Intelligent collection method, device and computer equipment for user memoirs

Abstract

The application relates to an intelligent collection method and device of a user memoir and a computer device. The method comprises the following steps: obtaining user identity information of a user, and identifying a current intention category of the user based on the user identity information; generating current session information corresponding to the user through a dialogue flow control strategy, and adjusting the current session information of the user through a dialogue flow adjustment strategy to obtain new session feedback content of the user and a new intention category of the user; replacing session feedback content of the user with the new session feedback content and the new intention category, and returning to execute the above steps until a session end condition is met to obtain a session feedback content sequence of the user, so that new memoir information of the user is obtained. The method can improve the memoir collection efficiency of the user.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to a method, apparatus and computer device for intelligent collection of user memoirs. Background Technology

[0002] Smart memoirs use AI dialogue to generate text or video, recording the life experiences and emotional stories of the elderly, helping them preserve precious memories. Among these, personalized smart memoirs can interact with the elderly using local dialects, generating customized content to meet their emotional expression needs. However, smart memoirs often fail to proactively and structurally guide users (especially the elderly) in recalling and recounting past experiences, resulting in loosely structured content lacking a central theme. Therefore, improving the guidance and narration of memories for the elderly is a current research focus.

[0003] Existing technologies use large language models and multi-turn prompts to guide further communication with users, acting as intelligent companion chatbots to collect users' memories and generate memoirs. However, they have poor perception of users' emotions, resulting in a stiff interactive experience and an inability to intelligently adjust dialogue strategies based on users' moods, leading to low efficiency in collecting users' memoirs. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, and computer device for intelligent collection of user memoirs to address the aforementioned technical problems.

[0005] Firstly, this application provides an intelligent method for collecting user memoirs, including:

[0006] Obtain the user's identity information, and based on the user's identity information, identify the user's current intent category through a session intent analysis strategy;

[0007] Based on the user's current intent category and the user's conversation feedback content, the current conversation information corresponding to the user is generated through the dialogue flow control strategy. In response to the user's conversation feedback operation, the current conversation information of the user is adjusted through the dialogue flow adjustment strategy to obtain the user's new conversation feedback content and the user's new intent category.

[0008] The new session feedback content and the new intent category replace the user's session feedback content and session feedback content. Then, the process returns to execute the step of generating the current session information corresponding to the user based on the user's current intent category and the user's session feedback content through the dialogue flow control strategy, until the session end condition is met, and the user's session feedback content sequence is obtained.

[0009] Based on the sequence of conversation feedback content, the user's historical memoir information is updated to obtain the user's new memoir information.

[0010] Optionally, identifying the user's current intent category based on the user identity information using a session intent analysis strategy includes:

[0011] Based on the user identity information, the user's recent historical conversation content and preferred conversation style are identified through historical dialogue information. Based on the user's recent historical conversation content, the user's initial conversation information is generated through a conversation generation model.

[0012] The initial session information is transmitted to the user through the user's preferred session method, the initial feedback content of the user is obtained, and based on the user's initial feedback content, the user's current intent category is identified through an intent classification strategy.

[0013] Optionally, the step of generating the current session information corresponding to the user based on the user's current intent category and the user's session feedback content, through a dialogue flow control strategy, includes:

[0014] Identify the current topic corresponding to the conversation feedback content, and based on the user's current intent category, identify the topic suitability of the current topic;

[0015] Based on the topic suitability of the current topic, the current dialogue flow control strategy that suits the user is selected, and based on the user's conversation feedback content, the current semantic content of the user is identified through a semantic recognition model.

[0016] Based on the current semantic content, the current communication sequence of the user is generated through the dialogue flow control strategy, and the current conversation information of the user is filtered in the current communication sequence.

[0017] Optionally, adjusting the user's current session information through a dialogue flow adjustment strategy to obtain the user's new session feedback content and the user's new intent category includes:

[0018] In response to the user's conversation feedback operation, the current conversation feedback content of the user is obtained, and based on the current conversation feedback content of the user, the user's memory content, the user's current emotional feedback content, the user's topic comment content on the current topic, and the user's current semantic information are identified through a semantic sentiment analysis strategy.

[0019] Based on the topic comments, a semantic analysis model is used to identify the user's intention to change the topic. Based on the user's current emotional feedback and the intention to change the topic, a dialogue flow adjustment strategy is used to adjust the user's current communication sequence to obtain the user's new communication sequence.

[0020] Based on the current semantic information of the user's current session feedback content, in the new communication sequence, new session information adapted to the user is filtered out, and the new session information replaces the initial session information;

[0021] Return to the execution step of propagating the initial session information to the user through the user's preferred session method, obtaining the initial feedback content of the user's feedback, obtaining the user's new session feedback content, and the user's new intent category.

[0022] Optionally, the step of adjusting the user's current communication sequence based on the user's current emotional feedback and the intention to change the topic of the current topic, through a dialogue flow adjustment strategy, to obtain the user's new communication sequence includes:

[0023] Based on the user's current emotional feedback, an emotional analysis network is used to identify the type of emotional response the user has to the memory content and the user's current memory trend.

[0024] Based on the intended change of the current topic and the current trend of memory, the current topic is adjusted to obtain a new topic and the direction of communication for the new topic;

[0025] Based on the new topic and the direction of communication related to the new topic, the user's communication sequence is regenerated through the dialogue flow control strategy, and this communication sequence is used as the user's new communication sequence.

