Visitor processing method and device, storage medium and electronic equipment

Abstract

The embodiment of the invention discloses a visitor processing method and device, a storage medium and electronic equipment, and the method comprises the steps: determining a visitor visiting event based on the multi-mode scene monitoring information of a monitoring scene region, generating a visitor conversation plan for a visitor object based on an intelligent conversation large model in a state that a user leaves home, autonomous dialogue processing is carried out based on the visitor dialogue plan and the visitor object to obtain autonomous dialogue content, visitor visiting item information is determined based on the autonomous dialogue content and the multi-mode scene monitoring information, and a predicted user reception intention is determined by adopting an intelligent dialogue large model based on the visitor visiting item information. And adjusting the visitor dialogue plan based on the predicted user reception intention to obtain a future visit negotiation dialogue plan, and recommending a future visit suggestion scheme to the visitor object based on the future visit negotiation dialogue plan.

Description

Technical Field

[0001] This specification relates to the field of computer technology, and in particular to a visitor processing method, device, storage medium, and electronic device. Background Art

[0002] As people's work and social lives become increasingly diverse, many people need to juggle multiple roles and diverse scenarios in their daily lives. Due to flexible work schedules, increased travel, and social activities, users are often unable to stay at home all day, so it is not uncommon for them to be away when visitors arrive. Summary of the Invention

[0003] The embodiments of this specification provide a visitor processing method, device, storage medium, and electronic device. The technical solutions are as follows:

[0004] In a first aspect, an embodiment of this specification provides a visitor processing method, the method comprising:

[0005] Determine a visitor arrival event based on multimodal scene monitoring information of the monitored scene area; generate a visitor dialogue plan for the visitor object based on the intelligent dialogue model when the user is away from home; and conduct an autonomous dialogue process with the visitor object based on the visitor dialogue plan to obtain autonomous dialogue content;

[0006] Determining visitor information based on the autonomous conversation content and multimodal scene monitoring information, predicting user reception intentions based on the visitor information using an intelligent conversation model, and adjusting the visitor conversation plan based on the predicted user reception intentions to obtain a future visit negotiation conversation plan;

[0007] Recommending future visit suggestions to the visitor object based on the future visit negotiation dialogue plan.

[0008] In a feasible implementation manner, the determining visitor visit information based on the autonomous conversation content and multimodal scene monitoring information includes:

[0009] The autonomous conversation content and the multimodal scene monitoring information are subjected to semantic feature fusion to obtain comprehensive features of the visiting scene, and visitor visit item information is generated based on the comprehensive features of the visiting scene and a preset visitor visit item structure.

[0010] In a feasible implementation, the semantic feature fusion of the autonomous conversation content and the multimodal scene monitoring information is performed to obtain a comprehensive feature of the visitor scene, and the visitor event information is generated based on the comprehensive feature of the visitor scene and a preset visitor event structure, including:

[0011] Performing semantic analysis on the autonomous conversation content to obtain visitor conversation semantic information, and performing scene semantic analysis on the multimodal scene monitoring information to obtain visitor scene semantic information;

[0012] Perform semantic fusion based on the visitor conversation semantic information and the visitor scene semantic information to obtain a visitor scene comprehensive feature, and determine the visitor identity, visit purpose, and environmental context information based on the visitor scene comprehensive feature;

[0013] Key visit item fields are extracted based on the visitor identity, the visit purpose and the environmental context information, and the key visit item fields are mapped to a preset visitor visit item structure to obtain visitor visit item information.

[0014] In a feasible implementation, the method of using an intelligent dialogue model to determine the predicted user reception intention based on the visitor's visit information includes:

[0015] Obtain the user's historical visit interaction records and user visitor setting preference information;

[0016] Based on the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, an intelligent dialogue large model is used to infer the predicted user reception intention for the visitor object.

[0017] In a feasible implementation, the predicting of the user reception intention for the visitor object based on the visitor event information, the historical visit interaction records, and the user visitor setting preference information using the intelligent dialogue large model inference includes:

[0018] Extracting prediction task elements from the visitor's visit event information, the historical visit interaction records, and the user visitor setting preference information to obtain visit prediction task elements, constructing a visit interest prediction reasoning chain for the visit prediction task elements using an interest prediction task thinking template, and generating a visit interest prediction task prompt word based on the visit interest prediction reasoning chain;

[0019] The visit interest prediction task prompt words are input into the intelligent dialogue model, and the user's visit interest is gradually inferred by the intelligent dialogue model according to the visit interest prediction reasoning chain to obtain the predicted user reception intention.

[0020] In a feasible implementation, adjusting the visitor dialogue plan based on the predicted user reception intention to obtain a future visit negotiation dialogue plan includes:

[0021] Determine the target user intention type corresponding to the predicted user reception intention, obtain the recommended conversation features corresponding to the target user intention type, and obtain the user's visitor reception preference information and home schedule information;

[0022] Based on the recommended dialogue characteristics, the visitor reception preference information, the home schedule information and the visitor visit matters information, a visit negotiation plan adjustment prompt is generated, the visit negotiation plan adjustment prompt is input into the intelligent dialogue big model, the intelligent dialogue big model is used to perform feature analysis processing to obtain a plan adjustment context vector, and based on the plan adjustment context vector, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0023] In a feasible implementation, the step of using the intelligent dialogue macro model to perform feature analysis to obtain a plan adjustment context vector includes:

[0024] The intelligent dialogue model is used to perform structured extraction of input data to obtain entity nodes and entity relationship edges of incoming items;

[0025] Constructing a plan adjustment knowledge graph based on the entity nodes and entity relationship edges of the visiting matter, performing node embedding coding processing on the plan adjustment knowledge graph to obtain a node encoding vector, and performing edge relationship association conversion processing on the plan adjustment knowledge graph to obtain a node relationship weight matrix;

[0026] Based on the node encoding vector and the node relationship weight matrix, an attention mechanism is used to perform optimization to obtain a plan adjustment context vector.

[0027] In a feasible implementation manner, adjusting the visitor dialogue plan based on the plan adjustment context vector to obtain a future visit negotiation dialogue plan includes:

[0028] Determining a multidimensional dialogue adjustment semantic slot corresponding to the dialogue plan structure of the visitor dialogue plan based on the plan adjustment context vector, and generating a visit negotiation dialogue based on the multidimensional dialogue adjustment semantic slot to obtain an adjusted visit negotiation dialogue;

[0029] Based on the adjustment of the visit negotiation dialogue, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0030] In a second aspect, an embodiment of this specification provides a visitor processing device, the device comprising:

[0031] A dialogue module is used to determine a visitor arrival event based on the multimodal scene monitoring information of the monitoring scene area, generate a visitor dialogue plan for the visitor object based on the intelligent dialogue model when the user is away from home, and conduct an autonomous dialogue with the visitor object based on the visitor dialogue plan to obtain autonomous dialogue content;

[0032] A negotiation module is configured to determine visitor information based on the autonomous conversation content and multimodal scene monitoring information, predict the user's reception intention based on the visitor information using an intelligent conversation model, and adjust the visitor conversation plan based on the predicted user reception intention to obtain a future visit negotiation conversation plan;

[0033] A recommendation module is used to recommend a future visit suggestion plan to the visitor object based on the future visit negotiation dialogue plan.

[0034] In a feasible implementation manner, the determining visitor visit information based on the autonomous conversation content and multimodal scene monitoring information includes:

[0035] The autonomous conversation content and the multimodal scene monitoring information are subjected to semantic feature fusion to obtain comprehensive features of the visiting scene, and visitor visit item information is generated based on the comprehensive features of the visiting scene and a preset visitor visit item structure.

[0036] In a feasible implementation, the semantic feature fusion of the autonomous conversation content and the multimodal scene monitoring information is performed to obtain a comprehensive feature of the visitor scene, and the visitor event information is generated based on the comprehensive feature of the visitor scene and a preset visitor event structure, including:

[0037] Performing semantic analysis on the autonomous conversation content to obtain visitor conversation semantic information, and performing scene semantic analysis on the multimodal scene monitoring information to obtain visitor scene semantic information;

[0038] Perform semantic fusion based on the visitor conversation semantic information and the visitor scene semantic information to obtain a visitor scene comprehensive feature, and determine the visitor identity, visit purpose, and environmental context information based on the visitor scene comprehensive feature;

[0039] Key visit item fields are extracted based on the visitor identity, the visit purpose and the environmental context information, and the key visit item fields are mapped to a preset visitor visit item structure to obtain visitor visit item information.

[0040] In a feasible implementation, the method of using an intelligent dialogue model to determine the predicted user reception intention based on the visitor's visit information includes:

[0041] Obtain the user's historical visit interaction records and user visitor setting preference information;

[0042] Based on the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, an intelligent dialogue large model is used to infer the predicted user reception intention for the visitor object.

[0043] In a feasible implementation, the predicting of the user reception intention for the visitor object based on the visitor event information, the historical visit interaction records, and the user visitor setting preference information using the intelligent dialogue large model inference includes:

[0044] Extracting prediction task elements from the visitor's visit event information, the historical visit interaction records, and the user visitor setting preference information to obtain visit prediction task elements, constructing a visit interest prediction reasoning chain for the visit prediction task elements using an interest prediction task thinking template, and generating a visit interest prediction task prompt word based on the visit interest prediction reasoning chain;

[0045] The visit interest prediction task prompt words are input into the intelligent dialogue model, and the user's visit interest is gradually inferred by the intelligent dialogue model according to the visit interest prediction reasoning chain to obtain the predicted user reception intention.

