choose to respond to some of them instead of responding to each one immediately. At the same time, the response time of the asynchronous AI interaction mode adapted to different users is very different. The reasoning requests of the AI ​​interaction mode are processed in a decentralized manner to avoid high concurrent calls in a short period of time and optimize the use of system resources. This design not only improves the system's response efficiency, but also reduces the system's operating costs, making AI's companion service more sustainable;

[0055] S5: Set the AI's response expectations based on the user's interaction frequency, including the response time and response form, so that users can have reasonable expectations of the AI's behavior; this design can reduce the user's anxiety caused by not receiving a response for a long time and enhance the interaction consistency between the user and the AI;

[0056] S6: AI generates personalized response content based on the expected response value, and pushes the response content to the user's APP client through the server. After receiving the push notification, the user opens the APP to view the AI's reply, completing a complete asynchronous interaction; through the construction and dynamic adjustment of the relationship model, a personalized response based on the strength of the relationship between the user and AI is generated, making the AI ​​interaction more real and credible. According to the user's interaction frequency and intimacy, the AI's response time and form are reasonably arranged to simulate a more natural interpersonal interaction and enhance the user's emotional experience. The asynchronous interaction mode disperses the reasoning requests of the AI ​​model, avoids high-frequency synchronous calls, and reduces the high concurrency burden of the system in a short period of time. By reasonably scheduling the reasoning time, not only the overall response efficiency of the system is improved, but also the resource consumption and operating costs are reduced, and the stability and scalability of the system are enhanced.

Claims

1. A relationship-based asynchronous AI interactive companionship method, characterized in that: The specific steps include: S1: The user logs in to the APP, selects an AI character, and interacts with the AI ​​character by posting notes. The relationship model is built based on the records and analysis of the user's posted notes, behaviors, preferences, and interaction history; S2: Design an adaptive asynchronous AI interaction mode based on the relationship model based on the user's own status, thoughts or daily records; S3: The AI ​​role generates personalized response content based on the user's relationship model and the adapted asynchronous AI interaction mode and the current context, and formulates the AI ​​response time through the expectation management module; S4: Monitor the user's interaction frequency and AI's response time, dynamically adjust the AI's response interval, avoid frequent interactions, reduce the user's psychological burden, and make the interaction appear more natural; S5: Set AI response expectations based on user interaction frequency, including response time and response format, so that users have reasonable expectations of AI behavior; S6: AI generates personalized response content based on the expected response value, and pushes the response content to the user's APP client through the server. After receiving the push notification, the user opens the APP to view the AI's reply, completing a complete asynchronous interaction.

2. According to claim 1, a relationship-based asynchronous AI interactive companionship method, characterized in that: The specific steps of constructing the relationship model are as follows: Step 1: Record and collect the interaction data between users and AI characters and the user's activity patterns through the APP; Step 2: Classify the recorded and collected interaction data into different behaviors and dimensions, calculate the intimacy and trust between the user and the AI, and record the frequency of interaction between the user and the AI; Step 3: Extract key features from different types of interaction data, train, calculate and predict key feature data based on the logistic regression mathematical model to infer the future interaction mode and response timing between users and AI; Step 4: The logistic regression mathematical model will be updated regularly based on the user's latest behavioral data to ensure that the AI's response content and interaction methods always meet the user's current emotional needs and behavioral characteristics.

3. The relationship-based asynchronous AI interactive companionship method according to claim 1, characterized in that: The specific steps of designing the asynchronous AI interaction mode are as follows: 1): Establish an asynchronous message queue for each user. When a user posts a note, the note will be added to the queue. The content in the queue is analyzed through the relational model to determine the time and content of the reply; 2): Set different priorities according to the intimacy calculated by the user's relationship model; 3): Preset AI response time based on user activity patterns; 4): AI generates highly personalized replies based on the content of the user's notes, contextual information, and relationship models. The replies include text, pictures, and voice, and will be dynamically adjusted based on the user's preferences; 5): After replying, AI can conduct multiple rounds of interaction based on the user's further feedback. This multiple rounds of interaction makes the user feel a sense of continuous companionship; 6): Regularly collect user feedback on AI interactions and use the feedback data to update the relationship model and optimize the AI ​​response content.

4. According to claim 2, the relationship-based asynchronous AI interactive companionship method is characterized in that: The interaction frequency records the average interaction time interval between the user and the AI. According to the interaction frequency between the user and the AI, the response time and content complexity of the AI ​​are adjusted to match the interaction rhythm of the user.

5. The relationship-based asynchronous AI interactive companionship method according to claim 2, characterized in that: The classification of different behaviors and dimensions includes the user's interactive dialogue length, the user's content preference, and the user's emotional tendency in the dialogue.

6. The relationship-based asynchronous AI interactive companionship method according to claim 2, characterized in that: The interactive data includes the frequency of users posting notes, content type, user reactions after AI replies, and other user activities in the APP. User reactions include likes and replies, and other activities include browsing history and like habits.

Images