[0026] Optionally, updating the user's historical memoir information based on the sequence of conversation feedback content to obtain the user's new memoir information includes:

[0027] For each memory content, based on the memory content, identify the user's memory stage information and the user's memory event, and query the sub-historical memoir corresponding to the memory stage information in the historical memoir information;

[0028] Based on the sub-historical memoirs, the user's memory events, and the emotional information of the memory content, a new memoir is generated through a memoir writing model. Each of the new memoirs replaces each of the sub-historical memoirs to obtain the user's new memoir information.

[0029] Secondly, this application also provides a smart collection device for user memoirs, comprising:

[0030] The acquisition module is used to acquire the user's identity information and, based on the user's identity information, identify the user's current intent category through a session intent analysis strategy;

[0031] The adjustment module is used to generate the current session information corresponding to the user based on the user's current intent category and the user's session feedback content through a dialogue flow control strategy, and in response to the user's session feedback operation, adjust the user's current session information through a dialogue flow adjustment strategy to obtain the user's new session feedback content and the user's new intent category.

[0032] The iteration module is used to replace the user's conversation feedback content and conversation feedback content with the new conversation feedback content and the new intent category, and return to execute the step of generating the current conversation information corresponding to the user based on the user's current intent category and the user's conversation feedback content through the dialogue flow control strategy, until the conversation end condition is met, and obtain the user's conversation feedback content sequence;

[0033] The update module is used to update the user's historical memoir information based on the sequence of session feedback content, so as to obtain the user's new memoir information.

[0034] Optionally, the acquisition module is specifically used for:

[0035] Based on the user identity information, the user's recent historical conversation content and preferred conversation style are identified through historical dialogue information. Based on the user's recent historical conversation content, the user's initial conversation information is generated through a conversation generation model.

[0036] The initial session information is transmitted to the user through the user's preferred session method, the initial feedback content of the user is obtained, and based on the user's initial feedback content, the user's current intent category is identified through an intent classification strategy.

[0037] Optionally, the adjustment module is specifically used for:

[0038] Identify the current topic corresponding to the conversation feedback content, and based on the user's current intent category, identify the topic suitability of the current topic;

[0039] Based on the topic suitability of the current topic, the current dialogue flow control strategy that suits the user is selected, and based on the user's conversation feedback content, the current semantic content of the user is identified through a semantic recognition model.

[0040] Based on the current semantic content, the current communication sequence of the user is generated through the dialogue flow control strategy, and the current conversation information of the user is filtered in the current communication sequence.

[0041] Optionally, the adjustment module is specifically used for:

[0042] In response to the user's conversation feedback operation, the current conversation feedback content of the user is obtained, and based on the current conversation feedback content of the user, the user's memory content, the user's current emotional feedback content, the user's topic comment content on the current topic, and the user's current semantic information are identified through a semantic sentiment analysis strategy.

[0043] Based on the topic comments, a semantic analysis model is used to identify the user's intention to change the topic. Based on the user's current emotional feedback and the intention to change the topic, a dialogue flow adjustment strategy is used to adjust the user's current communication sequence to obtain the user's new communication sequence.

[0044] Based on the current semantic information of the user's current session feedback content, in the new communication sequence, new session information adapted to the user is filtered out, and the new session information replaces the initial session information;

[0045] Return to the execution step of propagating the initial session information to the user through the user's preferred session method, obtaining the initial feedback content of the user's feedback, obtaining the user's new session feedback content, and the user's new intent category.

[0046] Optionally, the adjustment module is specifically used for:

[0047] Based on the user's current emotional feedback, an emotional analysis network is used to identify the type of emotional response the user has to the memory content and the user's current memory trend.

[0048] Based on the intended change of the current topic and the current trend of memory, the current topic is adjusted to obtain a new topic and the direction of communication for the new topic;

[0049] Based on the new topic and the direction of communication related to the new topic, the user's communication sequence is regenerated through the dialogue flow control strategy, and this communication sequence is used as the user's new communication sequence.

[0050] Optionally, the update module is specifically used for:

[0051] For each memory content, based on the memory content, identify the user's memory stage information and the user's memory event, and query the sub-historical memoir corresponding to the memory stage information in the historical memoir information;

[0052] Based on the sub-historical memoirs, the user's memory events, and the emotional information of the memory content, a new memoir is generated through a memoir writing model. Each of the new memoirs replaces each of the sub-historical memoirs to obtain the user's new memoir information.

[0053] Thirdly, this application provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described in any one of the first aspects.

[0054] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0055] Fifthly, this application provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0056] The aforementioned intelligent collection method, apparatus, and computer device for user memoirs involves: acquiring user identity information; identifying the user's current intent category based on the user identity information using a conversation intent analysis strategy; generating corresponding current conversation information for the user based on the user's current intent category and conversation feedback content using a dialogue flow control strategy; adjusting the user's current conversation information in response to the user's conversation feedback operation using a dialogue flow adjustment strategy to obtain new conversation feedback content and a new intent category for the user; replacing the user's previous conversation feedback content with the new conversation feedback content and the new intent category; and returning to execute the step of generating corresponding current conversation information for the user based on the user's current intent category and conversation feedback content using a dialogue flow control strategy until a conversation end condition is met, thus obtaining a sequence of the user's conversation feedback content; and updating the user's historical memoir information based on the sequence of conversation feedback content to obtain the user's new memoir information. This solution identifies user intent categories within the current conversation based on their identity information and through conversation analysis. It then generates targeted conversation information tailored to that user, guiding the conversation and increasing user acceptance and engagement. This allows for a more complete, comprehensive, and effortless recollection of information, enhancing the user experience. Furthermore, the solution intelligently adjusts the conversation information based on user feedback, enabling real-time perception of user emotions and intentions. When a user indicates a shift in topic or touches on sensitive points, the system automatically and smoothly transitions to a new theme (advance), making the interaction more human-like and personalized, further improving the user experience. Finally, compared to the manual organization of existing technologies, this AuthorLLM (AU) achieves fully automated generation from dialogue to text through multiple steps (story_writer, The method of generating and optimizing memoirs by literary writers not only ensures the readability and quality of the memoir content, but also allows for textual coupling with historical memoirs, thereby improving the completeness and readability of the overall memoirs obtained by users and maximizing the efficiency of memoir collection for users. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 This is a flowchart illustrating a method for intelligently collecting user memoirs in one embodiment;