[0046] In a feasible implementation, adjusting the visitor dialogue plan based on the predicted user reception intention to obtain a future visit negotiation dialogue plan includes:

[0047] Determine the target user intention type corresponding to the predicted user reception intention, obtain the recommended conversation features corresponding to the target user intention type, and obtain the user's visitor reception preference information and home schedule information;

[0048] Based on the recommended dialogue characteristics, the visitor reception preference information, the home schedule information and the visitor visit matters information, a visit negotiation plan adjustment prompt is generated, the visit negotiation plan adjustment prompt is input into the intelligent dialogue big model, the intelligent dialogue big model is used to perform feature analysis processing to obtain a plan adjustment context vector, and based on the plan adjustment context vector, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0049] In a feasible implementation, the step of using the intelligent dialogue macro model to perform feature analysis to obtain a plan adjustment context vector includes:

[0050] The intelligent dialogue model is used to perform structured extraction of input data to obtain entity nodes and entity relationship edges of incoming items;

[0051] Constructing a plan adjustment knowledge graph based on the entity nodes and entity relationship edges of the visiting matter, performing node embedding coding processing on the plan adjustment knowledge graph to obtain a node encoding vector, and performing edge relationship association conversion processing on the plan adjustment knowledge graph to obtain a node relationship weight matrix;

[0052] Based on the node encoding vector and the node relationship weight matrix, an attention mechanism is used to perform optimization to obtain a plan adjustment context vector.

[0053] In a feasible implementation manner, adjusting the visitor dialogue plan based on the plan adjustment context vector to obtain a future visit negotiation dialogue plan includes:

[0054] Determining a multidimensional dialogue adjustment semantic slot corresponding to the dialogue plan structure of the visitor dialogue plan based on the plan adjustment context vector, and generating a visit negotiation dialogue based on the multidimensional dialogue adjustment semantic slot to obtain an adjusted visit negotiation dialogue;

[0055] Based on the adjustment of the visit negotiation dialogue, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0056] In a third aspect, an embodiment of this specification provides a computer storage medium, wherein the computer storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the above-mentioned method steps.

[0057] In a fourth aspect, an embodiment of this specification provides an electronic device, which may include: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the above-mentioned method steps.

[0058] The beneficial effects of the technical solutions provided by some embodiments of this specification include at least:

[0059] In one or more embodiments of the present specification, when a user is away from home, an electronic device generates a visitor conversation plan for a visitor based on an intelligent conversation model, conducts an autonomous conversation with the visitor to obtain autonomous conversation content, determines visitor arrival information based on the autonomous conversation content and multimodal scene monitoring information, and then uses the intelligent conversation model to predict the user's reception intention. Based on the predicted user reception intention, the visitor conversation plan is adjusted to obtain a future visit negotiation conversation plan, and then recommends a future visit suggestion plan to the visitor. This enables the electronic device to automatically identify visitor arrival events when the user is away from home, and utilizes multimodal data and the intelligent conversation model to generate and adjust the conversation plan, thereby accurately extracting visitor needs and predicting user reception intentions, thereby providing the visitor with clear and flexible future visit suggestion plans. This not only breaks through the limitations of traditional smart home systems such as access control and intercom systems that can only provide simple one-to-one conversations, effectively reducing the security and privacy risks caused by poor information flow, but also enables intelligent and personalized responses to diverse visitor needs such as express delivery, visits with relatives and friends, and business negotiations, greatly improving the communication efficiency and user experience of visitors when the user is away. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0061] Figure 1 This is a flow chart of a visitor processing method provided in an embodiment of this specification;

[0062] Figure 2 This is a flow chart of generating visitor information provided by an embodiment of this specification;

[0063] Figure 3 This is a flowchart of a method for predicting a user's reception intention provided by an embodiment of this specification;

[0064] Figure 4 This is a flowchart of intentional reasoning provided by an embodiment of this specification;

[0065] Figure 5 This is a flowchart of a plan adjustment provided in an embodiment of this specification;

[0066] Figure 6 This is a flow chart of a feature analysis process provided by an embodiment of this specification;

[0067] Figure 7 This is a flowchart of a visit negotiation process provided by an embodiment of this specification;

[0068] Figure 8 This is a schematic structural diagram of a visitor processing device provided in an embodiment of this specification;

[0069] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification;

[0070] Figure 10 This is a schematic diagram of the structure of the operating system and user space provided in the embodiments of this specification;

[0071] Figure 11 yes Figure 10 The architecture diagram of the Android operating system;

[0072] Figure 12 yes Figure 10 Architecture diagram of the IOS operating system. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of this specification to clearly and completely describe the technical solutions in the embodiments of this specification. Obviously, the embodiments described are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this specification.

[0074] In the description of this specification, it should be understood that the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance. In the description of this specification, it should be noted that, unless otherwise expressly specified and limited, "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices. For those of ordinary skill in the art, the specific meanings of the above terms in this specification can be understood according to the specific circumstances. In addition, in the description of this specification, unless otherwise specified, "multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0075] In related technologies, users often can't stay home all day, so it's not uncommon for them to be away when visitors arrive. The purpose of these visitors is also diverse—perhaps a courier making a last-minute package delivery, requiring a remote signature or confirmation; perhaps a friend or relative seeking a surprise visit, or an important business partner visiting for an urgent matter. In these scenarios, relying solely on traditional access control or intercom systems is insufficient. They typically only offer simple one-on-one calling or door lock functionality, lacking the ability to deeply interact with the visitor's intent or effectively integrate with the user's schedule or privacy preferences.

[0076] When the user is away from home, visitors or users may need to know when they can meet, whether they can reschedule, or leave a message. If the system fails to provide a reasonable response and negotiation mechanism, it will not only affect the visitor experience but also pose security and privacy risks. Therefore, how to achieve flexible and intelligent visitor handling when the user is away from home has become a key technical issue that has attracted widespread attention. This specification is explained in detail below with reference to specific embodiments.

[0077] In one embodiment, Figure 1As shown, a visitor processing method is proposed. This method can be implemented using a computer program and can be run on a visitor processing device based on the von Neumann architecture. The computer program can be integrated into an application or run as a standalone tool application. The visitor processing device can be an electronic device, including but not limited to: a smart doorbell, a smart camera, a smart door lock, a personal computer, a tablet computer, a handheld device, an in-vehicle device, a wearable device, a computing device, or other processing device connected to a wireless modem. Terminal devices can be called different names in different networks, such as user equipment, access terminal, subscriber unit, subscriber station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device, cellular phone, cordless phone, device in 5G network or future evolution network, etc.

[0078] Specifically, the visitor processing method includes:

[0079] S102: Determine a visitor arrival event based on multimodal scene monitoring information of the monitored scene area, generate a visitor dialogue plan for the visitor object based on the intelligent dialogue model when the user is away from home, and conduct an autonomous dialogue process with the visitor object based on the visitor dialogue plan to obtain autonomous dialogue content;

[0080] Multimodal scene monitoring information: This refers to comprehensive information about the state of a monitored scene area obtained by electronic devices through data such as video, audio, and sensor data (such as temperature, infrared, and motion detection). Multimodal scene monitoring information can be used to determine the presence of visitors in the area, visitor behavior characteristics, and environmental changes.

[0081] Visitor arrival event: refers to the situation where an electronic device detects a visitor entering or staying in the monitoring area based on multimodal monitoring data.

[0082] The intelligent dialogue big model can be obtained by adapting the visitor processing scenario based on the basic big language model. The basic big language model includes but is not limited to the GPT series big model, DeepSeek series big model, Tongyi Qianwen big model, etc. In some embodiments, the existing basic big language model can also be directly used as the intelligent dialogue big model to combine with different processing task prompts in the intelligent dialogue scenario to process visitor dialogues.

[0083] Visitor Conversation Plan: A generated or pre-set initial conversation script for a specific visitor, including but not limited to pre-set responses such as greetings, inquiries about the visitor's purpose, and information about their current status. Visitor Conversation Plans ensure friendly and effective interactions with visitors while the user is away from home.

[0084] Autonomous conversation content: This refers to the recorded conversation between a visitor and the user, generated after the visitor arrives, based on a pre-generated or pre-set conversation plan. This content reflects information such as the visitor's needs, inquiry methods, and initial responses.

[0085] User away-from-home status: This refers to the system determining that the user is not currently at home through the positioning information of the user's smart devices (such as mobile phones, wearable devices, etc.), door lock status, or feedback from the home security system.