[0059] Figure 2 This is a flowchart illustrating an example of intelligent collection of user memoirs in one embodiment;

[0060] Figure 3 This is a structural block diagram of a smart collection device for user memoirs in one embodiment;

[0061] Figure 4 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0063] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0064] The intelligent collection method for user memoirs provided in this application embodiment can be applied to a system for intelligent collection of user memoirs. This system can be applied to a terminal, which includes: 1. MemoirManager (MM): Memoir Manager. As the system's central controller, it coordinates the work of various modules (TM, DM, CL, MC, AU) and handles user requests and responses. 2. TopicManager (TM): Topic Manager. Responsible for loading, persisting, and managing the topic state of the dialogue. Its core function is to maintain a {current_topic, next_topic} state machine and manage a topic library. 3. DialogueManager (DM): Dialogue Manager. Responsible for initializing and maintaining the context history of the dialogue. 4. MemoirClassifier (CL): Memoir Classifier. Responsible for classifying the user's input intent, particularly emotion, sensitivity, and transit intent. 5. MemoirChat (MC): Memoir Chatter. Responsible for calling the Large Language Model (LLM) API to generate guiding dialogues and welcome messages. 6. AuthorLLM (AU): Memoir author. Responsible for using the LLM API to process the complete dialogue history into the final memoir manuscript after the dialogue ends. 7. LLM_API: Large Language Model Interface (e.g., GPT, Claude, etc.). 8. FS (File System): Responsible for persistently storing user information, dialogue messages, state, and the final manuscript. This terminal can be, but is not limited to, various personal computers, laptops, mid-range computers, etc.The terminal, based on different user identity information, identifies the user's intent category in the current conversation through conversation analysis, thereby generating targeted conversation information suitable for that user. This not only guides the user's topic but also increases the user's acceptance and recognition of the topic, enabling the user to recall information more completely, comprehensively, and easily, thus improving the user experience. Furthermore, this solution intelligently adjusts the current conversation information based on user feedback, thereby sensing the user's emotions and intentions in real time. When the user indicates a shift in topic or touches on sensitive points, the system can automatically and smoothly switch the theme (advance), making the interaction more human-like and the content more considerate during intelligent interaction, further enhancing the user experience of recalling information. Finally, compared to the "manual organization" of existing technologies, this invention's AuthorLLM (AU)... It achieves fully automated generation from dialogue to text. Through a multi-step (story_writer, literature_writer) memoir generation and optimization method, it not only ensures the readability and quality of the memoir content, but also enables textual coupling with historical memoirs, thereby improving the completeness and readability of the overall memoirs obtained by users, and thus maximizing the efficiency of memoir collection for users.

[0065] In one exemplary embodiment, such as Figure 1 As shown, a method for intelligently collecting user memoirs is provided. Taking the application of this method to a terminal as an example, the method includes the following steps S101 to S104. Wherein:

[0066] Step S101: Obtain the user's identity information, and based on the user's identity information, identify the user's current intent category through a session intent analysis strategy.

[0067] In this embodiment, the terminal responds to the user's authentication operation and obtains the identity information of the user currently communicating. Then, the terminal generates different initial session information for users with different identity information, thereby proactively guiding the user to communicate, and then analyzes the current intent category corresponding to the user's initial session feedback information. The current intent category mainly represents the user's topic suitability to the initial session information. This current intent category includes, but is not limited to, no-intent category, no-interest category, and interested-category category. The topic suitability corresponding to each current intent category is as follows: no-intent category corresponds to medium topic suitability, no-interest category corresponds to low topic suitability, and interested-category category corresponds to high topic suitability. The specific identification process will be explained in detail later.

[0068] Step S102: Based on the user's current intent category and the user's conversation feedback content, generate the user's current conversation information through the dialogue flow control strategy, and adjust the user's current conversation information through the dialogue flow adjustment strategy in response to the user's conversation feedback operation, so as to obtain the user's new conversation feedback content and the user's new intent category.

[0069] In this embodiment, the terminal generates current conversation information corresponding to the user based on the user's current intent category and conversation feedback content through a dialogue flow control strategy. Responding to the user's conversation feedback operation, the terminal adjusts the user's current conversation information through a dialogue flow adjustment strategy to obtain new conversation feedback content and a new intent category. The dialogue flow control strategy generates a communication sequence for further communication with the user based on different topic suitability levels. This communication sequence includes multiple sequentially arranged conversation pieces, each guiding the user's recollection and narration. The dialogue adjustment strategy combines emotional feedback content and topic comments from the user's current feedback content to adjust the current topic. It also adjusts the actual conversation content of each conversation piece in the generated current communication sequence based on the emotional feedback content, thus better meeting the applicability of the user's emotional feedback content. The specific adjustment process will be explained in detail later.

[0070] Step S103: Take the new session feedback content, the new intent category, the session feedback content that replaces the user, and the session feedback content, and return to execute the step of generating the current session information corresponding to the user based on the user's current intent category and the user's session feedback content, through the dialogue flow control strategy, until the session end condition is met, and obtain the user's session feedback content sequence.

[0071] In this embodiment, the terminal will take the new session feedback content, the new intent category, the session feedback content that replaces the user, and the session feedback content, and return to execute the steps of generating the current session information corresponding to the user based on the user's current intent category and the user's session feedback content, through the dialogue flow control strategy, until the session end condition is met, and obtain the user's session feedback content sequence.

[0072] Step S104: Based on the sequence of conversation feedback content, update the user's historical memoir information to obtain the user's new memoir information.

[0073] In this embodiment, the terminal updates the user's historical memoir information based on the conversation feedback content sequence to obtain the user's new memoir information. This update process involves updating the sub-memoirs recorded at different time stages within the historical memoir information, thereby adding the memory content from each time interval stage in the current conversation feedback content sequence to the sub-memoirs. This update method ensures the readability and narrative completeness of the sub-memoirs. The specific update process will be described in detail later.

[0074] Based on the above scheme, the system identifies the user's intent category in the current conversation through conversation analysis, based on the user's identity information. This allows for the generation of targeted conversation information suitable for that user, not only guiding the topic but also increasing the user's acceptance and recognition of the topic. This enables users to recount information more completely, comprehensively, and easily, improving the user experience. Furthermore, the scheme intelligently adjusts the current conversation information based on user feedback, allowing for real-time perception of user emotions and intentions. When the user indicates a shift in topic or touches on sensitive points, the system automatically and smoothly switches the theme (advance), making the interaction more human-like and the content more considerate, further enhancing the user experience of recounting information. Finally, compared to the "manual organization" of existing technologies, the AuthorLLM (AU) of this invention achieves fully automated generation from dialogue to text through multiple steps (story_writer, The method of generating and optimizing memoirs by literary writers not only ensures the readability and quality of the memoir content, but also allows for textual coupling with historical memoirs, thereby improving the completeness and readability of the overall memoirs obtained by users and maximizing the efficiency of memoir collection for users.

[0075] Optionally, based on user identity information, a conversation intent analysis strategy is used to identify the user's current intent category, including: based on user identity information, through historical dialogue information, identifying the user's recent historical conversation content and the user's preferred conversation style; and based on the user's recent historical conversation content, generating the user's initial conversation information through a conversation generation model; propagating the initial conversation information to the user through the user's preferred conversation style to obtain the user's initial feedback content; and based on the user's initial feedback content, identifying the user's current intent category through an intent classification strategy.

[0076] In this embodiment, the terminal, based on user identity information and historical dialogue information, identifies the user's recent historical conversation content and preferred conversation style. Based on this content, the terminal generates the user's initial conversation information using a conversation generation model. This preferred conversation style includes conversation voice type (different pronunciations), playback speed, and tone. The conversation generation model can use the MM to call MemoirChat (MC) to generate personalized welcome messages / initial introductions. MC calls the LLM_API in a streaming manner, and the MM pushes the welcome message chunk returned by the LLM to the user (U) in real time.

[0077] Then, the terminal transmits initial session information to the user through the user's preferred session method, obtains the initial feedback content from the user, and identifies the user's current intent category based on the initial feedback content using an intent classification strategy. This intent analysis strategy involves the terminal first identifying the current topic after obtaining the initial feedback content, then injecting the current topic into the initial feedback content, and then asynchronously calling MemoirClassifier (CL) via MM to perform intent classification (classify_intent_async). Next, the terminal initiates three parallel calls to LLM_API via CL 4. to detect emotion, sensitivity, and transition (topic-change intent), respectively. The emotion database includes the emotion range, sensitivity range, and topic-change intent range corresponding to each intent category. Based on this emotion database, the terminal identifies the user's current emotion, sensitivity, and the current intent category corresponding to the topic-change intent through range adaptation.

[0078] Based on the above scheme, by conducting an initial conversation with the user and then identifying the current intent based on the content of the conversation, the system can intelligently and quickly identify the user's current intent information. This effectively combines the user's intent to guide the conversation, improving the efficiency of guiding the conversation and the user's conversation experience.

[0079] Optionally, based on the user's current intent category and the user's conversation feedback content, a dialogue flow control strategy is used to generate the user's current conversation information, including: identifying the current topic corresponding to the conversation feedback content, and identifying the topic suitability of the current topic based on the user's current intent category; filtering the current dialogue flow control strategy that suits the user based on the topic suitability of the current topic, and identifying the user's current semantic content based on the user's conversation feedback content through a semantic recognition model; generating the user's current communication sequence based on the current semantic content through the dialogue flow control strategy, and filtering the user's current conversation information in the current communication sequence.

[0080] In this embodiment, the terminal identifies the current topic corresponding to the conversation feedback content and, based on the user's current intent category, identifies the topic suitability of the current topic. As mentioned above, each current intent category corresponds to a topic suitability, and the terminal identifies the topic suitability of the current topic based on this correspondence. Then, the terminal filters the current conversation flow control strategy that suits the user based on the topic suitability of the current topic. Specifically, when the topic suitability is low, the corresponding conversation flow control strategy is that if the tags contain "transit" (the user wants to switch topics) or "sensitive" (triggering sensitive content), the MM calls the TM to execute "advance(persist=True)," advancing the topic state to "next_topic" and persisting the change. When the topic suitability is medium, the corresponding conversation flow control strategy is that the MM calls the MC to execute "chat." When the topic fit is high, the corresponding dialogue flow control strategy is as follows: while chatting, the system checks (opt) whether the current topic has already gone through multiple rounds (e.g., > 10 rounds) and has not been extended. If the condition is met, MM starts an extend_topic task in the background in parallel (par).