[0086] Schematically, the electronic device integrates and processes the data from various sensors, and uses algorithms such as image recognition, sound signal analysis, and motion detection to determine whether a visitor appears in the monitoring area to obtain a comprehensive judgment result. When the comprehensive judgment result indicates that there is visitor behavior (for example, a human figure is detected, the doorbell rings, or there is continuous movement), the system marks it as a "visitor arrival event." In advance, the electronic device can determine that the user is not at home or is away from home through the positioning information of the user's smart device (such as a mobile phone, wearable device, etc.), the door lock status, or feedback from the home security system, and then trigger the automatic visitor reception mode;

[0087] Furthermore, basic visitor characteristics data is collected from multimodal scene monitoring information to analyze visitor status. This basic visitor characteristics (such as visit time and possible behavior patterns) are integrated with the user's away-from-home status to form contextual information. Based on this contextual information, a conversation plan is constructed and generated as an input prompt for the intelligent conversation model. The conversation plan generation prompt describes the current scenario (such as "user away from home" or "visitor detected") and requires the generation of an appropriate conversation plan. The intelligent conversation model automatically generates a visitor conversation plan based on the provided prompt. For example, the plan includes a standard greeting, inquiries about the visitor's identity and purpose, and guiding statements for subsequent information collection. The electronic device then broadcasts the visitor conversation plan to the visitor object in voice or text form, and conducts an autonomous conversation with the visitor object to obtain autonomous conversation content. During this process, the visitor object responds according to the broadcast prompt, and the system converts the visitor's answers into text information through voice recognition (or direct text input). The electronic device's system records the conversation content of both parties in real time and performs semantic analysis on the visitor's answers, providing data support for subsequent determination of the visitor's visit matters and the user's reception intentions.

[0088] Example: The system uses sensors to confirm the arrival of a visitor and the user has left home. It then creates a prompt: "The user is not at home. Please generate a conversation plan to greet the visitor and confirm the purpose of the visit." The intelligent conversation model returns the plan content, for example: "Hello, welcome. Is your main purpose of visit to deliver a package, visit relatives or friends, or other matters?" The visitor responds: "I am here to deliver a package." The system uses voice recognition to convert the answer into text and record it as autonomous conversation content.

[0089] Through the above method, efficient and intelligent initial interaction with visitors can be automatically carried out when the user leaves home, ensuring that the needs of visitors are met and facilitating subsequent more in-depth interactions and negotiations.

[0090] S104: Determining visitor information based on the autonomous conversation content and multimodal scene monitoring information, predicting a user reception intention based on the visitor information using an intelligent conversation model, and adjusting the visitor conversation plan based on the predicted user reception intention to obtain a future visit negotiation conversation plan;

[0091] Visitor information: This is key visitor information extracted from autonomous conversations and multimodal monitoring data. This information describes the visitor's purpose, needs, and urgency. For example, whether the visitor is delivering a package, visiting relatives or friends, or conducting business negotiations.

[0092] Predicting user reception intention: This means that the system analyzes visitor information and uses the intelligent dialogue model to predict the user's attitude towards the visit and possible reception strategies (such as reception, suggestion to reschedule, or rejection).

[0093] Future Visit Negotiation Dialogue Plan: This is a negotiation interaction plan that is formed by adjusting and optimizing the original visitor dialogue plan based on the predicted user reception intention. This plan is intended to provide guidance for subsequent negotiations with the visitor on the time, method, or other details of the future visit.

[0094] In an illustrative example, combining autonomous conversation content with multimodal monitoring information, using natural language processing (NLP) and data fusion technology, we extract visitor keywords (such as "express delivery," "visit," and "meeting") from the text. At the same time, we use image and sound data to analyze visitor behavior characteristics, classify the extracted information, and generate visitor event information. The structure of visitor event information may include:

[0095] Purpose of visit: for example, delivering express, visiting relatives and friends, business negotiation.

[0096] Urgency: Is it an urgent matter (such as a courier delivery with a tight deadline) or a normal visit?

[0097] Other visitor needs: For example, whether to wait for the user, leave a package, or request a change of appointment.

[0098] Furthermore, the organized visitor information is used to construct an input context. In some embodiments, the input context can incorporate historical reception records or user-preset preference data to assist in prediction. The constructed input context is then integrated into a prompt to construct a plan adjustment, which is then given to the intelligent dialogue model. This prompt instructs the intelligent dialogue model to predict the user's reception intention based on the contextual information. Based on the intelligent dialogue model's predicted reception intention, the original visitor dialogue plan is dynamically adjusted. If the user is predicted to be inclined to not receive the visitor or to suggest rescheduling, the dialogue plan will include corresponding prompts and alternatives. If the user's reception intention is more positive, the plan will include confirmation and further arrangements. The dynamically adjusted text forms a "future visit negotiation dialogue plan" that not only responds to the current visitor's needs but also guides the visitor through the further negotiation process. Through the above process, the visitor's needs can be accurately grasped, the user's intention can be predicted, and ultimately a negotiation plan that meets the needs of both parties can be generated, thereby achieving smarter, safer, and more efficient visitor management when the user is away from home.

[0099] Optionally, the dynamic adjustment process of the intelligent dialogue model can be:

[0100] Text modification: Automatically replace or add key information in the plan. For example, change "Please wait for the user to be received" to "Since the user is currently unable to receive the delivery, we recommend that you choose to reschedule or retain the courier."

[0101] Adjust the tone: Ensure that the tone is consistent with the user's style preferences and also conveys the requirements of security and privacy protection.

[0102] Follow-up instructions: Provide a clear follow-up consultation plan, allowing visitors to choose to make an appointment at a future time or contact an alternative plan.

[0103] Example: The original visitor dialogue plan was: "Hello, is this a courier delivery?" After predicting that the user would like to reschedule, the dialogue was adjusted to: "Hello, I'm currently out and unable to receive a package. Would you like to schedule a delivery this afternoon or tomorrow morning? You can also contact a collection service."

[0104] This adjustment not only clearly explains the user's current status, but also provides visitors with specific next steps, thus forming a plan for future visit negotiation dialogues.

[0105] S106: Recommending a future visit suggestion plan to the visitor object based on the future visit negotiation dialogue plan.

[0106] Future visit negotiation dialogue plan: This refers to the negotiation text generated by adjusting the original visitor dialogue plan in the previous step (S104) based on the autonomous dialogue content, the visitor's visit matters, and the predicted user reception intentions. This plan clearly defines the initial communication and arrangement strategy for the future visitor's visit, such as rescheduling the visit time, proposing alternative plans, or suggesting other response methods.

[0107] Based on the pre-planned future visit negotiation dialogue, the electronic device provides specific recommendations to the visitor for actual coordination, including available visit time periods, reception methods, or other action guidelines. This approach is more practical, making it easier for visitors to make informed choices and providing clear direction for subsequent arrangements.

[0108] As an example, the intelligent conversational model first generates and adjusts a future visit negotiation dialogue plan based on multimodal data and autonomous conversation content. This plan includes the visitor's background, the user's current status, and potential negotiation directions. Based on this plan, the intelligent conversational model negotiates a future visit plan with the visitor. During this process, the intelligent conversational model recommends a specific visit plan based on the plan.

[0109] In an embodiment of the present specification, when a user is away from home, an electronic device generates a visitor conversation plan for a visitor based on an intelligent conversation model, conducts an autonomous conversation with the visitor to obtain autonomous conversation content, determines visitor arrival information based on the autonomous conversation content and multimodal scene monitoring information, then uses the intelligent conversation model to predict the user's reception intention. Based on the predicted user reception intention, the visitor conversation plan is adjusted to obtain a future visit negotiation conversation plan, and then recommends a future visit suggestion plan to the visitor. This enables the electronic device to automatically identify visitor arrival events when the user is away from home, and utilizes multimodal data and the intelligent conversation model to generate and adjust the conversation plan, thereby accurately extracting visitor needs and predicting user reception intentions, thereby providing the visitor with clear and flexible future visit suggestion plans. This not only overcomes the limitations of traditional smart home systems such as access control and intercom systems that can only provide simple one-to-one conversations, effectively reducing the security and privacy risks caused by poor information flow, but also enables intelligent and personalized responses to diverse visitor needs such as express delivery, visits with relatives and friends, and business negotiations, greatly improving communication efficiency and user experience for visitors when the user is away.

[0110] Optionally, the following implementations may be referred to for determining visitor event information based on the autonomous conversation content and multimodal scene monitoring information:

[0111] The autonomous conversation content and the multimodal scene monitoring information are subjected to semantic feature fusion to obtain comprehensive features of the visiting scene, and visitor visit item information is generated based on the comprehensive features of the visiting scene and a preset visitor visit item structure.

[0112] Comprehensive features of the visit scene: refers to the multi-dimensional information representation obtained after the fusion of semantic features, which can fully reflect the current visitor's visit scene. The comprehensive features of the visit scene may include key information in the visitor's language expression (such as keywords such as "express delivery" and "visit"), as well as visitor behavior characteristics detected in multimodal data (such as length of stay, movement patterns, facial expressions, etc.).

[0113] Preset visitor event structure: refers to a predefined set of standardized structures used to classify and describe different types of visitor events, such as "express delivery", "visit by relatives and friends", "business negotiation", etc., and may include additional attributes such as urgency and priority.

[0114] Illustratively, task prompts can be used to leverage the intelligent dialogue model to segment, identify entities, extract keywords, and analyze sentiment in the conversation text between visitors and the system. This extracts key conversation semantic information describing the visitor's purpose, needs, and urgency. Multimodal scene monitoring information, such as images and audio, can be used to extract visitor semantic information, including when the visitor appeared at the door, rang the doorbell, and how long they stayed. Key conversation semantic information and visitor semantic information extracted from the conversation text and multimodal data are then integrated to form a unified "visit scene comprehensive feature." This integrated "visit scene comprehensive feature" not only retains the visitor's linguistically expressed intentions but also includes supplementary information on behavior, environment, and other aspects, thereby more accurately reflecting the overall situation of the current visitor's visit.