[0081] a. MM calls MC, and MC calls LLM to generate a list of subtopics related to the current topic (returning in JSON format).

[0082] b. MM calls TM's add_topic to insert a list of new topics after the current subtopic.

[0083] c. MM calls TM's set_state to set next_topic to the first subtopic of this newly expanded topic.

[0084] Among them, the extended_topic task mentioned above is the "background topic extension" mechanism (extend_topic) built by the inventor of this solution. When a user delves into a topic, it can intelligently prepare subsequent related topics in the background, avoiding the embarrassment of "having nothing to talk about after finishing the conversation" and making the memoir content richer.

[0085] Then, based on the user's conversation feedback, the terminal identifies the user's current semantic content using a semantic recognition model. Finally, based on this current semantic content, the terminal generates the user's current communication sequence through a dialogue flow control strategy, and filters the user's current conversation information within this sequence. This semantic recognition model can be implemented using the Large Language Model (LLM) API. The filtered current conversation information is selected from the unfiltered conversation information, following the order of the current communication sequence from beginning to end, choosing the conversation information that appears first in the sequence.

[0086] Based on the above scheme, by combining different topic adaptability, different dialogue flow control strategies are used to guide the user's conversation, thereby achieving strong guidance of the dialogue, ensuring that the conversation revolves around the main theme of the memoir, automatically advancing the process, and improving the efficiency of collecting the user's memoir.

[0087] Optionally, by adjusting the dialogue flow adjustment strategy, the user's current conversation information is adjusted to obtain the user's new conversation feedback content and the user's new intent category. This includes: responding to the user's conversation feedback operation, obtaining the user's current conversation feedback content, and based on the user's current conversation feedback content, identifying the user's memory content, the user's current emotional feedback content, the user's topic comment content on the current topic, and the user's current semantic information through a semantic sentiment analysis strategy; based on the topic comment content, identifying the user's intent to change the topic on the current topic through a semantic analysis model, and based on the user's current emotional feedback content and the intent to change the topic on the current topic, adjusting the user's current communication sequence through the dialogue flow adjustment strategy to obtain the user's new communication sequence; based on the current semantic information of the user's current conversation feedback content, filtering the new conversation information that is suitable for the user in the new communication sequence, and replacing the initial conversation information with the new conversation information; returning to the execution of the initial feedback content step of propagating the initial conversation information to the user through the user's preferred conversation method to obtain the user's feedback, thus obtaining the user's new conversation feedback content and the user's new intent category.

[0088] In this embodiment, the terminal responds to the user's conversation feedback operation, obtains the user's current conversation feedback content, and based on this content, uses a semantic sentiment analysis strategy to identify the user's memories, current emotional feedback, topic comments on the current topic, and current semantic information. This semantic sentiment analysis strategy involves using an LLM (Local Language Management) call to analyze the user's emotions, sensitivity, and topic transition intentions in real time to obtain the user's current emotional feedback content. A semantic recognition network is then used to identify the user's memories and topic comments. This semantic recognition network can be generated using a large language model based on natural language processing technology.

[0089] Then, based on the topic comments, the terminal uses a semantic analysis model to identify the user's intention to change the topic. Based on the user's current emotional feedback and this intention, it adjusts the user's current communication sequence using a dialogue flow adjustment strategy, resulting in a new communication sequence. The intention to change the topic is identified directly through the user's feedback, demonstrating their adaptability to the current topic. This method allows for efficient analysis of the user's adaptability to the current topic.

[0090] Then, based on the current semantic information of the user's current conversation feedback content, the terminal filters new conversation information that is suitable for the user in the new communication sequence and replaces the initial conversation information with the new conversation information. The method of filtering new conversation information is the same as the method of filtering conversation information from the communication sequence described above, and will not be repeated here. Finally, the terminal returns to the step of propagating the initial conversation information to the user through the user's preferred conversation method and obtaining the initial feedback content of the user, thus obtaining the user's new conversation feedback content and the user's new intent category.

[0091] Based on the above solution, by sensing the user's emotions and intentions in real time, the system can automatically and smoothly switch topics (advance) when the user shows signs of shifting the topic or touching on sensitive points, making the interaction more human-like and considerate.

[0092] Optionally, based on the user's current emotional feedback and the intention to change the current topic, a dialogue flow adjustment strategy is used to adjust the user's current communication sequence to obtain a new communication sequence. This includes: based on the user's current emotional feedback, using an emotional analysis network to identify the user's emotional type regarding the recalled content and the user's current recall trend; based on the intention to change the current topic and the current recall trend, adjusting the current topic to obtain a new topic and its communication direction; and based on the new topic and its communication direction, using a dialogue flow control strategy to regenerate the user's communication sequence and use this new communication sequence as the user's new communication sequence.

[0093] In this embodiment, the terminal, based on the user's current emotional feedback, uses an emotion analysis network to identify the user's emotional type regarding the recalled content and the user's current recall trend. This emotion analysis network can be a neural network that uses an LLM (Limited Language Management) to analyze the user's emotions, sensitivity, and topic transition intentions in real time. The emotional type of the recalled content represents the user's actual emotional response to the recalled content. The current recall trend indicates the user's tendency to further elaborate on the current recalled content.

[0094] Then, based on the user's intention to change the topic and the current recall trend, the terminal identifies the user's topic adaptability and adapts a new dialogue flow control strategy based on the topic adaptability. Finally, based on the new dialogue flow control strategy, the terminal regenerates the user's communication sequence and uses the communication sequence as the user's new communication sequence.

[0095] Based on the above solution, the system can perceive the user's emotions and intentions in real time. When the user shows signs of topic shifting or touching on sensitive points, the system can automatically and smoothly switch topics, improving the fluency of the user's topic communication and the efficiency of outputting recalled content.

[0096] Optionally, based on the sequence of conversational feedback content, the user's historical memoir information is updated to obtain the user's new memoir information, including: for each memoir content, based on the memoir content, identifying the user's memoir stage information and the user's memoir events, and querying the sub-historical memoirs corresponding to the memoir stage information in the historical memoir information; based on the sub-historical memoirs, the user's memoir events, and the memoir emotional information of the memoir content, a new memoir is generated through a memoir writing model, and each new memoir replaces each sub-historical memoir to obtain the user's new memoir information.

[0097] In this embodiment, for each piece of memory content, the terminal identifies the user's memory stage information and the user's memory event based on the memory content, and queries the sub-historical memoirs corresponding to the memory stage information in the historical memoir information. The memory stage includes, but is not limited to, the time period in which the memory content is located, and the memory theme corresponding to the meeting content (e.g., childhood anecdotes, emotional experiences, life experiences, work content, and life experiences, etc.). The memory event refers to the specific event content.

[0098] Then, based on the sub-historical memoirs, the user's recalled events, and the emotional information of the recalled content, the terminal generates new memoirs through a memoir writing model, and replaces each sub-historical memoir with a new memoir to obtain the user's new memoir information. The meeting writing model, implemented by the AuthorLLM (AU) designed by the staff of this solution, achieves fully automated generation from dialogue to text, ensuring the quality and readability of the text through multiple steps (story_writer, literary_writer). Specifically, the terminal controls the AU to call the LLM (story_writer role) to generate a "story version" memoir (story), which is stored in the FS. The AU again calls the LLM (literary_writer role) to refine the story literaryally, generating a "literary version" (literary), which is also stored in the FS. The AU converts the literary content into HTML format and stores it in the FS. The AU calls the Memory_API (such as external storage) for archiving (post_memory). The AU returns the final HTML text to the MM, which presents it to the user (U) and sends a farewell message.

[0099] Based on the above approach, this solution employs a multi-step automatic memoir generation method (AuthorLLM). By calling LLMs (story writer, literary writer) for different roles, the original dialogue is refined, story-driven, and literary polished, ultimately automatically generating a structured (e.g., HTML) memoir manuscript. This achieves fully automated generation from dialogue to manuscript, and the multi-step (story writer, literary writer) process ensures the quality and readability of the memoir.

[0100] This application also provides an example of intelligent collection of user memoirs, such as Figure 2 As shown, the specific processing procedure includes the following steps:

[0101] Step S201: Obtain the user's identity information.

[0102] Step S202: Based on user identity information, identify the user's recent historical conversation content and preferred conversation style through historical dialogue information, and generate the user's initial conversation information based on the user's recent historical conversation content through a conversation generation model.

[0103] Step S203: Through the user's preferred conversation method, the initial conversation information is transmitted to the user to obtain the initial feedback content of the user, and based on the user's initial feedback content, the user's current intent category is identified through an intent classification strategy.

[0104] Step S204: Identify the current topic corresponding to the conversation feedback content, and identify the topic suitability of the current topic based on the user's current intent category.

[0105] Step S205: Based on the topic suitability of the current topic, filter the current dialogue flow control strategy that suits the user, and based on the user's conversation feedback content, identify the user's current semantic content through a semantic recognition model.

[0106] Step S206: Based on the current semantic content, generate the user's current communication sequence through the dialogue flow control strategy, and filter the user's current conversation information in the current communication sequence.

[0107] Step S207: In response to the user's conversation feedback operation, obtain the user's current conversation feedback content, and based on the user's current conversation feedback content, identify the user's memory content, the user's current emotional feedback content, the user's topic comments on the current topic, and the user's current semantic information through a semantic sentiment analysis strategy.

[0108] Step S208: Based on the topic comment content, the semantic analysis model is used to identify the user's intention to change the topic of the current topic. Based on the user's current emotional feedback and the intention to change the topic of the current topic, the dialogue flow adjustment strategy is used to adjust the user's current communication sequence to obtain the user's new communication sequence.

[0109] Step S209: Based on the user's current emotional feedback content, the emotional type of the user's recall of the recalled content and the user's current recall trend tendency are identified through the emotion analysis network.

[0110] Step S210: Based on the intention to change the current topic and the current recall trend, adjust the current topic to obtain a new topic and the direction of communication for the new topic.

[0111] Step S211: Based on the new topic and the communication direction of the new topic, the user's communication sequence is regenerated through the dialogue flow control strategy, and the communication sequence is used as the user's new communication sequence.

[0112] Step S212: Replace the initial session information with the new session information.

[0113] Step S213: Return to the step of transmitting initial session information to the user through the user's preferred session method and obtaining the initial feedback content of the user's feedback, and obtain the user's new session feedback content and the user's new intent category.