[0115] Furthermore, a set of standard structures for visitor matters are pre-defined, for example, visitor matters are divided into categories (express delivery, visit, business, etc.), urgency, expected interaction methods and other fields. The intelligent dialogue model can map the integrated features of the visit scene into this preset structure based on the preset visitor matter structure, thereby generating visitor matter information.

[0116] As you can see, by integrating semantic features from autonomous conversation content with multimodal scene monitoring information, the system is able to generate a comprehensive and accurate comprehensive profile of visitor scenarios. This comprehensive profile is then matched with the pre-defined visitor activity structure, ultimately generating structured visitor activity information. This process not only enhances understanding of visitors' true needs but also provides a solid data foundation for subsequent intelligent predictions and dialogue plan adjustments, enabling more accurate, intelligent, and personalized visitor management.

[0117] In one possible implementation, see Figure 2 , Figure 2 This is a flow chart of a visitor event information generation process proposed in this specification. Specifically, the semantic feature fusion of the autonomous conversation content and multimodal scene monitoring information is performed to obtain the visitor scene comprehensive features, and the visitor event information is generated based on the visitor scene comprehensive features and the preset visitor event structure. The following methods can be used:

[0118] S202: performing semantic analysis on the autonomous conversation content to obtain visitor conversation semantic information, and performing scene semantic analysis on the multimodal scene monitoring information to obtain visitor scene semantic information;

[0119] Visitor conversation semantic information: This refers to information obtained by semantically analyzing autonomous conversation content (the text of interactions between visitors and the system) using natural language processing technology or intelligent conversation models. This information may include key information such as the visitor's purpose, request, and emotional state expressed in the conversation.

[0120] Visitor scene semantic information: This refers to semantic information about visitor behavior, environmental status, and scene characteristics extracted by parsing multimodal scene monitoring information (such as video, audio, and sensor data). For example, the length of time a visitor stays at the door, their movement patterns, or background environmental characteristics.

[0121] S204: performing semantic fusion based on the visitor conversation semantic information and the visitor scene semantic information to obtain a visitor scene comprehensive feature, and determining the visitor identity, visitor purpose, and environmental context information based on the visitor scene comprehensive feature;

[0122] In schematic form, the semantic information of the visiting conversation and the semantic information of the visiting scene are semantically fused through the feature fusion mechanism of the intelligent dialogue model (such as multimodal Transformer, neural network fusion model) to obtain the comprehensive features of the visiting scene, and the visitor identity and purpose of the visit and the environmental context information (such as length of stay, activity trajectory, door status, etc.) are analyzed to obtain the visitor identity, purpose of visit and environmental context information.

[0123] S206: extracting key visit item fields based on the visitor identity, the visit purpose and the environmental context information, mapping the key visit item fields to a preset visitor visit item structure, and obtaining visitor visit item information.

[0124] Key visit item fields: refers to the core information fields extracted from the comprehensive characteristics of the visit scene. These fields usually include visitor identity, visit purpose, urgency, and environmental context information.

[0125] Preset Visitor Item Structure: This is a predefined, standardized structure used to categorize and describe visitor items. This structure may contain multiple fields, such as "Visit Category," "Visitor Identity," "Urgency," and "Related Notes."

[0126] Indicatively, the core fields related to visitor arrivals are automatically extracted from the integrated features, and the extracted key fields are mapped according to the preset visitor arrival event structure to generate a structured visitor arrival event information.

[0127] In the embodiments of this specification, the system can combine the intelligent dialogue model to first analyze the autonomous dialogue content and multimodal scene monitoring information separately to obtain the semantic information of the visitor's language expression and environmental behavior; then fuse the two to generate comprehensive features including the visitor's identity, purpose of visit and environmental context; finally, extract the key visit matter fields from them, and map them according to the preset visitor visit matter structure to generate structured visitor visit matter information, which not only improves the accurate identification of visitor information, but also provides solid data support for subsequent intelligent decision-making and negotiation.

[0128] Optional, see Figure 3 , Figure 3 This is a flowchart of predicting the user's reception intention. Specifically, the method of using the intelligent dialogue model to predict the user's reception intention based on the visitor's visit information can be referred to as follows:

[0129] S302: Obtain the user's historical visit interaction records and user visitor setting preference information;

[0130] Historical visitor interaction records: This refers to all conversation records, interaction logs, and feedback data generated during past interactions between users and visitors. These records reflect the user's handling and response tendencies in different visitor scenarios.

[0131] User visitor preference information: refers to the preferences and rules regarding visitor reception that the user pre-sets in the system, such as the types of visitors allowed, the list of visitors who are rejected, special requirements for express delivery or visits from relatives and friends, reception time restrictions, etc.

[0132] Illustratively, historical interaction records are retrieved from the database. These records may include user feedback from previous visitors, records of automatic system processing, and user evaluation feedback. At the same time, the user's visitor reception preferences set in the smart home app or system backend are read. These preferences are determined by the user during initial configuration or subsequent updates.

[0133] S304: Based on the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, the intelligent dialogue big model is used to infer the predicted user reception intention for the visitor object.

[0134] Schematically, historical interaction records and user visitor preference information are integrated into a unified context input, task prompt words are generated based on the context input, and the task prompt words are input into the intelligent dialogue big model to instruct the intelligent dialogue big model to infer the predicted user reception intention for the visitor object. The intelligent dialogue big model uses semantic understanding and reasoning, combined with context information for internal calculation to obtain the predicted user reception intention. The predicted user reception intention can be a text description (such as "the user reception intention is to suggest changing the appointment or using the collection service"), or it can be a structured result (such as the probability distribution marked as "refuse on-site reception").

[0135] In this specification, in step S302, the electronic device system extracts key information from historical visitor interaction records and user visitor preferences, providing data support for subsequent intelligent reasoning. Then, in step S304, the visitor's visit information is combined with the integrated historical records and preference information, and the intelligent dialogue model is used to perform reasoning and predict the user's reception intention. This process ensures that the system can intelligently predict and adjust visitor response strategies based on the user's past behavior and current settings, achieving more personalized, secure, and efficient visitor management.

[0136] In one possible implementation, see Figure 4 , Figure 4 This is a flowchart of intention reasoning, specifically implementing the predicted user reception intention for the visitor object based on the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, using the intelligent dialogue large model to reason, which can refer to the following methods:

[0137] S402: Extracting prediction task elements from the visitor's visit event information, the historical visit interaction records, and the user visitor setting preference information to obtain visitor prediction task elements, constructing a visitor interest prediction reasoning chain based on the visitor interest prediction reasoning chain, and generating a visitor interest prediction task prompt based on the visitor interest prediction reasoning chain;

[0138] Prediction task elements: refers to the core factors extracted from the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, which have a key impact on predicting the user's reception intention, including the main characteristics of the visitor's visit, the decision-making pattern in the historical interaction and the user's preference settings.

[0139] Interest prediction task thinking template: It is a pre-designed logical framework or thinking model used to guide the intelligent dialogue model to evaluate the visitor's visit situation from multiple perspectives, so as to determine the user's possible interest in or reception intention of the current visitor.

[0140] For example, a preset thinking template for interest prediction tasks is as follows:

[0141] 1. Analyze the matching degree between the visitor's visit purpose and the user's historical reception behavior;

[0142] 2. Evaluate user preferences for handling strategies for specific visitor types;

[0143] 3. Consider the impact of the current situation (such as waiting time and environmental conditions) on the user's reception intention.

[0144] Visitor interest prediction reasoning chain: refers to a series of reasoning steps constructed based on prediction task elements and thinking templates, forming a logical chain, which is used to gradually analyze and deduce the user's reception intention in the current situation.

[0145] Visitor Interest Prediction Task Prompts: These are the prompts generated through the aforementioned reasoning chain, guiding the intelligent conversation model's step-by-step reasoning. These prompts guide the model along the pre-set reasoning chain, ultimately outputting a prediction of the user's reception intention.

[0146] Predicting user reception intentions: This means that the system uses large-scale model reasoning to predict the reception strategies that users may adopt in the current situation, such as on-site reception, suggestion to reschedule, or refusal to receive the user.

[0147] In a nutshell, we filter visitor information, historical visitor interaction records, and user visitor preference settings to identify key factors influencing user decisions, such as visitor identity, visit purpose, wait time, past user responses in similar scenarios, and explicit user preferences. These key factors are then integrated into a structured set of prediction task elements. The "Interest Prediction Task Thinking Template" is then used to construct a visitor interest prediction reasoning chain. Specifically, following the steps of the thinking template, the extracted task elements are organized into a clear visitor interest prediction reasoning chain, ensuring that each step has a logical basis and ultimately deriving the user's reception intention.

[0148] Furthermore, based on the constructed reasoning chain, the system generates a set of prompt words for the visitor interest prediction task. These prompt words describe the current situation, key elements, and expected reasoning path in natural language, providing detailed contextual guidance for the intelligent dialogue model.

[0149] Example prompts: For example, "The current visitor is a courier. Historical records show that users generally do not support impromptu signatures, and their preferences explicitly reject on-site signatures. Based on this information, please infer that the user may suggest changing the appointment or using a collection service."

[0150] S404: Input the prompt words of the visit interest prediction task into the intelligent dialogue model, and use the intelligent dialogue model to gradually infer the user's visit interest according to the visit interest prediction reasoning chain to obtain the predicted user reception intention.

[0151] Furthermore, the prompt word for the visitor interest prediction task is input into the intelligent dialogue model. The model then considers the inference chain structure in the prompt word, evaluating the visitor's purpose, historical interaction patterns, and user preferences. It then conducts layer-by-layer reasoning based on the current context, and gradually generates: In this process, the model first identifies the visitor's category, then determines the user's typical reactions in similar situations, and finally, comprehensively analyzes all this information to predict the user's most likely reception intention. The intelligent dialogue model outputs the predicted user reception intention, which can be a specific text description or a structured decision, such as "recommend rescheduling" or "refuse on-site reception."

[0152] In this specification, through steps S402 and S404, the electronic device first extracts key prediction task elements from visitor information, historical interaction records, and user preferences. Using the interest prediction task thinking template, it constructs a clear chain of reasoning and generates detailed task prompts. This prompt is then input into the intelligent dialogue model, which gradually derives the user's reception intentions in the current context based on the pre-set chain of reasoning. This entire process integrates multi-dimensional complex information and logical reasoning, ensuring that the prediction results accurately reflect user preferences and provide a reliable basis for subsequent visitor negotiation and processing.

[0153] Optional, see Figure 5 , Figure 5 This is a flowchart of adjusting the plan. Specifically, the adjustment of the visitor dialogue plan based on the predicted user reception intention to obtain the future visit negotiation dialogue plan can be performed in the following manner:

[0154] S502: Determine the target user intention type corresponding to the predicted user reception intention, obtain recommended conversation features corresponding to the target user intention type, and obtain the user's visitor reception preference information and homecoming schedule information;

[0155] Target User Intention Type: This refers to the specific intent category mapped to the user's reception intention based on the intelligent conversation model's predictions, such as "rescheduling," "on-site reception," or "rejecting reception." This type reflects how the user would like to handle their visit request in the current context.

[0156] Recommended conversation features: These refer to the suggested conversation styles and wording requirements extracted by the system based on the target user's intention type, such as tone, politeness, information detail, and how to guide visitors to choose subsequent solutions.

[0157] User visitor reception preference information: refers to the personalized requirements for visitor reception pre-set by the user in the system, such as allowing or rejecting certain types of visitors, common response strategies, or special handling methods for certain visitor requests.

[0158] Home schedule information: refers to the user's home schedule or schedule information, which is used to determine when the user can receive visitors, thereby affecting the time options for negotiation plans.

[0159] Indicatively, the user reception intention predicted according to the previous steps corresponds to a specific target user intention type preset in the system. For example, if the prediction result is "recommendation to reschedule", then the target intention type is "reception for reschedule". Then, for each intention type, the system presets corresponding dialogue characteristics, such as tone, wording, and information guidance strategies. For example, for "reception for reschedule", recommended dialogue characteristics may include: being gentle and polite, explaining the user's current status, and proactively providing optional appointment time periods. Furthermore, the visitor reception preference information is extracted from the user settings to understand the user's handling principles in specific scenarios. At the same time, the user's home schedule information is obtained to determine when the user can receive visitors, so as to provide a time basis for the negotiation plan.

[0160] Example: Assuming the predicted user reception intention is "rescheduling reception," the system determines the target user's intention type as "rescheduling reception." The user's visitor reception preferences indicate that they typically prefer to schedule a specific appointment rather than sign for deliveries on short notice. Their homecoming schedule indicates that they expect to be home after 6 PM. Therefore, the system determines the recommended conversation characteristics to be gentle, explanatory, and clearly inform the visitor that they can schedule a reception after 6 PM.

[0161] S504: Generate a visit negotiation plan adjustment prompt based on the recommended conversation characteristics, the visitor reception preference information, the home schedule information and the visitor visit matters information, input the visit negotiation plan adjustment prompt into the intelligent conversation big model, use the intelligent conversation big model to perform feature analysis processing to obtain a plan adjustment context vector, and adjust the visitor conversation plan based on the plan adjustment context vector to obtain a future visit negotiation conversation plan.

[0162] Prompt words for adjusting visit negotiation plans: refers to a set of guiding prompts generated based on the characteristics of the recommended conversation, user preferences, home schedule information, and visitor visit information. These prompt words are used to guide the intelligent conversation model to perform reasoning and generate the context for plan adjustments.

[0163] Plan adjustment context vector: It is a multidimensional vector representation formed after the intelligent dialogue model undergoes feature analysis. It comprehensively reflects key information such as the current visitor context, user preferences, and future reception arrangements, providing a basis for subsequent adjustments to the original visitor dialogue plan.

[0164] Future visit negotiation dialogue plan: refers to the dialogue strategy adjusted based on the original visitor dialogue plan according to the predicted user reception intention, user preferences and return schedule, to guide the subsequent negotiation with the visitor for future visit arrangements.

[0165] In an illustrative manner, the recommended conversation features, user visitor reception preference information, homecoming schedule information and visitor visit information obtained in S502 are integrated to construct a set of detailed task prompts. These prompts describe the current visitor situation, user preferences and schedule restrictions in natural language, and explicitly require the generation of adjustment suggestions.

[0166] For example, the prompt might read: "The visitor is a courier and has been waiting for a long time. The user's preference is not to support temporary on-site signing, and the return time is after 6 pm. Please generate a schedule change negotiation plan to remind the visitor to make an appointment to receive the courier between 6 pm and 8 pm."

[0167] Furthermore, the generated prompts for adjusting the visit negotiation plan are fed into the intelligent conversation model. The model then analyzes the prompts to generate a context vector for adjusting the plan. This context vector is then used to adjust the original visitor conversation plan. These adjustments include updating the response, adding an appointment time option, and clarifying the user's current status. Ultimately, this creates a plan for future visit negotiation conversations. This plan guides the system in subsequent communications with the visitor, ensuring that the user is unable to meet in person while providing clear options for future appointments.

[0168] In this specification, in step S502, the system first determines the target user's intended type based on the predicted user's reception intentions, and then extracts recommended conversation features based on the user's visitor reception preferences and home schedule information. Then, in step S504, the system generates prompts for adjusting the visit negotiation plan based on this information and the visitor's visit details. The prompts are input into the intelligent conversation model, and the model is used to generate a plan adjustment context vector. This allows precise adjustments to the original visitor conversation plan, generating a future visit negotiation conversation plan that meets the needs of both parties. This process ensures that the system can intelligently coordinate visitor needs with the user's actual reception capabilities when the user is away from home, achieving efficient, secure, and personalized visitor management.

[0169] In one possible implementation, see Figure 6 , Figure 6 This is a flow chart of feature analysis processing. Specifically, the following methods can be used to perform feature analysis processing using the intelligent dialogue model to obtain a context vector for adjusting the plan:

[0170] S602: Using the intelligent dialogue model to perform structured extraction processing on input data to obtain entity nodes and entity relationship edges of incoming items;

[0171] The visit item entity node refers to the key entities related to the visit of the visitor extracted from the input data, such as “visitor identity”, “visit purpose”, “waiting time”, “urgency”, etc.

[0172] Entity relationship edges represent semantic associations or logical relationships between entities, such as the relationship between "visitor identity" and "visit purpose", or the relationship between "waiting time" and "urgency".

[0173] For example, the input data includes multimodal data such as autonomous conversation content, scene monitoring information, and user preferences. Leveraging the natural language processing and multimodal processing capabilities built into the intelligent conversation model, named entity recognition (NER) and relationship extraction are performed on the input data. For example, from "I'm here to deliver a courier," "courier" is extracted as the visit purpose entity; from image and sensor data, "long wait" is extracted as the wait duration entity.

[0174] According to the extraction results, the corresponding entity nodes are constructed, and the relationship edges between entities are identified and saved as structured data.

[0175] S604: Constructing a plan adjustment knowledge graph based on the entity nodes and entity relationship edges of the incoming event, performing node embedding coding processing on the plan adjustment knowledge graph to obtain a node encoding vector, and performing edge relationship association conversion processing on the plan adjustment knowledge graph to obtain a node relationship weight matrix;

[0176] Plan adjustment knowledge graph: The graph constructed using the entity nodes and relationship edges extracted in S602 can intuitively represent the relationship between the various elements of the visitor's visit.

[0177] Node embedding encoding vector: Each entity node in the graph is converted into a fixed-dimensional vector representation through an embedding method for subsequent calculations.

[0178] Node relationship weight matrix: The edges between entities in the graph are weighted to form a matrix that represents the strength of the relationship or degree of association between each node.