[0114] Step S214: Take the new session feedback content, the new intent category, the session feedback content that replaces the user, and the session feedback content, and return to execute the step of generating the current session information corresponding to the user based on the user's current intent category and the user's session feedback content, through the dialogue flow control strategy, until the session end condition is met, and obtain the user's session feedback content sequence.

[0115] Step S215: For each memory content, based on the memory content, identify the user's memory stage information and the user's memory events, and query the sub-historical memoirs corresponding to the memory stage information in the historical memoir information.

[0116] Step S216: Based on the sub-historical memoirs, the user's memory events, and the emotional information of the memory content, a new memoir is generated through a memoir writing model, and each new memoir replaces each sub-historical memoir to obtain the user's new memoir information.

[0117] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0118] Based on the same inventive concept, this application also provides an intelligent collection device for user memoirs to implement the intelligent collection method for user memoirs described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more embodiments of the intelligent collection device for user memoirs provided below can be found in the limitations of the intelligent collection method for user memoirs described above, and will not be repeated here.

[0119] In one exemplary embodiment, such as Figure 3 As shown, a smart collection device for user memoirs is provided, comprising: an acquisition module 310, an adjustment module 320, an iteration module 330, and an update module 340, wherein:

[0120] The acquisition module 310 is used to acquire the user's user identity information and, based on the user identity information, identify the user's current intent category through a session intent analysis strategy;

[0121] The adjustment module 320 is used to generate the current session information corresponding to the user based on the user's current intent category and the user's session feedback content through a dialogue flow control strategy, and in response to the user's session feedback operation, adjust the user's current session information through a dialogue flow adjustment strategy to obtain the user's new session feedback content and the user's new intent category.

[0122] The iteration module 330 is used to replace the user's conversation feedback content and conversation feedback content with the new conversation feedback content and the new intent category, and return to execute the step of generating the current conversation information corresponding to the user based on the user's current intent category and the user's conversation feedback content through the dialogue flow control strategy, until the conversation end condition is met, and obtain the user's conversation feedback content sequence.

[0123] The update module 340 is used to update the user's historical memoir information based on the sequence of conversation feedback content, so as to obtain the user's new memoir information.

[0124] Optionally, the acquisition module 310 is specifically used for:

[0125] Based on the user identity information, the user's recent historical conversation content and preferred conversation style are identified through historical dialogue information. Based on the user's recent historical conversation content, the user's initial conversation information is generated through a conversation generation model.

[0126] The initial session information is transmitted to the user through the user's preferred session method, the initial feedback content of the user is obtained, and based on the user's initial feedback content, the user's current intent category is identified through an intent classification strategy.

[0127] Optionally, the adjustment module 320 is specifically used for:

[0128] Identify the current topic corresponding to the conversation feedback content, and based on the user's current intent category, identify the topic suitability of the current topic;

[0129] Based on the topic suitability of the current topic, the current dialogue flow control strategy that suits the user is selected, and based on the user's conversation feedback content, the current semantic content of the user is identified through a semantic recognition model.

[0130] Based on the current semantic content, the current communication sequence of the user is generated through the dialogue flow control strategy, and the current conversation information of the user is filtered in the current communication sequence.

[0131] Optionally, the adjustment module 320 is specifically used for:

[0132] In response to the user's conversation feedback operation, the current conversation feedback content of the user is obtained, and based on the current conversation feedback content of the user, the user's memory content, the user's current emotional feedback content, the user's topic comment content on the current topic, and the user's current semantic information are identified through a semantic sentiment analysis strategy.

[0133] Based on the topic comments, a semantic analysis model is used to identify the user's intention to change the topic. Based on the user's current emotional feedback and the intention to change the topic, a dialogue flow adjustment strategy is used to adjust the user's current communication sequence to obtain the user's new communication sequence.

[0134] Based on the current semantic information of the user's current session feedback content, in the new communication sequence, new session information adapted to the user is filtered out, and the new session information replaces the initial session information;

[0135] Return to the execution step of propagating the initial session information to the user through the user's preferred session method, obtaining the initial feedback content of the user's feedback, obtaining the user's new session feedback content, and the user's new intent category.

[0136] Optionally, the adjustment module 320 is specifically used for:

[0137] Based on the user's current emotional feedback, an emotional analysis network is used to identify the type of emotional response the user has to the memory content and the user's current memory trend.

[0138] Based on the intended change of the current topic and the current recall trend, the current topic is adjusted to obtain a new topic and the direction of communication for the new topic;

[0139] Based on the new topic and the direction of communication related to the new topic, the user's communication sequence is regenerated through the dialogue flow control strategy, and this communication sequence is used as the user's new communication sequence.

[0140] Optionally, the update module 340 is specifically used for:

[0141] For each memory content, based on the memory content, identify the user's memory stage information and the user's memory event, and query the sub-historical memoir corresponding to the memory stage information in the historical memoir information;

[0142] Based on the sub-historical memoirs, the user's memory events, and the emotional information of the memory content, a new memoir is generated through a memoir writing model. Each of the new memoirs replaces each of the sub-historical memoirs to obtain the user's new memoir information.

[0143] The various modules in the aforementioned intelligent collection device for user memoirs can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can invoke and execute the corresponding operations of each module.

[0144] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements an intelligent method for collecting user memoirs. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0145] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0146] In one exemplary embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of a beer warehouse inventory optimization method.

[0147] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of a beer warehouse inventory optimization method.