[0179] Schematically, a contingency plan adjustment knowledge graph is constructed based on the obtained entity nodes and relationship edges. Nodes in the contingency plan adjustment knowledge graph represent key elements, and edges represent the logical or semantic connections between them. Using a graph neural network, each node is converted into a low-dimensional vector representation. These vectors capture the node's feature information and its structural information within the graph. The edges in the contingency plan adjustment knowledge graph are then transformed into relationship relationships, generating a node relationship weight matrix. Each value in the matrix reflects the strength of the relationship between the corresponding edge.

[0180] S606: Based on the node encoding vector and the node relationship weight matrix, an attention mechanism is used to perform optimization to obtain a plan adjustment context vector.

[0181] Plan adjustment context vector: A comprehensive vector obtained by fusing the node embedding vector and the node relationship weight matrix through the attention mechanism. It comprehensively represents the key information of the visitor's visit and the interaction between various elements, providing contextual support for subsequent plan adjustments.

[0182] Attention mechanism: A mechanism that assigns different weights to input features, aggregates multi-dimensional information into a vector representation through weighted summation, and thus highlights important features.

[0183] Schematically, the node embedding vector and the node relationship weight matrix are used as input, and the attention mechanism is used to calculate the importance of each node to the overall context. Specifically, each node's vector is weighted according to the relationship weights with other nodes, resulting in an attention distribution. Using the calculated attention weights, the embedding vectors of each node are weighted and summed to generate a comprehensive plan adjustment context vector. This context vector captures the global semantics of all key information and reflects the relevance between nodes. The final output plan adjustment context vector will serve as the basis for subsequent adjustments to the visitor dialogue plan by the large model, ensuring that the adjustment plan fully reflects the details of the visit and the user's requirements.

[0184] In one possible implementation, see Figure 7 , Figure 7 This is a flow chart of visit negotiation processing. Specifically, the process of adjusting the visitor dialogue plan based on the plan adjustment context vector to obtain a future visit negotiation dialogue plan can be performed in the following manner:

[0185] S702: Determine, based on the plan adjustment context vector, a multidimensional dialogue adjustment semantic slot corresponding to the dialogue plan structure of the visitor dialogue plan, and generate a visit negotiation dialogue based on the multidimensional dialogue adjustment semantic slot to obtain an adjusted visit negotiation dialogue;

[0186] The context vector for plan adjustment is a vector representation obtained in the previous steps (such as S602, S604, and S606) that comprehensively represents key visitor information (such as visitor identity, visit purpose, waiting time, etc.) as well as user-related preferences, schedule, and other information. It provides a global context basis for subsequent plan adjustments.

[0187] Multi-dimensional conversation adjustment semantic slots: These refer to multiple semantic fields or "slots" preset in the visitor conversation plan structure that require dynamic adjustment, such as reception instructions, time reminders, suggested appointment times, and voice tone. These slots are filled or updated based on the information in the plan adjustment context vector to meet the communication needs of the current situation.

[0188] Adjusted visit negotiation dialogue: refers to the temporary dialogue text generated based on the above semantic slots, which includes negotiation content adjusted according to the visitor's current visit situation, user preferences, return schedule and other factors, and reflects new response strategies and appointment suggestions.

[0189] Indicatively, the context vector is adjusted using the plan, and the structure of the original visitor dialogue plan is analyzed through the intelligent dialogue model to determine which parts need to be adjusted. For example, the original plan may contain a "waiting instructions" slot and a "reception time suggestion" slot. The intelligent dialogue model is used to determine the direction and focus of the adjustment based on the key information contained in the context vector (such as the user's current status, historical preferences, home schedule, etc.). For example, if the vector shows that the user is not at home and tends to reschedule, the "reception time suggestion" slot will need to be updated to a future appointment time, while the "waiting instructions" slot needs to include a gentle reminder to reschedule.

[0190] Furthermore, the intelligent dialogue model is controlled to construct multi-dimensional dialogue adjustment semantic slots. In this process, the slots that need to be adjusted can be divided into multiple dimensions based on the analysis results, such as:

[0191] Status slot: Describes the user's current status (e.g., "The user is out and expected to be home after 6 pm"). Appointment suggestion slot: Provides specific appointment time options (e.g., "It is recommended to make an appointment between 6 and 8 pm tonight"). Interaction tone slot: Determines the overall tone and wording style (e.g., "gentle and polite").

[0192] Furthermore, the intelligent dialogue model is controlled to adjust the semantic slots based on the multi-dimensional dialogue to generate a visit negotiation dialogue. The intelligent dialogue model generates an adjusted visit negotiation dialogue text according to the slot prompts and context information. The visit negotiation dialogue text reflects the response to the visitor's needs and the arrangement of the subsequent negotiation process.

[0193] Example: Assume that the original visitor dialogue plan is: "Hello, please wait, the user will be back to greet you soon."

[0194] The context vector representation of the plan adjustment obtained in the previous step indicates that the visitor is a courier and has a long waiting time. Neither the user's historical preferences nor the current settings support temporary signing, and the user's home time is after 6 pm.

[0195] The intelligent dialogue model determines the following semantic slots in S702:

[0196] Status description slot: Updated to "Hello, the user is currently out and expected to be home after 6pm tonight." Appointment suggestion slot: Added suggestion "Would you like to make an appointment to receive a delivery between 6pm and 8pm tonight, or contact a collection service?" Interaction tone slot: Keep the tone gentle and polite.

[0197] Based on these slots, the intelligent dialogue model generates the following adjustment visit negotiation dialogue:

[0198] "Hello, you are currently out and are expected to return home after 6pm tonight. To ensure that you receive your package on time, would you like to make an appointment to receive your package between 6pm and 8pm tonight? Alternatively, you can contact the collection service."

[0199] S704: Based on the adjusted visit negotiation dialogue, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0200] Schematically, the intelligent conversation model uses the adjusted visit negotiation dialogue generated in S702 as its core content, replacing or integrating the original visitor conversation plan. For example, the simple waiting prompt in the original plan is replaced with a new appointment suggestion and status description, forming a conversation plan more appropriate to the current situation. The adjusted negotiation dialogue text is embedded into the overall visitor conversation process, forming a structurally complete and logically coherent future visit negotiation dialogue plan.

[0201] Example: Continuing with the previous example, the original visitor dialogue plan is: "Hello, please wait, the user will be back to greet you soon."

[0202] The negotiation dialogue generated after the adjustment in S702 is: "Hello, the user is currently out and is expected to return home after 6 pm tonight. To ensure that you can receive the express delivery in time, would you like to make an appointment to receive the express delivery between 6 pm and 8 pm tonight? Alternatively, you can choose to contact the collection service."

[0203] In S704, the system replaces the simple prompt in the original plan with this adjusted text, and the final generated future visit negotiation dialogue plan is:

[0204] "Hello, the user is currently out and is expected to return home after 6pm tonight. To ensure that you can receive the express delivery in time, would you like to make an appointment to receive the express delivery between 6pm and 8pm tonight? Alternatively, you can choose to contact the collection service. Please reply with your choice so that the system can make further arrangements."

[0205] In this specification, in step S702, the system uses the plan adjustment context vector to determine the multi-dimensional semantic slots that need to be adjusted in the original visitor dialogue plan, and generates an adjusted visit negotiation dialogue that fully integrates information such as the user's status, the visitor's needs, preferences, and return schedule. Next, in step S704, the system integrates this adjusted negotiation dialogue back into the original dialogue plan to generate the final future visit negotiation dialogue plan, ensuring that the system can intelligently and accurately conduct subsequent communication and appointment negotiations with the visitor while the user is away from home.

[0206] The following will be combined Figure 8 , the visitor processing device provided in the embodiment of this specification is introduced in detail. It should be noted that, Figure 8 The visitor processing device shown is used to execute this instruction Figures 1 to 7 For the convenience of explanation, only the part related to the embodiment of this specification is shown. For the specific technical details not disclosed, please refer to this specification. Figures 1 to 7 The embodiment shown.

[0207] See Figure 8 , which shows a schematic diagram of the structure of a visitor processing device according to an embodiment of this specification. The visitor processing device 1 can be implemented as all or part of a device through software, hardware, or a combination of both. According to some embodiments, the visitor processing device 1 includes a dialogue module 11, a negotiation module 12, and a recommendation module 13, which are specifically used to:

[0208] The dialogue module 11 is configured to determine a visitor arrival event based on the multimodal scene monitoring information of the monitoring scene area, generate a visitor dialogue plan for the visitor object based on the intelligent dialogue model when the user is away from home, and conduct an autonomous dialogue with the visitor object based on the visitor dialogue plan to obtain autonomous dialogue content;

[0209] Negotiation module 12, configured to determine visitor visit matter information based on the autonomous conversation content and multimodal scene monitoring information, predict the user reception intention based on the visitor visit matter information using an intelligent conversation model, and adjust the visitor conversation plan based on the predicted user reception intention to obtain a future visit negotiation conversation plan;

[0210] The recommendation module 13 is configured to recommend a future visit suggestion plan to the visitor object based on the future visit negotiation dialogue plan.

[0211] In a feasible implementation manner, the determining visitor visit information based on the autonomous conversation content and multimodal scene monitoring information includes:

[0212] The autonomous conversation content and the multimodal scene monitoring information are subjected to semantic feature fusion to obtain comprehensive features of the visiting scene, and visitor visit item information is generated based on the comprehensive features of the visiting scene and a preset visitor visit item structure.

[0213] In a feasible implementation, the semantic feature fusion of the autonomous conversation content and the multimodal scene monitoring information is performed to obtain a comprehensive feature of the visitor scene, and the visitor event information is generated based on the comprehensive feature of the visitor scene and a preset visitor event structure, including:

[0214] Performing semantic analysis on the autonomous conversation content to obtain visitor conversation semantic information, and performing scene semantic analysis on the multimodal scene monitoring information to obtain visitor scene semantic information;

[0215] Perform semantic fusion based on the visitor conversation semantic information and the visitor scene semantic information to obtain a visitor scene comprehensive feature, and determine the visitor identity, visit purpose, and environmental context information based on the visitor scene comprehensive feature;

[0216] Key visit item fields are extracted based on the visitor identity, the visit purpose and the environmental context information, and the key visit item fields are mapped to a preset visitor visit item structure to obtain visitor visit item information.

[0217] In a feasible implementation, the method of using an intelligent dialogue model to determine the predicted user reception intention based on the visitor's visit information includes:

[0218] Obtain the user's historical visit interaction records and user visitor setting preference information;

[0219] Based on the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, an intelligent dialogue large model is used to infer the predicted user reception intention for the visitor object.

[0220] In a feasible implementation, the predicting of the user reception intention for the visitor object based on the visitor event information, the historical visit interaction records, and the user visitor setting preference information using the intelligent dialogue large model inference includes:

[0221] Extracting prediction task elements from the visitor's visit event information, the historical visit interaction records, and the user visitor setting preference information to obtain visit prediction task elements, constructing a visit interest prediction reasoning chain for the visit prediction task elements using an interest prediction task thinking template, and generating a visit interest prediction task prompt word based on the visit interest prediction reasoning chain;

[0222] The visit interest prediction task prompt words are input into the intelligent dialogue model, and the user's visit interest is gradually inferred by the intelligent dialogue model according to the visit interest prediction reasoning chain to obtain the predicted user reception intention.

[0223] In a feasible implementation, adjusting the visitor dialogue plan based on the predicted user reception intention to obtain a future visit negotiation dialogue plan includes:

[0224] Determine the target user intention type corresponding to the predicted user reception intention, obtain the recommended conversation features corresponding to the target user intention type, and obtain the user's visitor reception preference information and home schedule information;

[0225] Based on the recommended dialogue characteristics, the visitor reception preference information, the home schedule information and the visitor visit matters information, a visit negotiation plan adjustment prompt is generated, the visit negotiation plan adjustment prompt is input into the intelligent dialogue big model, the intelligent dialogue big model is used to perform feature analysis processing to obtain a plan adjustment context vector, and based on the plan adjustment context vector, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0226] In a feasible implementation, the step of using the intelligent dialogue macro model to perform feature analysis to obtain a plan adjustment context vector includes:

[0227] The intelligent dialogue model is used to perform structured extraction of input data to obtain entity nodes and entity relationship edges of incoming items;

[0228] Constructing a plan adjustment knowledge graph based on the entity nodes and entity relationship edges of the visiting matter, performing node embedding coding processing on the plan adjustment knowledge graph to obtain a node encoding vector, and performing edge relationship association conversion processing on the plan adjustment knowledge graph to obtain a node relationship weight matrix;

[0229] Based on the node encoding vector and the node relationship weight matrix, an attention mechanism is used to perform optimization to obtain a plan adjustment context vector.

[0230] In a feasible implementation manner, adjusting the visitor dialogue plan based on the plan adjustment context vector to obtain a future visit negotiation dialogue plan includes:

[0231] Determining a multidimensional dialogue adjustment semantic slot corresponding to the dialogue plan structure of the visitor dialogue plan based on the plan adjustment context vector, and generating a visit negotiation dialogue based on the multidimensional dialogue adjustment semantic slot to obtain an adjusted visit negotiation dialogue;

[0232] Based on the adjustment of the visit negotiation dialogue, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

[0233] It should be noted that the visitor processing device provided in the above embodiment, when executing the visitor processing method, is merely illustrated by the division of the aforementioned functional modules. In actual applications, the aforementioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the visitor processing device provided in the above embodiment and the visitor processing method embodiment are based on the same concept. The implementation process is detailed in the method embodiment and will not be repeated here.

[0234] The serial numbers of the embodiments in this specification are for description only and do not represent the advantages or disadvantages of the embodiments.

[0235] The embodiment of this specification also provides a computer storage medium, which can store multiple instructions, and the instructions are suitable for being loaded and executed by a processor as described above. Figures 1 to 7 The visitor processing method of the embodiment shown in the figure can be found in the specific execution process. Figures 1 to 7 The detailed description of the illustrated embodiment will not be repeated here.

[0236] This specification also provides a computer program product, which stores at least one instruction, and the at least one instruction is loaded and executed by the processor as described above. Figures 1 to 7 The visitor processing method of the embodiment shown in the figure can be found in the specific execution process. Figures 1 to 7 The detailed description of the illustrated embodiment will not be repeated here.

[0237] Please refer to Figure 9 , which shows a block diagram of the structure of an electronic device provided by an exemplary embodiment of this specification. The electronic device described in this specification may include one or more of the following components: a processor 110, a memory 120, an input device 130, an output device 140, and a bus 150. The processor 110, the memory 120, the input device 130, and the output device 140 may be connected via the bus 150.

[0238] The processor 110 may include one or more processing cores. The processor 110 utilizes various interfaces and circuits to connect various components within the electronic device. It executes instructions, programs, code sets, or instruction sets stored in the memory 120, as well as accesses data stored in the memory 120, to perform various functions of the electronic device and process data. Optionally, the processor 110 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 110 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem handles wireless communications. It is understood that the modem may not be integrated into the processor 110 and may be implemented separately via a communications chip.

[0239] The memory 120 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 120 includes a non-transitory computer-readable storage medium. The memory 120 may be used to store instructions, programs, codes, code sets, or instruction sets. The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc. The operating system may be an Android system, including a system deeply developed based on the Android system, an IOS system developed by Apple, including a system deeply developed based on the IOS system or other systems. The data storage area may also store data created by the electronic device during use, such as a phone book, audio and video data, chat record data, etc.

[0240] See also Figure 10 As shown, the memory 120 can be divided into operating system space and user space. The operating system runs in the operating system space, and native and third-party applications run in the user space. In order to ensure that different third-party applications can achieve better operating results, the operating system allocates corresponding system resources to different third-party applications. However, the requirements for system resources in different application scenarios in the same third-party application are also different. For example, in the local resource loading scenario, the third-party application has higher requirements for disk reading speed; in the animation rendering scenario, the third-party application has higher requirements for GPU performance. The operating system and the third-party application are independent of each other, and the operating system often cannot perceive the current application scenario of the third-party application in a timely manner, resulting in the operating system being unable to perform targeted system resource adaptation according to the specific application scenario of the third-party application.

[0241] In order for the operating system to distinguish the specific application scenarios of third-party applications, it is necessary to open up data communication between third-party applications and the operating system so that the operating system can obtain the current scenario information of third-party applications at any time, and then perform targeted system resource adaptation based on the current scenario.

[0242] Taking the Android operating system as an example, the programs and data stored in the memory 120 are as follows: Figure 11As shown, the memory 120 may store a Linux kernel layer 320, a system runtime library layer 340, an application framework layer 360, and an application layer 380. The Linux kernel layer 320, the system runtime library layer 340, and the application framework layer 360 belong to the operating system space, and the application layer 380 belongs to the user space. The Linux kernel layer 320 provides underlying drivers for various hardware components of electronic devices, such as display drivers, audio drivers, camera drivers, Bluetooth drivers, Wi-Fi drivers, power management, etc. The system runtime library layer 340 provides major feature support for the Android system through some C / C++ libraries. For example, the SQLite library provides database support, the OpenGL / ES library provides 3D drawing support, and the Webkit library provides browser kernel support. The system runtime library layer 340 also provides the Android runtime library (Android runtime), which mainly provides some core libraries that allow developers to write Android applications using the Java language. The application framework layer 360 provides various APIs that may be used when building applications. Developers can also use these APIs to build their own applications, such as activity management, window management, view management, notification management, content provider management, package management, call management, resource management, and location management. The application layer 380 runs at least one application. These applications can be native applications that come with the operating system, such as contacts, SMS, clock, and camera applications, or third-party applications developed by third-party developers, such as games, instant messaging programs, and photo enhancement programs.

[0243] Taking the operating system as the IOS system as an example, the programs and data stored in the memory 120 are as follows: Figure 10As shown, the IOS system includes: a core operating system layer 420 (Core OS layer), a core service layer 440 (Core Services layer), a media layer 460 (Media layer), and a touchable layer 480 (Cocoa Touch Layer). The core operating system layer 420 includes the operating system kernel, drivers, and underlying program frameworks. These underlying program frameworks provide functions closer to the hardware for use by the program framework located in the core service layer 440. The core service layer 440 provides system services and / or program frameworks required by applications, such as the foundation framework, account framework, advertising framework, data storage framework, network connection framework, geographic location framework, motion framework, etc. The media layer 460 provides applications with audio-visual interfaces, such as graphics and image-related interfaces, audio technology-related interfaces, video technology-related interfaces, and wireless playback (AirPlay) interfaces for audio and video transmission technologies. The touchable layer 480 provides various commonly used interface-related frameworks for application development. The touchable layer 480 is responsible for user touch interaction operations on electronic devices. For example, local notification service, remote push service, advertising framework, game tool framework, message user interface (UI) framework, user interface UIKit framework, map framework, etc.