[0148] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of a beer warehouse inventory optimization method.

[0149] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0150] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0151] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0152] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for intelligently collecting a user memoir, characterized by, The method comprises: obtaining user identity information of a user, and identifying a current intention category of the user based on the user identity information through a conversation intention analysis strategy; generating current conversation information corresponding to the user based on the current intention category of the user and conversation feedback content of the user through a dialogue flow control strategy; in response to a conversation feedback operation of the user, obtaining current conversation feedback content of the user, and identifying recall content of the user, current emotional feedback content of the user, topic comment content of the user on a current topic, and current semantic information of the user based on the current conversation feedback content of the user through a semantic emotion analysis strategy; identifying a topic change intention of the user on the current topic based on the topic comment content through a semantic analysis model, and adjusting a current communication and exchange sequence of the user based on the current emotional feedback content of the user and the topic change intention of the current topic through a dialogue flow adjustment strategy to obtain a new communication and exchange sequence of the user; based on the current semantic information of the current conversation feedback content of the user, screening new conversation information suitable for the user in the new communication and exchange sequence to obtain new conversation feedback content of the user and a new intention category of the user; replacing the conversation feedback content of the user and the current intention category with the new conversation feedback content and the new intention category, and returning to execute the step of generating the current conversation information corresponding to the user based on the current intention category of the user and the conversation feedback content of the user through the dialogue flow control strategy until a conversation end condition is met to obtain a conversation feedback content sequence of the user; updating historical diary information of the user based on the conversation feedback content sequence to obtain new diary information of the user.

2. The method of claim 1, wherein, The method comprises: based on the user identity information, identifying recent historical conversation content of the user and a preferred conversation mode of the user through historical dialogue information, and generating initial conversation information of the user based on the recent historical conversation content of the user through a conversation generation model; propagating the initial conversation information to the user through the preferred conversation mode of the user to obtain initial feedback content fed back by the user, and identifying the current intention category of the user based on the initial feedback content of the user through an intention classification strategy.

3. The method of claim 2, wherein, The method comprises: identifying a current topic corresponding to the conversation feedback content, and identifying a topic adaptation degree of the current topic based on the current intention category of the user; based on the topic adaptation degree of the current topic, screening a current dialogue flow control strategy suitable for the user, and identifying current semantic content of the user based on the conversation feedback content of the user through a semantic recognition model; Based on the current semantic content, a current communication sequence of the user is generated through the dialogue flow control strategy, and current session information of the user is screened in the current communication sequence.

4. The method of claim 1, wherein, Based on the current emotional feedback content of the user and the topic change intention of the current topic, the current communication sequence of the user is adjusted through a dialogue flow adjustment strategy to obtain a new communication sequence of the user, including: Based on the current emotional feedback content of the user, the recall emotional type of the user to the recall content and the current recall trend tendency of the user are identified through an emotional analysis network. Based on the topic change intention of the current topic and the current recall trend tendency, the current topic is adjusted to obtain a new topic and a topic communication direction of the new topic. Based on the new topic and the topic communication direction of the new topic, the communication sequence of the user is regenerated through the dialogue flow control strategy, and the communication sequence is taken as the new communication sequence of the user.

5. The method of claim 4, wherein, Based on the sequence of the session feedback content, the historical memoir information of the user is updated to obtain new memoir information of the user, including: For each recall content, based on the recall content, the recall stage information of the user and the recall event of the user are identified, and a sub historical memoir corresponding to the recall stage information is queried in the historical memoir information. Based on the sub historical memoir, the recall event of the user and the recall emotional information of the recall content, a new memoir is generated through a memoir writing model, and each new memoir replaces each sub historical memoir to obtain new memoir information of the user.

6. A smart collection device for user memoirs, characterized by, The device comprises: The acquisition module is configured to acquire user identity information of a user, and identify a current intention category of the user based on the user identity information through a session intention analysis strategy. The adjustment module is configured to generate current session information corresponding to the user based on the current intention category of the user and session feedback content of the user through a dialogue flow control strategy, acquire current session feedback content of the user in response to a session feedback operation of the user, and identify recall content of the user, current emotional feedback content of the user, topic comment content of the user to a current topic and current semantic information of the user based on the current session feedback content of the user through a semantic emotional analysis strategy, identify a topic change intention of the user to the current topic based on the topic comment content through a semantic analysis model, and adjust a current communication sequence of the user to obtain a new communication sequence of the user based on the current emotional feedback content of the user and the topic change intention of the current topic through a dialogue flow adjustment strategy. an iteration module, configured to replace the new session feedback content and the new intention category with the session feedback content of the user and the current intention category, and return to perform a step of generating current session information corresponding to the user based on the current intention category of the user and the session feedback content of the user by means of a dialogue flow control strategy until a session end condition is met, to obtain a session feedback content sequence of the user; an updating module, configured to update historical memoir information of the user based on the session feedback content sequence, to obtain new memoir information of the user.

7. The apparatus of claim 6, wherein, The obtaining module is specifically configured to: identify recent historical session content of the user and a preferred session mode of the user based on the user identity information by means of historical dialogue information, and generate initial session information of the user based on the recent historical session content of the user by means of a session generation model; propagate the initial session information to the user by means of the preferred session mode of the user to obtain initial feedback content fed back by the user, and identify a current intention category of the user based on the initial feedback content of the user by means of an intention classification strategy.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor implements the steps of the method of any one of claims 1 to 5 when executing the computer program.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 5.

10. A computer program product comprising a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 5. The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 5.

Images