[0244] exist Figure 12 Among the frameworks shown, those relevant to most applications include, but are not limited to, the Foundation framework in the core services layer 440 and the UIKit framework in the touchable layer 480. The Foundation framework provides many basic object classes and data types, offering fundamental system services for all applications and having nothing to do with the UI. The classes provided by the UIKit framework are the foundational UI class library for creating touch-based user interfaces. iOS applications can use the UIKit framework to provide their UIs, providing the application infrastructure for building user interfaces, drawing, handling user interaction events, responding to gestures, and so on.

[0245] Among them, the method and principle of implementing data communication between third-party applications and the operating system in the IOS system can be referred to the Android system, and this manual will not go into details here.

[0246] Among them, the input device 130 is used to receive input instructions or data, and the input device 130 includes but is not limited to a keyboard, a mouse, a camera, a microphone or a touch device. The output device 140 is used to output instructions or data, and the output device 140 includes but is not limited to a display device and a speaker. In one example, the input device 130 and the output device 140 can be combined, and the input device 130 and the output device 140 are touch screen displays, which are used to receive touch operations on or near the user using any suitable objects such as fingers and touch pens, and to display the user interface of each application. The touch screen display is usually provided on the front panel of the electronic device. The touch screen display can be designed as a full screen, a curved screen or a special-shaped screen. The touch screen display can also be designed as a combination of a full screen and a curved screen, or a combination of a special-shaped screen and a curved screen, which is not limited in the embodiments of this specification.

[0247] In addition, those skilled in the art will understand that the structures of the electronic devices shown in the above figures do not limit the electronic devices. The electronic devices may include more or fewer components than shown, or may combine certain components, or arrange the components differently. For example, the electronic devices may also include radio frequency circuits, input units, sensors, audio circuits, wireless fidelity (WiFi) modules, power supplies, Bluetooth modules, and other components, which are not described in detail here.

[0248] In the embodiments of this specification, the execution entity of each step can be the electronic device described above. Optionally, the execution entity of each step is the operating system of the electronic device. The operating system can be Android, iOS, or other operating systems, and this embodiment of this specification does not limit this.

[0249] The electronic device of the embodiment of this specification may further be equipped with a display device, and the display device may be any device capable of realizing a display function, such as a cathode ray tube display (CR), a light-emitting diode display (LED), an electronic ink screen, a liquid crystal display (LCD), a plasma display panel (PDP), etc. The user may use the display device on the electronic device to view displayed text, images, videos and other information. The electronic device may be a smart phone, a tablet computer, a gaming device, an AR (Augmented Reality) device, a car, a data storage device, an audio playback device, a video playback device, a notebook, a desktop computing device, a wearable device such as an electronic watch, electronic glasses, an electronic helmet, an electronic bracelet, an electronic necklace, electronic clothing and the like.

[0250] exist Figure 9 In the electronic device shown, the processor 110 can be used to call the application stored in the memory 120 and specifically execute the visitor processing method involved in one or more embodiments of this specification.

[0251] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0252] The above disclosure is only a preferred embodiment of this specification, and certainly cannot be used to limit the scope of rights of this specification. Therefore, equivalent changes made according to the claims of this specification are still within the scope covered by this specification.

Claims

1. A visitor processing method, characterized in that: The method comprises: Determine a visitor arrival event based on multimodal scene monitoring information of the monitored scene area; generate a visitor dialogue plan for the visitor object based on the intelligent dialogue model when the user is away from home; and conduct an autonomous dialogue process with the visitor object based on the visitor dialogue plan to obtain autonomous dialogue content; Determining visitor information based on the autonomous conversation content and multimodal scene monitoring information, predicting user reception intentions based on the visitor information using an intelligent conversation model, and adjusting the visitor conversation plan based on the predicted user reception intentions to obtain a future visit negotiation conversation plan; Recommending future visit suggestions to the visitor object based on the future visit negotiation dialogue plan.

2. The method according to claim 1, wherein determining visitor visit information based on the autonomous conversation content and multimodal scene monitoring information comprises: The autonomous conversation content and the multimodal scene monitoring information are subjected to semantic feature fusion to obtain comprehensive features of the visiting scene, and visitor visit item information is generated based on the comprehensive features of the visiting scene and a preset visitor visit item structure.

3. The method according to claim 2, wherein the step of fusing the autonomous conversation content and the multimodal scene monitoring information to obtain a comprehensive visitor scene feature, and generating visitor visit item information based on the comprehensive visitor scene feature and a preset visitor visit item structure, comprises: Performing semantic analysis on the autonomous conversation content to obtain visitor conversation semantic information, and performing scene semantic analysis on the multimodal scene monitoring information to obtain visitor scene semantic information; Perform semantic fusion based on the visitor conversation semantic information and the visitor scene semantic information to obtain a visitor scene comprehensive feature, and determine the visitor identity, visit purpose, and environmental context information based on the visitor scene comprehensive feature; Key visit item fields are extracted based on the visitor identity, the visit purpose and the environmental context information, and the key visit item fields are mapped to a preset visitor visit item structure to obtain visitor visit item information.

4. The method according to claim 1, wherein the step of using an intelligent dialogue model to predict a user's reception intention based on the visitor's visit information comprises: Obtain the user's historical visit interaction records and user visitor setting preference information; Based on the visitor's visit information, the historical visit interaction records and the user visitor setting preference information, an intelligent dialogue large model is used to infer the predicted user reception intention for the visitor object.

5. The method according to claim 4, wherein the method uses an intelligent dialogue model to infer the predicted user reception intention for the visitor object based on the visitor's visit event information, the historical visit interaction records, and the user visitor setting preference information, including: Extracting prediction task elements from the visitor's visit event information, the historical visit interaction records, and the user visitor setting preference information to obtain visit prediction task elements, constructing a visit interest prediction reasoning chain for the visit prediction task elements using an interest prediction task thinking template, and generating a visit interest prediction task prompt word based on the visit interest prediction reasoning chain; The visit interest prediction task prompt words are input into the intelligent dialogue model, and the user's visit interest is gradually inferred by the intelligent dialogue model according to the visit interest prediction reasoning chain to obtain the predicted user reception intention.

6. The method according to claim 1, wherein adjusting the visitor dialogue plan based on the predicted user reception intention to obtain a future visit negotiation dialogue plan comprises: Determine the target user intention type corresponding to the predicted user reception intention, obtain the recommended conversation features corresponding to the target user intention type, and obtain the user's visitor reception preference information and home schedule information; Based on the recommended dialogue characteristics, the visitor reception preference information, the home schedule information and the visitor visit matters information, a visit negotiation plan adjustment prompt is generated, the visit negotiation plan adjustment prompt is input into the intelligent dialogue big model, the intelligent dialogue big model is used to perform feature analysis processing to obtain a plan adjustment context vector, and based on the plan adjustment context vector, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

7. The method according to claim 6, wherein the step of performing feature analysis using the intelligent dialogue model to obtain a plan adjustment context vector comprises: The intelligent dialogue model is used to perform structured extraction of input data to obtain entity nodes and entity relationship edges of incoming items; Constructing a plan adjustment knowledge graph based on the entity nodes and entity relationship edges of the visiting matter, performing node embedding coding processing on the plan adjustment knowledge graph to obtain a node encoding vector, and performing edge relationship association conversion processing on the plan adjustment knowledge graph to obtain a node relationship weight matrix; Based on the node encoding vector and the node relationship weight matrix, an attention mechanism is used to perform optimization to obtain a plan adjustment context vector.

8. The method according to claim 6, wherein adjusting the visitor dialogue plan based on the plan adjustment context vector to obtain a future visit negotiation dialogue plan comprises: Determining a multidimensional dialogue adjustment semantic slot corresponding to the dialogue plan structure of the visitor dialogue plan based on the plan adjustment context vector, and generating a visit negotiation dialogue based on the multidimensional dialogue adjustment semantic slot to obtain an adjusted visit negotiation dialogue; Based on the adjustment of the visit negotiation dialogue, the visitor dialogue plan is adjusted to obtain a future visit negotiation dialogue plan.

9. A visitor processing device, characterized in that: The device comprises: A dialogue module is used to determine a visitor arrival event based on the multimodal scene monitoring information of the monitoring scene area, generate a visitor dialogue plan for the visitor object based on the intelligent dialogue model when the user is away from home, and conduct an autonomous dialogue with the visitor object based on the visitor dialogue plan to obtain autonomous dialogue content; A negotiation module is configured to determine visitor information based on the autonomous conversation content and multimodal scene monitoring information, predict the user's reception intention based on the visitor information using an intelligent conversation model, and adjust the visitor conversation plan based on the predicted user reception intention to obtain a future visit negotiation conversation plan; A recommendation module is used to recommend a future visit suggestion plan to the visitor object based on the future visit negotiation dialogue plan.

10. An electronic device, characterized in that: include: A processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the method steps according to any one of claims 1 to 8.

Images